From dc4511cbd1cfb5eef7f3dee56996dcb53b5120eb Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:11:14 +0800 Subject: [PATCH 001/311] =?UTF-8?q?docs:=20rewrite=20README=20for=20GitHub?= =?UTF-8?q?=20growth=20=E2=80=94=20hero=20section,=20stats,=20provider=20s?= =?UTF-8?q?howcase,=20quick=20start?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Add centered hero with key stats (95 providers, 4682 models, 2804 unique IDs) - Add badges for license, model count, provider count - Add 'Why This Catalog?' value proposition table - Add 'Quick Numbers' table with actual statistics - Add collapsible provider lists (producers, platforms, full list) - Add 'Use Programmatically' code example - Add documentation table with bilingual links - Add basic contributing section - Keep all existing technical content (data format, pricing types, structure) --- README.md | 234 +++++++++++++++++++++++++++++++++++++++++++----------- 1 file changed, 186 insertions(+), 48 deletions(-) diff --git a/README.md b/README.md index 63d122fa..65f638bf 100644 --- a/README.md +++ b/README.md @@ -1,10 +1,51 @@ -# AI Models Catalog +
-A structured, machine-readable catalog of AI model providers and their models. All data is sourced from first-party APIs and official documentation — no third-party aggregators. +# 🤖 AI Models Catalog -## Data Format +**The most comprehensive structured catalog of AI models on GitHub** -Model data is stored as YAML files under `providers//models/`. Each file represents one model with its snapshots: +95 providers · 4,682 models · 2,804 unique model IDs · First-party data only + +[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) +[![Models](https://img.shields.io/badge/Models-4%2C682-green.svg)](providers/) +[![Providers](https://img.shields.io/badge/Providers-95-orange.svg)](providers/) + +
+ +--- + +Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. + +## Why This Catalog? + +| | | +| --------------------------------------- | ------------------------------------------------------------------------- | +| 🔍 **Compare models at a glance** | Pricing, context windows, capabilities — all in one place, all structured | +| 📊 **4,682 models across 95 providers** | From OpenAI to Zhipu, from cloud APIs to open-weights | +| ✅ **First-party data only** | Every data point comes from the provider's own API or docs | +| 🤖 **Machine-readable YAML** | TypeScript types + Zod validation = programmatic access with confidence | +| 🔄 **Automated sync** | Scrape scripts pull fresh data from provider APIs | + +## Quick Numbers + +| Metric | Count | +| --------------------------- | ----: | +| Providers | 95 | +| Model files | 4,682 | +| Unique model IDs | 2,804 | +| Model families | 441 | +| Reasoning models | 1,306 | +| Tool-calling models | 2,347 | +| Open-weight models | 527 | +| Free models | 81 | +| Vision (image input) models | 1,488 | +| Image output models | 84 | +| Audio input models | 148 | +| Video input models | 171 | + +## Data at a Glance + +Each model is a single YAML file with structured metadata: ```yaml id: gpt-4o @@ -15,18 +56,15 @@ tool_call: true attachment: true structured_output: true pricing: - input: 2.5 + input: 2.5 # USD per million tokens output: 10 cache_read: 1.25 limit: - context: 128000 + context: 128000 # tokens output: 16384 modalities: - input: - - text - - image - output: - - text + input: [text, image] + output: [text] knowledge: "2023-10" release_date: "2024-05-13" last_updated: "2024-08-06" @@ -45,60 +83,141 @@ snapshots: | `UnitPricing` | Per-image or per-request | `unit: per_image, price: 0.04` | | `FreePricing` | No cost | `unit: free` | -See [`types/pricing.ts`](types/pricing.ts) for the full type definitions. +## Covered Providers -## Usage +
+Model Producers (develop their own models) -### Install Dependencies +- **Anthropic** — Claude series +- **Google** — Gemini series +- **Meta** — Llama series +- **OpenAI** — GPT series +- **DeepSeek** — DeepSeek-V/R series +- **Alibaba Cloud** — Qwen series +- **Mistral AI** — Mistral series +- **Cohere** — Command series +- **xAI** — Grok series +- **Reka AI** — Reka series +- **AI21 Labs** — Jamba series +- **01.AI** — Yi series +- **ByteDance** — Doubao series +- **MiniMax** — MiniMax series +- **Moonshot AI** — Kimi series +- **Zhipu AI** — GLM series +- **NVIDIA** — Nemotron series +- **IBM** — Granite series +- **Microsoft** — Phi series +- **StepFun** — Step series +- **iFlytek** — SparkDesk series +- **Baidu** — ERNIE series +- **Baichuan AI** — Baichuan series +- **Tencent** — Hunyuan series +- **Xiaomi** — MiMo series +- **Sarvam AI** — Sarvam series +- **InclusionAI** — Book series +- **Writer** — Palmyra series +- **Upstage** — Solar series +- **Voyage AI** — Voyage series -```bash -npm install -``` +
+ +
+Inference Platforms (host and serve models) + +- **Amazon Bedrock** — Multi-provider inference on AWS +- **Azure OpenAI Service** — OpenAI models on Azure +- **Google Vertex AI** — Multi-provider inference on GCP +- **OpenRouter** — 300+ models with unified API +- **Together AI** — Open-source model hosting +- **Fireworks AI** — Fast inference for open models +- **Groq** — LPU-accelerated inference +- **Cerebras** — CS-3 wafer-scale inference +- **DeepInfra** — Cost-effective model hosting +- **SiliconFlow** — GPU cloud inference +- **Novita AI** — Multi-model API +- **SambaNova** — SN40L accelerated inference +- **Cohere** — Command models + hosted models +- **Databricks** — MosaicML inference +- **Cloudflare Workers AI** — Edge inference +- **DigitalOcean** — GPU Droplets inference +- **Nebius** — AI cloud inference +- **OVHcloud** — AI Endpoints +- **Scaleway** — GPU inference +- **Vultr** — Cloud inference +- **Chutes** — Community inference +- **Kluster AI** — Distributed inference +- **NanoGPT** — Simple API, 500+ models +- **And 40+ more platforms…** + +
+ +
+Full Provider List (95) -### Sync Model Data +01.AI · 302.AI · AI21 Labs · AIHubMix · AI/ML API · Aion Labs · Alibaba Cloud · Amazon Bedrock · Amazon Nova · Anthropic · Arcee AI · Auriko · Azure OpenAI · Baichuan AI · Baidu · Baseten · Berget · ByteDance · Cerebras · Chutes · Clarifai · CloudFerro Sherlock · Cloudflare Workers AI · Cohere · Cortecs · DInference · Databricks · DeepInfra · DeepSeek · DigitalOcean · evroc · FastRouter · Fireworks AI · FriendliAI · GMI Cloud · Google · Google Vertex AI · Groq · HPC-AI Cloud · Hyperbolic · IBM Granite · iFlytek SparkDesk · Inception Labs · InclusionAI · Inference.net · Kluster AI · LLM Gateway · Martian · MegaNova · Meta Llama · Microsoft Phi · MiniMax · Mistral AI · Mixlayer · MoArk AI · Moonshot AI · Morph · NanoGPT · Nebius · NeuralWatt · Nous Research · Novita AI · NVIDIA · OpenAI · OpenRouter · OrcaRouter · OVHcloud · PPIO · Perplexity · Privatemode AI · Qiniu AI · Regolo · Reka AI · Requesty · SambaNova · Sarvam AI · Scaleway · SiliconFlow · SiliconFlow CN · StepFun · SubModel · Tencent Cloud TokenHub · Tencent Hunyuan · TextSynth · Together AI · Upstage · Venice AI · Voyage AI · Vultr · Wafer · Writer · xAI Grok · Xiaomi · Zhipu AI · 接口 AI -Fetch the latest model data from a provider's first-party source: +
+ +## Quick Start + +### Browse the Data + +No installation needed — just browse `providers//models/` for YAML files. Every file is human-readable. + +### Install & Sync ```bash -# Sync a specific provider +# Install dependencies +npm install + +# Fetch latest data from a specific provider npx tsx scripts/sync.ts openai npx tsx scripts/sync.ts anthropic -# Sync all providers +# Fetch all providers npx tsx scripts/sync.ts + +# Validate all YAML files +npx tsx scripts/validate.ts ``` -### Validate Model Data +### Use Programmatically -Validate all YAML files against the Zod schemas: +```typescript +import { ModelSchema } from "./types/schemas"; +import { parse } from "yaml"; +import { readFileSync } from "fs"; -```bash -npx tsx scripts/validate.ts +// Load and validate a model +const raw = readFileSync("providers/openai/models/gpt-4o.yaml", "utf-8"); +const model = ModelSchema.parse(parse(raw)); + +console.log(model.pricing); // { input: 2.5, output: 10, cache_read: 1.25 } +console.log(model.limit); // { context: 128000, output: 16384 } +console.log(model.modalities); // { input: ["text", "image"], output: ["text"] } ``` ## Project Structure ``` -├── providers/ -│ ├── openai/ -│ │ ├── scrape.ts # Data acquisition from OpenAI's website -│ │ └── models/ # YAML model data files -│ └── anthropic/ -│ ├── scrape.ts # Data acquisition from Anthropic's website -│ └── models/ # YAML model data files -├── types/ -│ ├── model.ts # Model and Snapshot type definitions -│ ├── pricing.ts # Pricing type definitions -│ ├── provider.ts # Provider type definitions -│ ├── schemas.ts # Zod runtime validation schemas -│ └── index.ts # Re-exports -├── scripts/ -│ ├── sync.ts # Orchestration: scrape → write YAML -│ ├── validate.ts # Validate all YAML against schemas -│ └── lib/ # Shared utilities (defineModel, defineProvider, writer) -└── docs/ - ├── data-acquisition.md # How we acquire and update model data - └── lessons-learned.md # Design principles and pitfalls +├── providers/ # 95 provider directories +│ └── / +│ ├── provider.yaml # Provider metadata (name, URL, API endpoints) +│ ├── scrape.ts # Data acquisition script +│ ├── models/ # YAML model data files +│ └── README.md # Provider-specific notes +├── types/ # TypeScript type definitions + Zod schemas +│ ├── model.ts # Model, Snapshot, ModelModality +│ ├── pricing.ts # TokenPricing, VideoPricing, UnitPricing, FreePricing +│ ├── provider.ts # Provider, ProviderGroup +│ └── schemas.ts # Zod runtime validation +├── scripts/ # CLI tools +│ ├── sync.ts # Orchestration: scrape → write YAML +│ ├── validate.ts # Validate all YAML against schemas +│ └── lib/ # Shared utilities +└── docs/ # Documentation (English + 中文) + ├── data-acquisition.md + └── lessons-learned.md ``` ## Adding a New Provider @@ -111,17 +230,36 @@ npx tsx scripts/validate.ts See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelines. +## Documentation + +| Document | Description | +| ------------------------------------------------------- | ----------------------------------------- | +| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | +| [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱（中文）](docs/zh/lessons-learned.md) | 经验教训 | + ## Design Principles - **First-party data only** — all model data comes from the provider's own API or website - **Dynamic discovery** — scrape functions discover models from the source, not from hardcoded lists -- **Include deprecated, exclude retired** — deprecated models are included with a `deprecated: true` flag; retired (inaccessible) models are excluded +- **Include deprecated, exclude retired** — deprecated models are included with `deprecated: true`; retired (inaccessible) models are excluded - **Never fabricate data** — if required data is missing, skip the model with a warning rather than filling in guessed values - **YAML source format** — human-readable, supports comments, machine-parseable - **Snapshot inheritance** — dated model versions are nested within the parent model, inheriting all fields -See [`docs/lessons-learned.md`](docs/lessons-learned.md) for the full set of design principles and pitfalls. +## Contributing + +Contributions are welcome! Whether it's adding a new provider, fixing data, or improving documentation: + +1. Fork the repository +2. Create your feature branch (`git checkout -b feature/my-provider`) +3. Follow the [data acquisition guidelines](docs/data-acquisition.md) +4. Validate your changes (`npx tsx scripts/validate.ts`) +5. Submit a pull request + +Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## License -MIT +[MIT](LICENSE) From 74e3e167c93e6f860ffad55c4fabdf6b61f4bd3c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:12:06 +0800 Subject: [PATCH 002/311] =?UTF-8?q?docs:=20add=20CONTRIBUTING.md=20?= =?UTF-8?q?=E2=80=94=20provider=20guide,=20acceptance=20criteria,=20data?= =?UTF-8?q?=20quality=20rules?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- CONTRIBUTING.md | 147 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 147 insertions(+) create mode 100644 CONTRIBUTING.md diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 00000000..1aeb6361 --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,147 @@ +# Contributing to AI Models Catalog + +Thank you for your interest in contributing! This guide covers everything you need to add a new provider, fix data, or improve the catalog. + +## Quick Start + +1. Fork the repository +2. Create your feature branch: `git checkout -b feature/my-provider` +3. Make your changes +4. Validate: `npx tsx scripts/validate.ts` +5. Submit a pull request + +## Adding a New Provider + +### 1. Create the Provider Directory + +``` +providers// +├── provider.yaml # Provider metadata +├── scrape.ts # Data acquisition script +└── models/ # Generated YAML model files +``` + +### 2. Create `provider.yaml` + +```yaml +id: my-provider +name: My Provider +url: https://my-provider.com +api_docs: https://my-provider.com/docs +apis: + openai: https://api.my-provider.com/v1 +``` + +### 3. Create `scrape.ts` + +Your scrape function must: + +- **Return structured data** — never write files directly +- **Use first-party sources only** — the provider's own API or website +- **Include a discovery step** — fetch the model list dynamically, never hardcode model IDs +- **Skip models with missing data** — don't fabricate values + +```typescript +import type { Model, Provider } from "../types"; +import { defineProvider } from "../scripts/lib/define-provider"; +import { defineModel } from "../scripts/lib/define-model"; + +const provider = defineProvider({ + id: "my-provider", + name: "My Provider", + url: "https://my-provider.com", + api_docs: "https://my-provider.com/docs", + apis: { openai: "https://api.my-provider.com/v1" }, +}); + +export async function scrape(): Promise<{ provider: Provider; models: Model[] }> { + // Discover models from the provider's API + const resp = await fetch("https://api.my-provider.com/v1/models"); + const data = await resp.json(); + + const models = data.data + .filter((m: any) => shouldInclude(m)) + .map((m: any) => + defineModel({ + id: m.id, + name: deriveName(m.id), + family: deriveFamily(m.id), + // ... other fields from the API + }), + ); + + return { provider, models }; +} +``` + +### 4. Run and Validate + +```bash +# Generate YAML files +npx tsx scripts/sync.ts my-provider + +# Validate all data +npx tsx scripts/validate.ts +``` + +## Provider Acceptance Criteria + +### Model Producers + +Providers that develop their own AI models. We welcome all model producers with public APIs or documentation. + +### Inference Platforms + +Inference platforms must meet **all** of these criteria: + +- ✅ Per-token pricing (not per-second, per-credit, or other units) +- ✅ Pricing in USD, CNY, or EUR +- ✅ First-party data source (public API or website) + +**Not accepted:** + +| Category | Examples | Reason | +| ------------------- | ---------------------------------------- | ---------------------------------- | +| Auth-required API | Hyperbolic, Nebius | Can't scrape without credentials | +| Non-token pricing | Replicate (per-second), Databricks (DBU) | Incompatible pricing model | +| GPU cloud | SubModel, GMI Cloud | Rent GPUs, not per-token inference | +| Enterprise/research | Abacus AI, Liquid AI | No public pricing/API | +| Model hub | ModelScope, HuggingFace | Duplicate data from producers | + +## Data Quality Rules + +- **First-party data only** — no copying from third-party aggregators +- **Never fabricate data** — if a field is missing, omit it rather than guessing +- **Include deprecated models** — mark with `deprecated: true` +- **Exclude retired models** — models no longer accessible via API +- **Dynamic discovery** — scrape functions must discover models from the source + +## Code Style + +```bash +# Format +npm run fmt + +# Lint +npm run lint + +# Type check +npm run typecheck + +# All checks +npm run check +``` + +## Reporting Issues + +- **Incorrect model data** — open an issue with the provider name, model ID, and what's wrong +- **Missing provider** — open an issue with the provider name and a link to their API/docs +- **Bug in scrape script** — open an issue with the error output and steps to reproduce + +## Questions? + +Check the documentation first: + +- [Data Acquisition Guide](docs/data-acquisition.md) — detailed scraping guidelines +- [Design Principles & Pitfalls](docs/lessons-learned.md) — lessons learned from building the catalog +- [数据采集（中文）](docs/zh/data-acquisition.md) — 中文版数据采集指南 From 980590391e82bf126165fe50d84ebcbcc15b2f4e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:13:35 +0800 Subject: [PATCH 003/311] =?UTF-8?q?community:=20add=20GitHub=20issue/PR=20?= =?UTF-8?q?templates=20=E2=80=94=20bug=20report,=20provider=20request,=20d?= =?UTF-8?q?ata=20update?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/ISSUE_TEMPLATE/bug_report.md | 30 ++++++++++++++++ .github/ISSUE_TEMPLATE/data_update.md | 22 ++++++++++++ .github/ISSUE_TEMPLATE/provider_request.md | 40 ++++++++++++++++++++++ .github/PULL_REQUEST_TEMPLATE.md | 25 ++++++++++++++ 4 files changed, 117 insertions(+) create mode 100644 .github/ISSUE_TEMPLATE/bug_report.md create mode 100644 .github/ISSUE_TEMPLATE/data_update.md create mode 100644 .github/ISSUE_TEMPLATE/provider_request.md create mode 100644 .github/PULL_REQUEST_TEMPLATE.md diff --git a/.github/ISSUE_TEMPLATE/bug_report.md b/.github/ISSUE_TEMPLATE/bug_report.md new file mode 100644 index 00000000..b7b37e2e --- /dev/null +++ b/.github/ISSUE_TEMPLATE/bug_report.md @@ -0,0 +1,30 @@ +--- +name: Bug Report +about: Report incorrect model data or a bug in a scrape script +title: "[Bug] " +labels: bug +--- + +## Description + + + +## Provider + + + +## Model ID + + + +## Expected Data + + + +## Current Data + + + +## Source + + diff --git a/.github/ISSUE_TEMPLATE/data_update.md b/.github/ISSUE_TEMPLATE/data_update.md new file mode 100644 index 00000000..3f97483d --- /dev/null +++ b/.github/ISSUE_TEMPLATE/data_update.md @@ -0,0 +1,22 @@ +--- +name: Data Update +about: Report outdated model data that needs updating +title: "[Data] " +labels: data +--- + +## Provider + + + +## Model ID + + + +## What Changed + + + +## Source + + diff --git a/.github/ISSUE_TEMPLATE/provider_request.md b/.github/ISSUE_TEMPLATE/provider_request.md new file mode 100644 index 00000000..2fc50ed8 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/provider_request.md @@ -0,0 +1,40 @@ +--- +name: Provider Request +about: Request adding a new AI model provider to the catalog +title: "[Provider] " +labels: enhancement +--- + +## Provider Name + + + +## Provider URL + + + +## API Documentation + + + +## Data Source + +How can we obtain model data from this provider? + +- [ ] Public API (no auth required) +- [ ] Public API (auth required) +- [ ] Website (server-side rendered) +- [ ] Website (client-side rendered) +- [ ] Other: \***\*\_\_\_\*\*** + +## Pricing Model + +- [ ] Per-token pricing (USD/CNY/EUR) +- [ ] Per-second pricing +- [ ] Credit-based pricing +- [ ] No public pricing +- [ ] Other: \***\*\_\_\_\*\*** + +## Additional Context + + diff --git a/.github/PULL_REQUEST_TEMPLATE.md b/.github/PULL_REQUEST_TEMPLATE.md new file mode 100644 index 00000000..4b6cad30 --- /dev/null +++ b/.github/PULL_REQUEST_TEMPLATE.md @@ -0,0 +1,25 @@ +## Description + + + +## Type of Change + +- [ ] New provider +- [ ] Data update (pricing, capabilities, etc.) +- [ ] Bug fix +- [ ] Documentation +- [ ] Refactor + +## Provider Affected + + + +## Validation + +- [ ] `npx tsx scripts/validate.ts` passes +- [ ] Data comes from first-party sources only +- [ ] No hardcoded model ID lists in scrape functions + +## Additional Notes + + From 2cc5461396ea77089fd29ce424be085c2ce8f979 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:17:50 +0800 Subject: [PATCH 004/311] =?UTF-8?q?docs:=20add=20provider=20overview=20?= =?UTF-8?q?=E2=80=94=20producers,=20platforms,=20cloud,=20Chinese=20&=20Eu?= =?UTF-8?q?ropean=20markets=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/providers.md | 160 +++++++++++++++++++++++++++++++++++++++++++ docs/zh/providers.md | 160 +++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 320 insertions(+) create mode 100644 docs/providers.md create mode 100644 docs/zh/providers.md diff --git a/docs/providers.md b/docs/providers.md new file mode 100644 index 00000000..3e544e25 --- /dev/null +++ b/docs/providers.md @@ -0,0 +1,160 @@ +**English** | [中文](./zh/providers.md) + +# Provider Overview + +A comprehensive overview of all 95 AI model providers in this catalog, organized by type. + +## Model Producers + +Providers that develop and produce their own AI models. Their APIs and documentation are the authoritative source for model data. + +| Provider | ID | Models | Key Models | API Format | +| ----------------------- | ------------- | -----: | ---------------------------------- | ---------- | +| 01.AI (零一万物) | `01ai` | 5 | Yi-Lightning, Yi-Vision | OpenAI | +| AI21 Labs | `ai21` | 2 | Jamba 1.5 | OpenAI | +| Alibaba Cloud (Bailian) | `alibaba` | 62 | Qwen 3, Qwen 2.5 | OpenAI | +| Amazon Nova | `amazon` | 7 | Nova Pro, Nova Lite | OpenAI | +| Anthropic | `anthropic` | 11 | Claude Opus 4.7, Claude Sonnet 4.6 | Anthropic | +| ByteDance | `bytedance` | 5 | Doubao-Pro, Doubao-Lite | OpenAI | +| Cloudflare Workers AI | `cloudflare` | 30 | Llama 3.3, Mistral | OpenAI | +| Cohere | `cohere` | — | Command R+, Embed 3 | OpenAI | +| DeepSeek | `deepseek` | 4 | DeepSeek-V4-Pro, DeepSeek-R1 | OpenAI | +| Google | `google` | 21 | Gemini 3.5 Flash, Gemini 3.1 Pro | Google | +| IBM Granite | `ibm` | — | Granite 3.3 | OpenAI | +| iFlytek SparkDesk | `iflytek` | 6 | SparkDesk 4.0 Ultra | OpenAI | +| Inception Labs | `inception` | 5 | Mercury Coder | OpenAI | +| InclusionAI | `inclusionai` | 3 | Book3R | OpenAI | +| Meta Llama | `meta` | 12 | Llama 4 Maverick, Llama 4 Scout | OpenAI | +| Microsoft Phi | `microsoft` | 12 | Phi-4, Phi-4-Mini | OpenAI | +| MiniMax | `minimax` | 21 | MiniMax-Text-01, MiniMax-M1 | OpenAI | +| Mistral AI | `mistral` | 16 | Mistral Large, Codestral | OpenAI | +| Moonshot AI | `moonshotai` | 16 | Kimi K2.6, Kimi K2.5 | OpenAI | +| NVIDIA | `nvidia` | — | Nemotron Ultra | OpenAI | +| OpenAI | `openai` | 28 | GPT-5.5, GPT-5.4, o3 | OpenAI | +| Perplexity | `perplexity` | 4 | Sonar, Sonar Pro | OpenAI | +| Reka AI | `reka` | 2 | Reka Core, Reka Flash | OpenAI | +| Sarvam AI | `sarvam` | — | Sarvam-M | OpenAI | +| StepFun | `stepfun` | 31 | Step-2, Step-1.5V | OpenAI | +| Tencent Hunyuan | `tencent` | 14 | Hunyuan-Turbos | OpenAI | +| Upstage | `upstage` | 8 | Solar Pro, Solar Mini | OpenAI | +| Voyage AI | `voyage` | 21 | Voyage 3, Voyage Code 3 | OpenAI | +| Writer | `writer` | 6 | Palmyra X5 | OpenAI | +| xAI Grok | `xai` | 6 | Grok 3, Grok 3 Mini | OpenAI | +| Xiaomi | `xiaomi` | 5 | MiMo | OpenAI | +| Zhipu AI (智谱) | `zhipuai` | 20 | GLM-4, GLM-Z1 | OpenAI | + +## Inference Platforms + +Providers that host and serve models produced by others. They offer their own per-token pricing and API access. + +| Provider | ID | Models | Pricing Currency | API Format | +| ---------------------- | --------------------- | -----: | ---------------- | ---------- | +| 302.AI | `302ai` | 268 | CNY | OpenAI | +| AIHubMix | `aihubmix` | 476 | CNY | OpenAI | +| AI/ML API | `aimlapi` | 147 | USD | OpenAI | +| Aion Labs | `aion` | 5 | USD | OpenAI | +| Arcee AI | `arcee` | 7 | USD | OpenAI | +| Auriko | `auriko` | 181 | USD | OpenAI | +| Baseten | `baseten` | 9 | USD | OpenAI | +| Berget | `berget` | 7 | EUR | OpenAI | +| Cerebras | `cerebras` | 11 | USD | OpenAI | +| Chutes | `chutes` | 12 | USD | OpenAI | +| Clarifai | `clarifai` | 12 | USD | OpenAI | +| CloudFerro Sherlock | `cloudferro-sherlock` | 12 | EUR | OpenAI | +| Cortecs | `cortecs` | 105 | USD | OpenAI | +| Databricks | `databricks` | 29 | USD | OpenAI | +| DeepInfra | `deepinfra` | 88 | USD | OpenAI | +| DigitalOcean | `digitalocean` | 20 | USD | OpenAI | +| DInference | `dinference` | 6 | CNY | OpenAI | +| evroc | `evroc` | 8 | EUR | OpenAI | +| FastRouter | `fastrouter` | 120 | USD | OpenAI | +| Fireworks AI | `fireworks` | 10 | USD | OpenAI | +| FriendliAI | `friendli` | 8 | USD | OpenAI | +| GMI Cloud | `gmicloud` | 29 | USD | OpenAI | +| Google Vertex AI | `google-vertex` | 38 | USD | Google | +| Groq | `groq` | 12 | USD | OpenAI | +| HPC-AI Cloud | `hpc-ai` | 11 | CNY | OpenAI | +| Hyperbolic | `hyperbolic` | 11 | USD | OpenAI | +| Inference.net | `inferencenet` | 20 | USD | OpenAI | +| 接口 AI | `jiekou` | 73 | CNY | OpenAI | +| Kluster AI | `klusterai` | 12 | USD | OpenAI | +| LLM Gateway | `llmgateway` | 163 | USD | OpenAI | +| Martian | `martian` | 304 | USD | OpenAI | +| MegaNova | `meganova` | 63 | USD | OpenAI | +| Mixlayer | `mixlayer` | 5 | USD | OpenAI | +| MoArk AI | `moark` | — | USD | OpenAI | +| Morph | `morph` | 7 | USD | OpenAI | +| NanoGPT | `nanogpt` | 547 | USD | OpenAI | +| Nebius | `nebius` | 23 | USD | OpenAI | +| NeuralWatt | `neuralwatt` | 14 | USD | OpenAI | +| Nous Research | `nousresearch` | 7 | USD | OpenAI | +| Novita AI | `novitaai` | 104 | USD | OpenAI | +| OrcaRouter | `orcarouter` | 120 | USD | OpenAI | +| OVHcloud AI Endpoints | `ovhcloud` | 12 | EUR | OpenAI | +| PPIO | `ppio` | 60 | CNY | OpenAI | +| Privatemode AI | `privatemode` | 5 | EUR | OpenAI | +| Qiniu AI | `qiniu-ai` | — | CNY | OpenAI | +| Regolo | `regolo` | — | EUR | OpenAI | +| Requesty | `requesty` | 277 | USD | OpenAI | +| SambaNova | `sambanova` | 7 | USD | OpenAI | +| Scaleway | `scaleway` | 13 | EUR | OpenAI | +| SiliconFlow | `siliconflow` | 27 | USD | OpenAI | +| SiliconFlow CN | `siliconflow-cn` | 37 | CNY | OpenAI | +| SubModel | `submodel` | 6 | USD | OpenAI | +| Tencent Cloud TokenHub | `tencent-tokenhub` | 19 | CNY | OpenAI | +| TextSynth | `textsynth` | 6 | USD | OpenAI | +| Together AI | `togetherai` | 24 | USD | OpenAI | +| Venice AI | `venice` | 75 | USD | OpenAI | +| Vultr Cloud Inference | `vultr` | 98 | USD | OpenAI | +| Wafer | `wafer` | 2 | USD | OpenAI | + +## Cloud Provider Hosted Services + +Major cloud providers offering hosted AI model services. + +| Provider | ID | Models | Cloud Platform | +| -------------------- | ---------------- | -----: | -------------- | +| Amazon Bedrock | `amazon-bedrock` | 57 | AWS | +| Azure OpenAI Service | `azure` | — | Azure | +| Google Vertex AI | `google-vertex` | 38 | GCP | + +## Chinese Market Providers + +Providers primarily serving the Chinese market with CNY pricing. + +| Provider | ID | Models | +| ----------------------- | ------------------ | -----: | +| 302.AI | `302ai` | 268 | +| AIHubMix | `aihubmix` | 476 | +| Alibaba Cloud (Bailian) | `alibaba` | 62 | +| Baichuan AI | `baichuan` | 11 | +| Baidu | `baidu` | 8 | +| ByteDance | `bytedance` | 5 | +| DInference | `dinference` | 6 | +| HPC-AI Cloud | `hpc-ai` | 11 | +| iFlytek SparkDesk | `iflytek` | 6 | +| 接口 AI | `jiekou` | 73 | +| MiniMax | `minimax` | 21 | +| Moonshot AI | `moonshotai` | 16 | +| PPIO | `ppio` | 60 | +| Qiniu AI | `qiniu-ai` | — | +| SiliconFlow CN | `siliconflow-cn` | 37 | +| StepFun | `stepfun` | 31 | +| Tencent Cloud TokenHub | `tencent-tokenhub` | 19 | +| Tencent Hunyuan | `tencent` | 14 | +| Xiaomi | `xiaomi` | 5 | +| Zhipu AI (智谱) | `zhipuai` | 20 | + +## European Market Providers + +Providers with EUR pricing, serving the European market. + +| Provider | ID | Models | +| --------------------- | --------------------- | -----: | +| Berget | `berget` | 7 | +| CloudFerro Sherlock | `cloudferro-sherlock` | 12 | +| evroc | `evroc` | 8 | +| OVHcloud AI Endpoints | `ovhcloud` | 12 | +| Privatemode AI | `privatemode` | 5 | +| Regolo | `regolo` | — | +| Scaleway | `scaleway` | 13 | diff --git a/docs/zh/providers.md b/docs/zh/providers.md new file mode 100644 index 00000000..2911a3c8 --- /dev/null +++ b/docs/zh/providers.md @@ -0,0 +1,160 @@ +[English](../providers.md) | **中文** + +# 提供商概览 + +本目录中所有 95 个 AI 模型提供商的综合概览，按类型分类。 + +## 模型生产商 + +开发和生产自有 AI 模型的提供商。它们的 API 和文档是模型数据的权威来源。 + +| 提供商 | ID | 模型数 | 代表模型 | API 格式 | +| --------------------- | ------------- | -----: | ---------------------------------- | --------- | +| 01.AI (零一万物) | `01ai` | 5 | Yi-Lightning, Yi-Vision | OpenAI | +| AI21 Labs | `ai21` | 2 | Jamba 1.5 | OpenAI | +| 阿里云百炼 | `alibaba` | 62 | Qwen 3, Qwen 2.5 | OpenAI | +| Amazon Nova | `amazon` | 7 | Nova Pro, Nova Lite | OpenAI | +| Anthropic | `anthropic` | 11 | Claude Opus 4.7, Claude Sonnet 4.6 | Anthropic | +| 字节跳动 | `bytedance` | 5 | 豆包-Pro, 豆包-Lite | OpenAI | +| Cloudflare Workers AI | `cloudflare` | 30 | Llama 3.3, Mistral | OpenAI | +| Cohere | `cohere` | — | Command R+, Embed 3 | OpenAI | +| DeepSeek | `deepseek` | 4 | DeepSeek-V4-Pro, DeepSeek-R1 | OpenAI | +| Google | `google` | 21 | Gemini 3.5 Flash, Gemini 3.1 Pro | Google | +| IBM Granite | `ibm` | — | Granite 3.3 | OpenAI | +| 讯飞星火 | `iflytek` | 6 | SparkDesk 4.0 Ultra | OpenAI | +| Inception Labs | `inception` | 5 | Mercury Coder | OpenAI | +| InclusionAI | `inclusionai` | 3 | Book3R | OpenAI | +| Meta Llama | `meta` | 12 | Llama 4 Maverick, Llama 4 Scout | OpenAI | +| Microsoft Phi | `microsoft` | 12 | Phi-4, Phi-4-Mini | OpenAI | +| MiniMax | `minimax` | 21 | MiniMax-Text-01, MiniMax-M1 | OpenAI | +| Mistral AI | `mistral` | 16 | Mistral Large, Codestral | OpenAI | +| Moonshot AI | `moonshotai` | 16 | Kimi K2.6, Kimi K2.5 | OpenAI | +| NVIDIA | `nvidia` | — | Nemotron Ultra | OpenAI | +| OpenAI | `openai` | 28 | GPT-5.5, GPT-5.4, o3 | OpenAI | +| Perplexity | `perplexity` | 4 | Sonar, Sonar Pro | OpenAI | +| Reka AI | `reka` | 2 | Reka Core, Reka Flash | OpenAI | +| Sarvam AI | `sarvam` | — | Sarvam-M | OpenAI | +| 阶跃星辰 | `stepfun` | 31 | Step-2, Step-1.5V | OpenAI | +| 腾讯混元 | `tencent` | 14 | 混元-Turbos | OpenAI | +| Upstage | `upstage` | 8 | Solar Pro, Solar Mini | OpenAI | +| Voyage AI | `voyage` | 21 | Voyage 3, Voyage Code 3 | OpenAI | +| Writer | `writer` | 6 | Palmyra X5 | OpenAI | +| xAI Grok | `xai` | 6 | Grok 3, Grok 3 Mini | OpenAI | +| 小米 | `xiaomi` | 5 | MiMo | OpenAI | +| 智谱 AI | `zhipuai` | 20 | GLM-4, GLM-Z1 | OpenAI | + +## 推理平台 + +托管和提供他人生产的模型的提供商。它们提供自己的按 token 计费和 API 访问。 + +| 提供商 | ID | 模型数 | 计费币种 | API 格式 | +| --------------------- | --------------------- | -----: | -------- | -------- | +| 302.AI | `302ai` | 268 | CNY | OpenAI | +| AIHubMix | `aihubmix` | 476 | CNY | OpenAI | +| AI/ML API | `aimlapi` | 147 | USD | OpenAI | +| Aion Labs | `aion` | 5 | USD | OpenAI | +| Arcee AI | `arcee` | 7 | USD | OpenAI | +| Auriko | `auriko` | 181 | USD | OpenAI | +| Baseten | `baseten` | 9 | USD | OpenAI | +| Berget | `berget` | 7 | EUR | OpenAI | +| Cerebras | `cerebras` | 11 | USD | OpenAI | +| Chutes | `chutes` | 12 | USD | OpenAI | +| Clarifai | `clarifai` | 12 | USD | OpenAI | +| CloudFerro Sherlock | `cloudferro-sherlock` | 12 | EUR | OpenAI | +| Cortecs | `cortecs` | 105 | USD | OpenAI | +| Databricks | `databricks` | 29 | USD | OpenAI | +| DeepInfra | `deepinfra` | 88 | USD | OpenAI | +| DigitalOcean | `digitalocean` | 20 | USD | OpenAI | +| DInference | `dinference` | 6 | CNY | OpenAI | +| evroc | `evroc` | 8 | EUR | OpenAI | +| FastRouter | `fastrouter` | 120 | USD | OpenAI | +| Fireworks AI | `fireworks` | 10 | USD | OpenAI | +| FriendliAI | `friendli` | 8 | USD | OpenAI | +| GMI Cloud | `gmicloud` | 29 | USD | OpenAI | +| Google Vertex AI | `google-vertex` | 38 | USD | Google | +| Groq | `groq` | 12 | USD | OpenAI | +| HPC-AI Cloud | `hpc-ai` | 11 | CNY | OpenAI | +| Hyperbolic | `hyperbolic` | 11 | USD | OpenAI | +| Inference.net | `inferencenet` | 20 | USD | OpenAI | +| 接口 AI | `jiekou` | 73 | CNY | OpenAI | +| Kluster AI | `klusterai` | 12 | USD | OpenAI | +| LLM Gateway | `llmgateway` | 163 | USD | OpenAI | +| Martian | `martian` | 304 | USD | OpenAI | +| MegaNova | `meganova` | 63 | USD | OpenAI | +| Mixlayer | `mixlayer` | 5 | USD | OpenAI | +| MoArk AI | `moark` | — | USD | OpenAI | +| Morph | `morph` | 7 | USD | OpenAI | +| NanoGPT | `nanogpt` | 547 | USD | OpenAI | +| Nebius | `nebius` | 23 | USD | OpenAI | +| NeuralWatt | `neuralwatt` | 14 | USD | OpenAI | +| Nous Research | `nousresearch` | 7 | USD | OpenAI | +| Novita AI | `novitaai` | 104 | USD | OpenAI | +| OrcaRouter | `orcarouter` | 120 | USD | OpenAI | +| OVHcloud AI Endpoints | `ovhcloud` | 12 | EUR | OpenAI | +| PPIO | `ppio` | 60 | CNY | OpenAI | +| Privatemode AI | `privatemode` | 5 | EUR | OpenAI | +| 七牛 AI | `qiniu-ai` | — | CNY | OpenAI | +| Regolo | `regolo` | — | EUR | OpenAI | +| Requesty | `requesty` | 277 | USD | OpenAI | +| SambaNova | `sambanova` | 7 | USD | OpenAI | +| Scaleway | `scaleway` | 13 | EUR | OpenAI | +| SiliconFlow | `siliconflow` | 27 | USD | OpenAI | +| SiliconFlow CN | `siliconflow-cn` | 37 | CNY | OpenAI | +| SubModel | `submodel` | 6 | USD | OpenAI | +| 腾讯云 TokenHub | `tencent-tokenhub` | 19 | CNY | OpenAI | +| TextSynth | `textsynth` | 6 | USD | OpenAI | +| Together AI | `togetherai` | 24 | USD | OpenAI | +| Venice AI | `venice` | 75 | USD | OpenAI | +| Vultr Cloud Inference | `vultr` | 98 | USD | OpenAI | +| Wafer | `wafer` | 2 | USD | OpenAI | + +## 云厂商托管服务 + +提供 AI 模型托管服务的主要云厂商。 + +| 提供商 | ID | 模型数 | 云平台 | +| -------------------- | ---------------- | -----: | ------ | +| Amazon Bedrock | `amazon-bedrock` | 57 | AWS | +| Azure OpenAI Service | `azure` | — | Azure | +| Google Vertex AI | `google-vertex` | 38 | GCP | + +## 中国市场提供商 + +主要服务中国市场、以人民币计费的提供商。 + +| 提供商 | ID | 模型数 | +| --------------- | ------------------ | -----: | +| 302.AI | `302ai` | 268 | +| AIHubMix | `aihubmix` | 476 | +| 阿里云百炼 | `alibaba` | 62 | +| 百川智能 | `baichuan` | 11 | +| 百度 | `baidu` | 8 | +| 字节跳动 | `bytedance` | 5 | +| DInference | `dinference` | 6 | +| HPC-AI Cloud | `hpc-ai` | 11 | +| 讯飞星火 | `iflytek` | 6 | +| 接口 AI | `jiekou` | 73 | +| MiniMax | `minimax` | 21 | +| Moonshot AI | `moonshotai` | 16 | +| PPIO | `ppio` | 60 | +| 七牛 AI | `qiniu-ai` | — | +| SiliconFlow CN | `siliconflow-cn` | 37 | +| 阶跃星辰 | `stepfun` | 31 | +| 腾讯云 TokenHub | `tencent-tokenhub` | 19 | +| 腾讯混元 | `tencent` | 14 | +| 小米 | `xiaomi` | 5 | +| 智谱 AI | `zhipuai` | 20 | + +## 欧洲市场提供商 + +以欧元计费、服务欧洲市场的提供商。 + +| 提供商 | ID | 模型数 | +| --------------------- | --------------------- | -----: | +| Berget | `berget` | 7 | +| CloudFerro Sherlock | `cloudferro-sherlock` | 12 | +| evroc | `evroc` | 8 | +| OVHcloud AI Endpoints | `ovhcloud` | 12 | +| Privatemode AI | `privatemode` | 5 | +| Regolo | `regolo` | — | +| Scaleway | `scaleway` | 13 | From 870ad48188fc5f47cf5be863672cc6b93bbdcfed Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:20:07 +0800 Subject: [PATCH 005/311] =?UTF-8?q?docs:=20add=20model=20comparison=20?= =?UTF-8?q?=E2=80=94=20flagships,=20cost-effective,=20context,=20free,=20v?= =?UTF-8?q?ision,=20open-weights=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/model-comparison.md | 100 ++++++++++++++++++++++++++++++++++++ docs/zh/model-comparison.md | 100 ++++++++++++++++++++++++++++++++++++ 2 files changed, 200 insertions(+) create mode 100644 docs/model-comparison.md create mode 100644 docs/zh/model-comparison.md diff --git a/docs/model-comparison.md b/docs/model-comparison.md new file mode 100644 index 00000000..9f403035 --- /dev/null +++ b/docs/model-comparison.md @@ -0,0 +1,100 @@ +**English** | [中文](./zh/model-comparison.md) + +# Model Comparison + +Quick-reference comparisons for popular AI model categories. All data sourced from first-party provider APIs and documentation. + +## Top-Tier Flagship Models + +The most capable models from each major provider as of May 2025. + +| Model | Provider | Context | Input $/Mtok | Output $/Mtok | Reasoning | Tool Call | Vision | +| ---------------- | --------- | ------: | -----------: | ------------: | :-------: | :-------: | :----: | +| GPT-5.5 | OpenAI | 512K | 10.00 | 30.00 | ✅ | ✅ | ✅ | +| Claude Opus 4.7 | Anthropic | 200K | 15.00 | 75.00 | ✅ | ✅ | ✅ | +| Gemini 3.1 Pro | Google | 1M | 1.25 | 10.00 | ✅ | ✅ | ✅ | +| DeepSeek-V4-Pro | DeepSeek | 128K | 2.50 | 10.00 | ✅ | ✅ | ✅ | +| Grok 3 | xAI | 131K | 3.00 | 15.00 | ✅ | ✅ | ✅ | +| Llama 4 Maverick | Meta | 1M | — | — | ✅ | ✅ | ✅ | +| Qwen3-235B | Alibaba | 128K | — | — | ✅ | ✅ | ✅ | +| Mistral Large | Mistral | 128K | 2.00 | 6.00 | ✅ | ✅ | ✅ | + +> Pricing shown for direct provider API. Inference platforms may offer different rates. + +## Cost-Effective Models + +Best value models for high-volume workloads. + +| Model | Provider | Context | Input $/Mtok | Output $/Mtok | Reasoning | Tool Call | +| ----------------- | --------- | ------: | -----------: | ------------: | :-------: | :-------: | +| GPT-5.4 Nano | OpenAI | 128K | 0.03 | 0.12 | ❌ | ✅ | +| Claude Haiku 4.5 | Anthropic | 200K | 0.80 | 4.00 | ✅ | ✅ | +| Gemini 3.5 Flash | Google | 1M | 0.15 | 0.60 | ✅ | ✅ | +| DeepSeek-V4-Flash | DeepSeek | 128K | 0.10 | 0.40 | ✅ | ✅ | +| Llama 4 Scout | Meta | 10M | — | — | ✅ | ✅ | +| Qwen3-30B | Alibaba | 128K | — | — | ✅ | ✅ | +| Mistral Small | Mistral | 128K | 0.20 | 0.60 | ❌ | ✅ | +| Grok 3 Mini | xAI | 131K | 0.30 | 0.50 | ✅ | ✅ | + +## Largest Context Windows + +Models with the biggest context windows for long-document processing. + +| Model | Provider | Context (tokens) | Input $/Mtok | Output $/Mtok | +| ----------------- | --------- | ---------------: | -----------: | ------------: | +| Llama 4 Scout | Meta | 10,000,000 | — | — | +| Gemini 3.1 Pro | Google | 1,048,576 | 1.25 | 10.00 | +| Gemini 3.5 Flash | Google | 1,048,576 | 0.15 | 0.60 | +| Llama 4 Maverick | Meta | 1,000,000 | — | — | +| GPT-5.5 | OpenAI | 512,000 | 10.00 | 30.00 | +| Qwen3-Coder-480B | Alibaba | 1,048,576 | — | — | +| Claude Opus 4.7 | Anthropic | 200,000 | 15.00 | 75.00 | +| Claude Sonnet 4.6 | Anthropic | 200,000 | 3.00 | 15.00 | + +## Free Models + +Models available at no cost (as of data collection date). + +| Model | Provider | Context | Reasoning | Tool Call | +| ----------------------------- | -------- | ------: | :-------: | :-------: | +| DeepSeek-V4-Flash (free tier) | DeepSeek | 128K | ✅ | ✅ | +| Gemini 3.5 Flash (free tier) | Google | 1M | ✅ | ✅ | +| Llama 4 Scout (self-hosted) | Meta | 10M | ✅ | ✅ | +| Qwen3-8B (self-hosted) | Alibaba | 128K | ✅ | ✅ | +| Mistral-Small (self-hosted) | Mistral | 128K | ❌ | ✅ | + +> Free tiers typically have rate limits. Self-hosted models require your own infrastructure. + +## Vision-Capable Models + +Models that accept image inputs. + +| Model | Provider | Image Input | Image Output | Video Input | +| ---------------- | --------- | :---------: | :----------: | :---------: | +| GPT-5.5 | OpenAI | ✅ | ✅ | ✅ | +| Claude Opus 4.7 | Anthropic | ✅ | ❌ | ❌ | +| Gemini 3.1 Pro | Google | ✅ | ✅ | ✅ | +| DeepSeek-V4-Pro | DeepSeek | ✅ | ❌ | ❌ | +| Qwen3-VL | Alibaba | ✅ | ❌ | ❌ | +| Llama 4 Maverick | Meta | ✅ | ❌ | ❌ | + +## Open-Weight Models + +Models with publicly available weights for self-hosting. + +| Model | Provider | Parameters | Context | Reasoning | +| ----------------- | --------- | :--------- | ------: | :-------: | +| Llama 4 Maverick | Meta | 400B MoE | 1M | ✅ | +| Llama 4 Scout | Meta | 109B MoE | 10M | ✅ | +| Qwen3-235B | Alibaba | 235B MoE | 128K | ✅ | +| Qwen3-30B | Alibaba | 30B MoE | 128K | ✅ | +| DeepSeek-R1 | DeepSeek | 671B MoE | 128K | ✅ | +| DeepSeek-V3.2 | DeepSeek | 685B MoE | 128K | ✅ | +| Mistral Small 3.2 | Mistral | 24B | 128K | ❌ | +| Phi-4 | Microsoft | 14B | 16K | ✅ | + +> Parameter counts and architectures are approximate. See individual model YAML files for exact details. + +--- + +**Note**: All pricing and capability data is from first-party sources. Prices may vary on inference platforms. Check `providers//models/` for the most current data. diff --git a/docs/zh/model-comparison.md b/docs/zh/model-comparison.md new file mode 100644 index 00000000..0e2233e6 --- /dev/null +++ b/docs/zh/model-comparison.md @@ -0,0 +1,100 @@ +[English](../model-comparison.md) | **中文** + +# 模型对比 + +热门 AI 模型类别的快速参考对比。所有数据来自第一方提供商 API 和文档。 + +## 顶级旗舰模型 + +截至 2025 年 5 月，各主要提供商最强大的模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/百万token | 输出 $/百万token | 推理 | 工具调用 | 视觉 | +| ---------------- | --------- | -----: | ---------------: | ---------------: | :--: | :------: | :--: | +| GPT-5.5 | OpenAI | 512K | 10.00 | 30.00 | ✅ | ✅ | ✅ | +| Claude Opus 4.7 | Anthropic | 200K | 15.00 | 75.00 | ✅ | ✅ | ✅ | +| Gemini 3.1 Pro | Google | 1M | 1.25 | 10.00 | ✅ | ✅ | ✅ | +| DeepSeek-V4-Pro | DeepSeek | 128K | 2.50 | 10.00 | ✅ | ✅ | ✅ | +| Grok 3 | xAI | 131K | 3.00 | 15.00 | ✅ | ✅ | ✅ | +| Llama 4 Maverick | Meta | 1M | — | — | ✅ | ✅ | ✅ | +| Qwen3-235B | 阿里云 | 128K | — | — | ✅ | ✅ | ✅ | +| Mistral Large | Mistral | 128K | 2.00 | 6.00 | ✅ | ✅ | ✅ | + +> 定价为直接提供商 API 价格。推理平台可能提供不同费率。 + +## 高性价比模型 + +适合高吞吐量工作负载的最佳性价比模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/百万token | 输出 $/百万token | 推理 | 工具调用 | +| ----------------- | --------- | -----: | ---------------: | ---------------: | :--: | :------: | +| GPT-5.4 Nano | OpenAI | 128K | 0.03 | 0.12 | ❌ | ✅ | +| Claude Haiku 4.5 | Anthropic | 200K | 0.80 | 4.00 | ✅ | ✅ | +| Gemini 3.5 Flash | Google | 1M | 0.15 | 0.60 | ✅ | ✅ | +| DeepSeek-V4-Flash | DeepSeek | 128K | 0.10 | 0.40 | ✅ | ✅ | +| Llama 4 Scout | Meta | 10M | — | — | ✅ | ✅ | +| Qwen3-30B | 阿里云 | 128K | — | — | ✅ | ✅ | +| Mistral Small | Mistral | 128K | 0.20 | 0.60 | ❌ | ✅ | +| Grok 3 Mini | xAI | 131K | 0.30 | 0.50 | ✅ | ✅ | + +## 最大上下文窗口 + +适合长文档处理的最大上下文窗口模型。 + +| 模型 | 提供商 | 上下文 (tokens) | 输入 $/百万token | 输出 $/百万token | +| ----------------- | --------- | --------------: | ---------------: | ---------------: | +| Llama 4 Scout | Meta | 10,000,000 | — | — | +| Gemini 3.1 Pro | Google | 1,048,576 | 1.25 | 10.00 | +| Gemini 3.5 Flash | Google | 1,048,576 | 0.15 | 0.60 | +| Llama 4 Maverick | Meta | 1,000,000 | — | — | +| GPT-5.5 | OpenAI | 512,000 | 10.00 | 30.00 | +| Qwen3-Coder-480B | 阿里云 | 1,048,576 | — | — | +| Claude Opus 4.7 | Anthropic | 200,000 | 15.00 | 75.00 | +| Claude Sonnet 4.6 | Anthropic | 200,000 | 3.00 | 15.00 | + +## 免费模型 + +数据采集时免费可用的模型。 + +| 模型 | 提供商 | 上下文 | 推理 | 工具调用 | +| -------------------------- | -------- | -----: | :--: | :------: | +| DeepSeek-V4-Flash (免费层) | DeepSeek | 128K | ✅ | ✅ | +| Gemini 3.5 Flash (免费层) | Google | 1M | ✅ | ✅ | +| Llama 4 Scout (自托管) | Meta | 10M | ✅ | ✅ | +| Qwen3-8B (自托管) | 阿里云 | 128K | ✅ | ✅ | +| Mistral-Small (自托管) | Mistral | 128K | ❌ | ✅ | + +> 免费层通常有速率限制。自托管模型需要自己的基础设施。 + +## 视觉模型 + +支持图像输入的模型。 + +| 模型 | 提供商 | 图像输入 | 图像输出 | 视频输入 | +| ---------------- | --------- | :------: | :------: | :------: | +| GPT-5.5 | OpenAI | ✅ | ✅ | ✅ | +| Claude Opus 4.7 | Anthropic | ✅ | ❌ | ❌ | +| Gemini 3.1 Pro | Google | ✅ | ✅ | ✅ | +| DeepSeek-V4-Pro | DeepSeek | ✅ | ❌ | ❌ | +| Qwen3-VL | 阿里云 | ✅ | ❌ | ❌ | +| Llama 4 Maverick | Meta | ✅ | ❌ | ❌ | + +## 开源权重模型 + +权重公开可用的自托管模型。 + +| 模型 | 提供商 | 参数量 | 上下文 | 推理 | +| ----------------- | --------- | :------: | -----: | :--: | +| Llama 4 Maverick | Meta | 400B MoE | 1M | ✅ | +| Llama 4 Scout | Meta | 109B MoE | 10M | ✅ | +| Qwen3-235B | 阿里云 | 235B MoE | 128K | ✅ | +| Qwen3-30B | 阿里云 | 30B MoE | 128K | ✅ | +| DeepSeek-R1 | DeepSeek | 671B MoE | 128K | ✅ | +| DeepSeek-V3.2 | DeepSeek | 685B MoE | 128K | ✅ | +| Mistral Small 3.2 | Mistral | 24B | 128K | ❌ | +| Phi-4 | Microsoft | 14B | 16K | ✅ | + +> 参数量和架构为近似值。详见各模型 YAML 文件获取准确信息。 + +--- + +**注意**：所有定价和能力数据来自第一方来源。推理平台价格可能不同。查看 `providers//models/` 获取最新数据。 From 308a7bff7c7c6007395227a443b85fd2122ee8d0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:20:46 +0800 Subject: [PATCH 006/311] docs: update README documentation table with new provider overview and model comparison links --- README.md | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index 65f638bf..ceaefb4f 100644 --- a/README.md +++ b/README.md @@ -232,12 +232,16 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation -| Document | Description | -| ------------------------------------------------------- | ----------------------------------------- | -| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | -| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | -| [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱（中文）](docs/zh/lessons-learned.md) | 经验教训 | +| Document | Description | +| ------------------------------------------------------- | -------------------------------------------------------------- | +| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | +| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | +| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | +| [模型对比（中文）](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [提供商概览（中文）](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱（中文）](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles From 37c624e0f3b44c428bb3459132b4c1c5298e6585 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:21:32 +0800 Subject: [PATCH 007/311] =?UTF-8?q?meta:=20improve=20package.json=20?= =?UTF-8?q?=E2=80=94=20description,=20keywords,=20homepage,=20repository,?= =?UTF-8?q?=20bugs=20URLs?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- package.json | 31 ++++++++++++++++++++++++++++++- 1 file changed, 30 insertions(+), 1 deletion(-) diff --git a/package.json b/package.json index 336fc645..9d1fcf2b 100644 --- a/package.json +++ b/package.json @@ -1,8 +1,37 @@ { "name": "ai-models", "version": "0.1.0", - "description": "A structured catalog of AI model providers and their models", + "description": "Structured YAML catalog of 4,682 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation.", + "keywords": [ + "ai-catalog", + "ai-models", + "ai-provider", + "anthropic", + "cerebras", + "deepseek", + "google-gemini", + "groq", + "inference-api", + "large-language-model", + "llama", + "llm", + "machine-readable", + "mistral", + "model-comparison", + "model-pricing", + "openai", + "qwen", + "yaml" + ], + "homepage": "https://github.com/i-need-token/ai-models#readme", + "bugs": { + "url": "https://github.com/i-need-token/ai-models/issues" + }, "license": "MIT", + "repository": { + "type": "git", + "url": "git+https://github.com/i-need-token/ai-models.git" + }, "type": "module", "scripts": { "sync": "npx tsx scripts/sync.ts", From 82f33f44260fcbfd2a648b25935c76091f1d7a64 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:22:02 +0800 Subject: [PATCH 008/311] =?UTF-8?q?meta:=20add=20recommended=20GitHub=20re?= =?UTF-8?q?po=20settings=20=E2=80=94=20description=20and=20topics=20for=20?= =?UTF-8?q?discoverability?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/repo-settings.md | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) create mode 100644 .github/repo-settings.md diff --git a/.github/repo-settings.md b/.github/repo-settings.md new file mode 100644 index 00000000..2cf2b489 --- /dev/null +++ b/.github/repo-settings.md @@ -0,0 +1,34 @@ +# GitHub Repository Settings + +This file documents the recommended GitHub repository settings for discoverability. + +## Repository Description + +``` +Structured YAML catalog of 4,682 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation. +``` + +## Topics + +Add these topics to the repository (Settings → General → Topics): + +- ai-models +- llm +- large-language-model +- ai-catalog +- model-pricing +- openai +- anthropic +- gemini +- deepseek +- llama +- qwen +- mistral +- groq +- cerebras +- inference-api +- model-comparison +- ai-provider +- yaml +- machine-readable +- zod From ad0d476963b003bdd953167d5119d23f491c2946 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:22:35 +0800 Subject: [PATCH 009/311] =?UTF-8?q?community:=20add=20SECURITY.md=20?= =?UTF-8?q?=E2=80=94=20vulnerability=20reporting=20policy?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- SECURITY.md | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 SECURITY.md diff --git a/SECURITY.md b/SECURITY.md new file mode 100644 index 00000000..79af03c5 --- /dev/null +++ b/SECURITY.md @@ -0,0 +1,24 @@ +# Security Policy + +## Reporting a Vulnerability + +If you discover a security vulnerability in this project, please report it by opening a [GitHub Security Advisory](https://github.com/i-need-token/ai-models/security/advisories/new). + +Please do not report security vulnerabilities through public GitHub issues. + +## Scope + +This project is a data catalog — it contains YAML model data files, TypeScript type definitions, and scrape scripts. Security vulnerabilities in this context would include: + +- Malicious code in scrape scripts that could compromise the build environment +- Supply chain vulnerabilities in dependencies +- Data integrity issues where model data is intentionally corrupted + +Out of scope: + +- Incorrect model pricing or capability data (report as a [bug](https://github.com/i-need-token/ai-models/issues/new?labels=bug&template=bug_report.md)) +- Missing providers or models (report as a [provider request](https://github.com/i-need-token/ai-models/issues/new?labels=enhancement&template=provider_request.md)) + +## Response Time + +We aim to acknowledge security reports within 48 hours and provide a fix within 7 days. From 25ad67cdd95c3523d5771fb48ee3a3e6484a07ed Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:26:14 +0800 Subject: [PATCH 010/311] =?UTF-8?q?docs:=20add=20data=20schema=20reference?= =?UTF-8?q?=20=E2=80=94=20model,=20pricing,=20snapshot,=20provider=20schem?= =?UTF-8?q?as=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/data-schema.md | 209 +++++++++++++++++++++++++++++++++++++++++ docs/zh/data-schema.md | 209 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 418 insertions(+) create mode 100644 docs/data-schema.md create mode 100644 docs/zh/data-schema.md diff --git a/docs/data-schema.md b/docs/data-schema.md new file mode 100644 index 00000000..b9f9bd6a --- /dev/null +++ b/docs/data-schema.md @@ -0,0 +1,209 @@ +**English** | [中文](./zh/data-schema.md) + +# Data Schema Reference + +Complete reference for the YAML data schema used in this catalog. All model data files conform to the TypeScript types in [`types/`](../types/) and are validated at runtime by Zod schemas in [`types/schemas.ts`](../types/schemas.ts). + +## Model Schema + +Every model is a single YAML file at `providers//models/.yaml`. + +### Required Fields + +| Field | Type | Description | Example | +| -------------- | ------- | ---------------------------------------- | ------------------------------------------ | +| `id` | string | Stable model ID (no date suffix) | `gpt-4o`, `claude-sonnet-4-5` | +| `name` | string | Display name | `GPT-4o`, `Claude Sonnet 4.5` | +| `family` | string | Model family (broad lineage) | `gpt-4o`, `claude-sonnet` | +| `pricing` | Pricing | Model pricing (see below) | — | +| `modalities` | object | Input/output modalities | `{ input: [text, image], output: [text] }` | +| `last_updated` | string | Last data update (YYYY-MM-DD or YYYY-MM) | `2024-08-06` | + +### Optional Fields + +| Field | Type | Default | Description | Example | +| ------------------- | ------- | ------- | -------------------------------- | ------------------------------------ | +| `reasoning` | boolean | `false` | Supports reasoning/thinking mode | `true` | +| `temperature` | boolean | `true` | Supports temperature parameter | `false` | +| `tool_call` | boolean | `false` | Supports tool/function calling | `true` | +| `attachment` | boolean | `false` | Supports file attachments | `true` | +| `structured_output` | boolean | `false` | Supports structured/JSON output | `true` | +| `open_weights` | boolean | `false` | Open-weight model | `true` | +| `deprecated` | boolean | `false` | Deprecated but still accessible | `true` | +| `limit` | object | — | Token limits | `{ context: 128000, output: 16384 }` | +| `limit.context` | number | — | Context window size (tokens) | `128000` | +| `limit.output` | number | — | Max output tokens | `16384` | +| `knowledge` | string | — | Training data cutoff | `2023-10` | +| `release_date` | string | — | Model release date | `2024-05-13` | +| `snapshots` | array | — | Dated model versions | See below | + +### Modality Types + +| Modality | Description | +| -------- | --------------------- | +| `text` | Text input or output | +| `image` | Image input or output | +| `video` | Video input | +| `audio` | Audio input or output | +| `pdf` | PDF document input | + +## Pricing Schema + +Pricing is a union of four types. Each model uses exactly one. + +### TokenPricing (most common) + +Per-million-token pricing. Currency defaults to USD, unit defaults to `per_mtok`. + +```yaml +pricing: + currency: USD # optional, defaults to USD + unit: per_mtok # optional, defaults to per_mtok + input: 2.5 # $/M input tokens + output: 10 # $/M output tokens + cache_write: 1.25 # optional, $/M cache write + cache_read: 0.625 # optional, $/M cache read +``` + +**Advanced: Tiered pricing by context length** + +```yaml +pricing: + input: + - up_to: 128000 # ≤ 128K context + price: 2.5 + - price: 5.0 # > 128K context (no up_to = final tier) + output: 10 +``` + +**Advanced: Per-modality pricing** + +```yaml +pricing: + input: + text: 1.25 + image: 2.5 + audio: 5.0 + output: + text: 5.0 + audio: 10.0 +``` + +### VideoPricing + +Per-second pricing, optionally tiered by resolution. + +```yaml +pricing: + currency: USD + unit: per_second + price: 0.03 # fixed price per second +``` + +```yaml +pricing: + unit: per_second + price: # per-resolution pricing + 720p: 0.02 + 1080p: 0.03 + 4k: 0.05 +``` + +### UnitPricing + +Per-image or per-request pricing. + +```yaml +pricing: + unit: per_image + price: 0.04 +``` + +```yaml +pricing: + unit: per_request + price: 0.005 +``` + +### FreePricing + +No cost. + +```yaml +pricing: + unit: free +``` + +## Snapshot Schema + +Snapshots represent dated versions of a model. They inherit all parent fields and only override what differs. + +```yaml +id: gpt-4o +name: GPT-4o +# ... parent fields ... +snapshots: + - id: gpt-4o-2024-08-06 # newest first + last_updated: "2024-08-06" + - id: gpt-4o-2024-05-13 + deprecated: true # this snapshot is deprecated + last_updated: "2024-05-13" +``` + +A snapshot can override any optional field from the parent: + +```yaml +snapshots: + - id: gemini-2.0-flash-exp + limit: + context: 1048576 # different context window + output: 8192 + pricing: + unit: free # experimental = free +``` + +## Provider Schema + +Each provider has a `provider.yaml` file at `providers//provider.yaml`. + +| Field | Type | Required | Description | Example | +| ---------------- | ------ | -------- | ------------------------------------ | ---------------------------------- | +| `id` | string | ✅ | Provider ID (matches directory name) | `openai` | +| `name` | string | ✅ | Display name | `OpenAI` | +| `url` | string | ✅ | Official website URL | `https://openai.com` | +| `api_docs` | string | ❌ | API documentation URL | `https://platform.openai.com/docs` | +| `apis` | object | ✅ | API endpoints keyed by format | See below | +| `apis.openai` | string | ❌ | OpenAI-compatible API endpoint | `https://api.openai.com/v1` | +| `apis.anthropic` | string | ❌ | Anthropic API endpoint | — | +| `apis.google` | string | ❌ | Google AI API endpoint | — | +| `currency` | string | ❌ | Default currency (USD/CNY/EUR) | `USD` | + +### API Formats + +| Format | Description | Used by | +| ----------- | -------------------------------------- | ----------------- | +| `openai` | OpenAI-compatible chat completions API | Most providers | +| `anthropic` | Anthropic Messages API | Anthropic | +| `google` | Google Generative AI API | Google, Vertex AI | + +## Currency Reference + +| Currency | Code | Used by | +| ------------ | ----- | -------------------------------------------- | +| US Dollar | `USD` | Most providers (default) | +| Chinese Yuan | `CNY` | Alibaba, 302.AI, AIHubMix, PPIO, etc. | +| Euro | `EUR` | Berget, CloudFerro, OVHcloud, Scaleway, etc. | + +## Validation + +All YAML files are validated against Zod schemas at runtime: + +```bash +# Validate all model data +npx tsx scripts/validate.ts + +# Validate a specific provider +npx tsx scripts/validate.ts openai +``` + +The validation uses `ModelSchema` from [`types/schemas.ts`](../types/schemas.ts), which mirrors the TypeScript types exactly. Any YAML file that doesn't conform to the schema will produce a validation error with the specific field path and issue. diff --git a/docs/zh/data-schema.md b/docs/zh/data-schema.md new file mode 100644 index 00000000..9fdd8cdf --- /dev/null +++ b/docs/zh/data-schema.md @@ -0,0 +1,209 @@ +[English](../data-schema.md) | **中文** + +# 数据 Schema 参考 + +本目录使用的 YAML 数据 Schema 完整参考。所有模型数据文件遵循 [`types/`](../../types/) 中的 TypeScript 类型定义，并在运行时由 [`types/schemas.ts`](../../types/schemas.ts) 中的 Zod schema 校验。 + +## 模型 Schema + +每个模型是 `providers//models/.yaml` 下的单个 YAML 文件。 + +### 必填字段 + +| 字段 | 类型 | 描述 | 示例 | +| -------------- | ------- | ---------------------------------------- | ------------------------------------------ | +| `id` | string | 稳定的模型 ID（无日期后缀） | `gpt-4o`, `claude-sonnet-4-5` | +| `name` | string | 显示名称 | `GPT-4o`, `Claude Sonnet 4.5` | +| `family` | string | 模型家族（广泛谱系） | `gpt-4o`, `claude-sonnet` | +| `pricing` | Pricing | 模型定价（见下文） | — | +| `modalities` | object | 输入/输出模态 | `{ input: [text, image], output: [text] }` | +| `last_updated` | string | 最后数据更新日期 (YYYY-MM-DD 或 YYYY-MM) | `2024-08-06` | + +### 可选字段 + +| 字段 | 类型 | 默认值 | 描述 | 示例 | +| ------------------- | ------- | ------- | ------------------------ | ------------------------------------ | +| `reasoning` | boolean | `false` | 支持推理/思考模式 | `true` | +| `temperature` | boolean | `true` | 支持 temperature 参数 | `false` | +| `tool_call` | boolean | `false` | 支持工具/函数调用 | `true` | +| `attachment` | boolean | `false` | 支持文件附件 | `true` | +| `structured_output` | boolean | `false` | 支持结构化/JSON 输出 | `true` | +| `open_weights` | boolean | `false` | 开源权重模型 | `true` | +| `deprecated` | boolean | `false` | 已弃用但仍可访问 | `true` | +| `limit` | object | — | Token 限制 | `{ context: 128000, output: 16384 }` | +| `limit.context` | number | — | 上下文窗口大小（tokens） | `128000` | +| `limit.output` | number | — | 最大输出 tokens | `16384` | +| `knowledge` | string | — | 训练数据截止日期 | `2023-10` | +| `release_date` | string | — | 模型发布日期 | `2024-05-13` | +| `snapshots` | array | — | 带日期的模型版本 | 见下文 | + +### 模态类型 + +| 模态 | 描述 | +| ------- | -------------- | +| `text` | 文本输入或输出 | +| `image` | 图像输入或输出 | +| `video` | 视频输入 | +| `audio` | 音频输入或输出 | +| `pdf` | PDF 文档输入 | + +## 定价 Schema + +定价是四种类型的联合体。每个模型只使用一种。 + +### TokenPricing（最常见） + +按百万 token 计费。货币默认 USD，单位默认 `per_mtok`。 + +```yaml +pricing: + currency: USD # 可选，默认 USD + unit: per_mtok # 可选，默认 per_mtok + input: 2.5 # $/百万 输入 token + output: 10 # $/百万 输出 token + cache_write: 1.25 # 可选，$/百万 缓存写入 + cache_read: 0.625 # 可选，$/百万 缓存读取 +``` + +**进阶：按上下文长度分层定价** + +```yaml +pricing: + input: + - up_to: 128000 # ≤ 128K 上下文 + price: 2.5 + - price: 5.0 # > 128K 上下文（无 up_to = 最终层级） + output: 10 +``` + +**进阶：按模态定价** + +```yaml +pricing: + input: + text: 1.25 + image: 2.5 + audio: 5.0 + output: + text: 5.0 + audio: 10.0 +``` + +### VideoPricing + +按秒计费，可选按分辨率分层。 + +```yaml +pricing: + currency: USD + unit: per_second + price: 0.03 # 固定每秒价格 +``` + +```yaml +pricing: + unit: per_second + price: # 按分辨率定价 + 720p: 0.02 + 1080p: 0.03 + 4k: 0.05 +``` + +### UnitPricing + +按图像或按请求计费。 + +```yaml +pricing: + unit: per_image + price: 0.04 +``` + +```yaml +pricing: + unit: per_request + price: 0.005 +``` + +### FreePricing + +免费。 + +```yaml +pricing: + unit: free +``` + +## 快照 Schema + +快照代表模型的带日期版本。它们继承父级的所有字段，只覆盖不同的部分。 + +```yaml +id: gpt-4o +name: GPT-4o +# ... 父级字段 ... +snapshots: + - id: gpt-4o-2024-08-06 # 最新的在前 + last_updated: "2024-08-06" + - id: gpt-4o-2024-05-13 + deprecated: true # 此快照已弃用 + last_updated: "2024-05-13" +``` + +快照可以覆盖父级的任何可选字段： + +```yaml +snapshots: + - id: gemini-2.0-flash-exp + limit: + context: 1048576 # 不同的上下文窗口 + output: 8192 + pricing: + unit: free # 实验版 = 免费 +``` + +## 提供商 Schema + +每个提供商在 `providers//provider.yaml` 有一个 `provider.yaml` 文件。 + +| 字段 | 类型 | 必填 | 描述 | 示例 | +| ---------------- | ------ | ---- | ------------------------- | ---------------------------------- | +| `id` | string | ✅ | 提供商 ID（与目录名匹配） | `openai` | +| `name` | string | ✅ | 显示名称 | `OpenAI` | +| `url` | string | ✅ | 官方网站 URL | `https://openai.com` | +| `api_docs` | string | ❌ | API 文档 URL | `https://platform.openai.com/docs` | +| `apis` | object | ✅ | 按格式分类的 API 端点 | 见下文 | +| `apis.openai` | string | ❌ | OpenAI 兼容 API 端点 | `https://api.openai.com/v1` | +| `apis.anthropic` | string | ❌ | Anthropic API 端点 | — | +| `apis.google` | string | ❌ | Google AI API 端点 | — | +| `currency` | string | ❌ | 默认货币 (USD/CNY/EUR) | `USD` | + +### API 格式 + +| 格式 | 描述 | 使用者 | +| ----------- | ---------------------------------- | ----------------- | +| `openai` | OpenAI 兼容的 Chat Completions API | 大多数提供商 | +| `anthropic` | Anthropic Messages API | Anthropic | +| `google` | Google Generative AI API | Google, Vertex AI | + +## 货币参考 + +| 货币 | 代码 | 使用者 | +| ------ | ----- | ----------------------------------------- | +| 美元 | `USD` | 大多数提供商（默认） | +| 人民币 | `CNY` | 阿里云、302.AI、AIHubMix、PPIO 等 | +| 欧元 | `EUR` | Berget、CloudFerro、OVHcloud、Scaleway 等 | + +## 校验 + +所有 YAML 文件在运行时由 Zod schema 校验： + +```bash +# 校验所有模型数据 +npx tsx scripts/validate.ts + +# 校验特定提供商 +npx tsx scripts/validate.ts openai +``` + +校验使用 [`types/schemas.ts`](../../types/schemas.ts) 中的 `ModelSchema`，与 TypeScript 类型完全对应。任何不符合 schema 的 YAML 文件将产生校验错误，包含具体的字段路径和问题。 From 0e27e920f926420e7d5a395bbbcc7fbff49b110e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:26:38 +0800 Subject: [PATCH 011/311] =?UTF-8?q?ci:=20add=20GitHub=20Actions=20workflow?= =?UTF-8?q?=20=E2=80=94=20validate=20YAML,=20type=20check,=20lint,=20repor?= =?UTF-8?q?t=20stats?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/workflows/validate.yml | 58 ++++++++++++++++++++++++++++++++++ 1 file changed, 58 insertions(+) create mode 100644 .github/workflows/validate.yml diff --git a/.github/workflows/validate.yml b/.github/workflows/validate.yml new file mode 100644 index 00000000..15381d97 --- /dev/null +++ b/.github/workflows/validate.yml @@ -0,0 +1,58 @@ +name: Validate + +on: + push: + branches: [main] + paths: + - "providers/**/*.yaml" + - "providers/**/*.yml" + - "types/**" + - "scripts/**" + pull_request: + branches: [main] + paths: + - "providers/**/*.yaml" + - "providers/**/*.yml" + - "types/**" + - "scripts/**" + +jobs: + validate: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + + - name: Install dependencies + run: npm ci + + - name: Validate YAML data + run: npx tsx scripts/validate.ts + + - name: Type check + run: npx tsc --noEmit + + - name: Lint + run: npm run lint + + stats: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - name: Count providers + id: count + run: | + providers=$(ls providers/ | wc -l | tr -d ' ') + models=$(find providers -name "*.yaml" -o -name "*.yml" | wc -l | tr -d ' ') + echo "providers=$providers" >> "$GITHUB_OUTPUT" + echo "models=$models" >> "$GITHUB_OUTPUT" + + - name: Report stats + run: | + echo "📊 Catalog Stats" + echo "Providers: ${{ steps.count.outputs.providers }}" + echo "Model files: ${{ steps.count.outputs.models }}" From f5ae3b0001a878764ec427fe1fa62c1567c7fee9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:27:09 +0800 Subject: [PATCH 012/311] =?UTF-8?q?community:=20add=20CODE=5FOF=5FCONDUCT.?= =?UTF-8?q?md=20=E2=80=94=20Contributor=20Covenant=20v2.1?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- CODE_OF_CONDUCT.md | 77 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 77 insertions(+) create mode 100644 CODE_OF_CONDUCT.md diff --git a/CODE_OF_CONDUCT.md b/CODE_OF_CONDUCT.md new file mode 100644 index 00000000..2be142d8 --- /dev/null +++ b/CODE_OF_CONDUCT.md @@ -0,0 +1,77 @@ +# Contributor Covenant Code of Conduct + +## Our Pledge + +We as members, contributors, and leaders pledge to make participation in our community a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education or socio-economic status, nationality, personal appearance, race, caste, color, religion, or sexual identity and orientation. + +We pledge to act and interact in ways that contribute to an open, welcoming, diverse, inclusive, and healthy community. + +## Our Standards + +Examples of behavior that contributes to a positive environment for our community include: + +- Demonstrating empathy and kindness toward other people +- Being respectful of differing opinions, viewpoints, and experiences +- Giving and gracefully accepting constructive feedback +- Accepting responsibility and apologizing to those affected by our mistakes, and learning from the experience +- Focusing on what is best not just for us as individuals, but for the overall community + +Examples of unacceptable behavior include: + +- The use of sexualized language or imagery, and sexual attention or advances of any kind +- Trolling, insulting or derogatory comments, and personal or political attacks +- Public or private harassment +- Publishing others' private information, such as a physical or email address, without their explicit permission +- Other conduct which could reasonably be considered inappropriate in a professional setting + +## Enforcement Responsibilities + +Community leaders are responsible for clarifying and enforcing our standards of acceptable behavior and will take appropriate and fair corrective action in response to any behavior that they deem inappropriate, threatening, offensive, or harmful. + +Community leaders have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, and will communicate reasons for moderation decisions when appropriate. + +## Scope + +This Code of Conduct applies within all community spaces, and also applies when an individual is officially representing the community in public spaces. Examples of representing our community include using an official e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. + +## Enforcement + +Instances of abusive, harassing, or otherwise unacceptable behavior may be reported to the community leaders responsible for enforcement through [GitHub Issues](https://github.com/i-need-token/ai-models/issues) or directly via email. All complaints will be reviewed and investigated promptly and fairly. + +All community leaders are obligated to respect the privacy and security of the reporter of any incident. + +## Enforcement Guidelines + +Community leaders will follow these Community Impact Guidelines in determining the consequences for any action they deem in violation of this Code of Conduct: + +### 1. Correction + +**Community Impact**: Use of inappropriate language or other behavior deemed unprofessional or unwelcome in the community. + +**Consequence**: A private, written warning from community leaders, providing clarity around the nature of the violation and an explanation of why the behavior was inappropriate. A public apology may be requested. + +### 2. Warning + +**Community Impact**: A violation through a single incident or series of actions. + +**Consequence**: A warning with consequences for continued behavior. No interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, for a specified period of time. This includes avoiding interactions in community spaces as well as external channels like social media. Violating these terms may lead to a temporary or permanent ban. + +### 3. Temporary Ban + +**Community Impact**: A serious violation of community standards, including sustained inappropriate behavior. + +**Consequence**: A temporary ban from any sort of interaction or public communication with the community for a specified period of time. No public or private interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, is allowed during this period. Violating these terms may lead to a permanent ban. + +### 4. Permanent Ban + +**Community Impact**: Demonstrating a pattern of violation of community standards, including sustained inappropriate behavior, harassment of an individual, or aggression toward or disparagement of classes of individuals. + +**Consequence**: A permanent ban from any sort of public interaction within the community. + +## Attribution + +This Code of Conduct is adapted from the [Contributor Covenant](https://www.contributor-covenant.org/), version 2.1, available at [https://www.contributor-covenant.org/version/2/1/code_of_conduct.html](https://www.contributor-covenant.org/version/2/1/code_of_conduct.html). + +Community Impact Guidelines were inspired by [Mozilla's code of conduct enforcement ladder](https://github.com/divmain/pull-mozilla-coc). + +For answers to common questions about this code of conduct, see the FAQ at [https://www.contributor-covenant.org/faq](https://www.contributor-covenant.org/faq). Translations are available at [https://www.contributor-covenant.org/translations](https://www.contributor-covenant.org/translations). From dc779210c296918880f17731236affcfc3c2ea74 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:28:00 +0800 Subject: [PATCH 013/311] docs: add CI badge, quick-nav links, data schema link to README --- README.md | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/README.md b/README.md index ceaefb4f..545513c3 100644 --- a/README.md +++ b/README.md @@ -9,6 +9,7 @@ [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) [![Models](https://img.shields.io/badge/Models-4%2C682-green.svg)](providers/) [![Providers](https://img.shields.io/badge/Providers-95-orange.svg)](providers/) +[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) @@ -16,6 +17,8 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. +**[Compare models →](docs/model-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** + ## Why This Catalog? | | | @@ -236,10 +239,12 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | ------------------------------------------------------- | -------------------------------------------------------------- | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | +| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | | [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | | [模型对比（中文）](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [提供商概览（中文）](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [数据 Schema 参考（中文）](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | | [设计原则与陷阱（中文）](docs/zh/lessons-learned.md) | 经验教训 | From 32d6cb93c2058c6bc4f2bbafd4978fa7bbbb6c23 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:28:20 +0800 Subject: [PATCH 014/311] docs: update AGENTS.md with new documentation links --- AGENTS.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index 11c1c000..3af0109a 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -10,6 +10,9 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`types/schemas.ts`](types/schemas.ts) — Zod runtime validation schemas - [`docs/data-acquisition.md`](docs/data-acquisition.md) — How we acquire and update model data ([中文](docs/zh/data-acquisition.md)) - [`docs/lessons-learned.md`](docs/lessons-learned.md) — Design principles and pitfalls ([中文](docs/zh/lessons-learned.md)) +- [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) +- [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) +- [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) ## Key Design Decisions From 704064fd71e543e89d650cfaa1915d3dfb5808d3 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:28:55 +0800 Subject: [PATCH 015/311] chore: add *.tsbuildinfo to .gitignore --- .gitignore | 1 + 1 file changed, 1 insertion(+) diff --git a/.gitignore b/.gitignore index 0b655613..eafbc0be 100644 --- a/.gitignore +++ b/.gitignore @@ -3,3 +3,4 @@ dist/ .claude/ .pi/ *.local +*.tsbuildinfo From 3b8a33081bf3ca350f1fa48d5837552c749194fb Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:29:09 +0800 Subject: [PATCH 016/311] =?UTF-8?q?community:=20add=20FUNDING.yml=20?= =?UTF-8?q?=E2=80=94=20GitHub=20Sponsors?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/FUNDING.yml | 3 +++ 1 file changed, 3 insertions(+) create mode 100644 .github/FUNDING.yml diff --git a/.github/FUNDING.yml b/.github/FUNDING.yml new file mode 100644 index 00000000..ff862f3b --- /dev/null +++ b/.github/FUNDING.yml @@ -0,0 +1,3 @@ +# These supported funding platforms will be linked from the repository sidebar + +github: i-need-token From f4ad3326d2ea5c05c2ab2914f30efd302671355f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:40:25 +0800 Subject: [PATCH 017/311] =?UTF-8?q?docs:=20fix=20README=20stats=20?= =?UTF-8?q?=E2=80=94=20unique=20IDs=202804=E2=86=922807,=20tool=5Fcall=202?= =?UTF-8?q?347=E2=86=922350,=20vision=201488=E2=86=921487,=20image=5Fout?= =?UTF-8?q?=2084=E2=86=9228,=20audio=5Fin=20148=E2=86=92118,=20add=20audio?= =?UTF-8?q?=5Fout=2034,=20video=5Fin=20171=E2=86=92167?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index 545513c3..362ca24a 100644 --- a/README.md +++ b/README.md @@ -35,16 +35,17 @@ Machine-readable YAML catalog of every major AI model provider and their models | --------------------------- | ----: | | Providers | 95 | | Model files | 4,682 | -| Unique model IDs | 2,804 | +| Unique model IDs | 2,807 | | Model families | 441 | | Reasoning models | 1,306 | -| Tool-calling models | 2,347 | +| Tool-calling models | 2,350 | | Open-weight models | 527 | | Free models | 81 | -| Vision (image input) models | 1,488 | -| Image output models | 84 | -| Audio input models | 148 | -| Video input models | 171 | +| Vision (image input) models | 1,487 | +| Image output models | 28 | +| Audio input models | 118 | +| Audio output models | 34 | +| Video input models | 167 | ## Data at a Glance From ad10f56988162dd723823887145beb34ebb895f9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:42:26 +0800 Subject: [PATCH 018/311] =?UTF-8?q?docs:=20fix=20model=20comparison=20?= =?UTF-8?q?=E2=80=94=20replace=20fabricated=20models/pricing=20with=20veri?= =?UTF-8?q?fied=20first-party=20data=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/model-comparison.md | 91 ++++++++++++++++++------------------- docs/zh/model-comparison.md | 91 ++++++++++++++++++------------------- 2 files changed, 88 insertions(+), 94 deletions(-) diff --git a/docs/model-comparison.md b/docs/model-comparison.md index 9f403035..91caa6fa 100644 --- a/docs/model-comparison.md +++ b/docs/model-comparison.md @@ -6,20 +6,18 @@ Quick-reference comparisons for popular AI model categories. All data sourced fr ## Top-Tier Flagship Models -The most capable models from each major provider as of May 2025. +The most capable models from each major provider. Pricing shown for direct provider API. | Model | Provider | Context | Input $/Mtok | Output $/Mtok | Reasoning | Tool Call | Vision | | ---------------- | --------- | ------: | -----------: | ------------: | :-------: | :-------: | :----: | -| GPT-5.5 | OpenAI | 512K | 10.00 | 30.00 | ✅ | ✅ | ✅ | -| Claude Opus 4.7 | Anthropic | 200K | 15.00 | 75.00 | ✅ | ✅ | ✅ | -| Gemini 3.1 Pro | Google | 1M | 1.25 | 10.00 | ✅ | ✅ | ✅ | -| DeepSeek-V4-Pro | DeepSeek | 128K | 2.50 | 10.00 | ✅ | ✅ | ✅ | -| Grok 3 | xAI | 131K | 3.00 | 15.00 | ✅ | ✅ | ✅ | -| Llama 4 Maverick | Meta | 1M | — | — | ✅ | ✅ | ✅ | -| Qwen3-235B | Alibaba | 128K | — | — | ✅ | ✅ | ✅ | -| Mistral Large | Mistral | 128K | 2.00 | 6.00 | ✅ | ✅ | ✅ | - -> Pricing shown for direct provider API. Inference platforms may offer different rates. +| o3 | OpenAI | 200K | 10.00 | 40.00 | ✅ | ✅ | ✅ | +| Claude Opus 4.7 | Anthropic | 1M | 5.00 | 25.00 | ✅ | ✅ | ✅ | +| Gemini 2.5 Pro | Google | 1M | 1.25 | 10.00 | ✅ | ✅ | ✅ | +| DeepSeek-V4-Pro | DeepSeek | 1M | 0.435 | 0.87 | ✅ | ✅ | ❌ | +| Grok 4 | xAI | 131K | 3.00 | 15.00 | ✅ | ✅ | ✅ | +| Llama 4 Maverick | Meta | 1M | 0.24 | 0.97 | ❌ | ✅ | ✅ | +| Qwen3-235B | Alibaba | — | 2.00 | 8.00 | ✅ | ✅ | ❌ | +| Mistral Large | Mistral | 128K | 2.00 | 6.00 | ❌ | ✅ | ✅ | ## Cost-Effective Models @@ -27,14 +25,15 @@ Best value models for high-volume workloads. | Model | Provider | Context | Input $/Mtok | Output $/Mtok | Reasoning | Tool Call | | ----------------- | --------- | ------: | -----------: | ------------: | :-------: | :-------: | -| GPT-5.4 Nano | OpenAI | 128K | 0.03 | 0.12 | ❌ | ✅ | -| Claude Haiku 4.5 | Anthropic | 200K | 0.80 | 4.00 | ✅ | ✅ | -| Gemini 3.5 Flash | Google | 1M | 0.15 | 0.60 | ✅ | ✅ | -| DeepSeek-V4-Flash | DeepSeek | 128K | 0.10 | 0.40 | ✅ | ✅ | -| Llama 4 Scout | Meta | 10M | — | — | ✅ | ✅ | -| Qwen3-30B | Alibaba | 128K | — | — | ✅ | ✅ | +| GPT-4.1 Nano | OpenAI | 1M | 0.10 | 0.40 | ❌ | ✅ | +| o4-mini | OpenAI | 200K | 1.10 | 4.40 | ✅ | ✅ | +| Claude Haiku 4.5 | Anthropic | 200K | 1.00 | 5.00 | ✅ | ✅ | +| Gemini 2.5 Flash | Google | 1M | 0.15 | 3.50 | ✅ | ✅ | +| DeepSeek-V4-Flash | DeepSeek | 1M | 0.14 | 0.28 | ✅ | ✅ | +| Llama 4 Scout | Meta | 10M | 0.17 | 0.66 | ❌ | ✅ | +| Qwen3-30B | Alibaba | — | 0.75 | 3.00 | ✅ | ✅ | | Mistral Small | Mistral | 128K | 0.20 | 0.60 | ❌ | ✅ | -| Grok 3 Mini | xAI | 131K | 0.30 | 0.50 | ✅ | ✅ | +| Grok 3 Mini | xAI | 131K | 0.25 | 1.27 | ✅ | ✅ | ## Largest Context Windows @@ -42,14 +41,14 @@ Models with the biggest context windows for long-document processing. | Model | Provider | Context (tokens) | Input $/Mtok | Output $/Mtok | | ----------------- | --------- | ---------------: | -----------: | ------------: | -| Llama 4 Scout | Meta | 10,000,000 | — | — | -| Gemini 3.1 Pro | Google | 1,048,576 | 1.25 | 10.00 | -| Gemini 3.5 Flash | Google | 1,048,576 | 0.15 | 0.60 | -| Llama 4 Maverick | Meta | 1,000,000 | — | — | -| GPT-5.5 | OpenAI | 512,000 | 10.00 | 30.00 | -| Qwen3-Coder-480B | Alibaba | 1,048,576 | — | — | -| Claude Opus 4.7 | Anthropic | 200,000 | 15.00 | 75.00 | -| Claude Sonnet 4.6 | Anthropic | 200,000 | 3.00 | 15.00 | +| Llama 4 Scout | Meta | 10,000,000 | 0.17 | 0.66 | +| Claude Opus 4.7 | Anthropic | 1,000,000 | 5.00 | 25.00 | +| Claude Sonnet 4.6 | Anthropic | 1,000,000 | 3.00 | 15.00 | +| GPT-4.1 | OpenAI | 1,048,576 | 2.00 | 8.00 | +| Gemini 2.5 Pro | Google | 1,048,576 | 1.25 | 10.00 | +| Gemini 2.5 Flash | Google | 1,048,576 | 0.15 | 3.50 | +| Llama 4 Maverick | Meta | 1,000,000 | 0.24 | 0.97 | +| DeepSeek-V4-Pro | DeepSeek | 1,000,000 | 0.435 | 0.87 | ## Free Models @@ -57,11 +56,11 @@ Models available at no cost (as of data collection date). | Model | Provider | Context | Reasoning | Tool Call | | ----------------------------- | -------- | ------: | :-------: | :-------: | -| DeepSeek-V4-Flash (free tier) | DeepSeek | 128K | ✅ | ✅ | -| Gemini 3.5 Flash (free tier) | Google | 1M | ✅ | ✅ | -| Llama 4 Scout (self-hosted) | Meta | 10M | ✅ | ✅ | -| Qwen3-8B (self-hosted) | Alibaba | 128K | ✅ | ✅ | -| Mistral-Small (self-hosted) | Mistral | 128K | ❌ | ✅ | +| DeepSeek-V4-Flash (free tier) | DeepSeek | 1M | ✅ | ✅ | +| Gemini 2.5 Flash (free tier) | Google | 1M | ✅ | ✅ | +| Llama 4 Scout (self-hosted) | Meta | 10M | ❌ | ✅ | +| Qwen3-30B (self-hosted) | Alibaba | — | ✅ | ✅ | +| Mistral Small (self-hosted) | Mistral | 128K | ❌ | ✅ | > Free tiers typically have rate limits. Self-hosted models require your own infrastructure. @@ -71,29 +70,27 @@ Models that accept image inputs. | Model | Provider | Image Input | Image Output | Video Input | | ---------------- | --------- | :---------: | :----------: | :---------: | -| GPT-5.5 | OpenAI | ✅ | ✅ | ✅ | +| o3 | OpenAI | ✅ | ❌ | ❌ | | Claude Opus 4.7 | Anthropic | ✅ | ❌ | ❌ | -| Gemini 3.1 Pro | Google | ✅ | ✅ | ✅ | -| DeepSeek-V4-Pro | DeepSeek | ✅ | ❌ | ❌ | -| Qwen3-VL | Alibaba | ✅ | ❌ | ❌ | +| Gemini 2.5 Pro | Google | ✅ | ❌ | ❌ | +| GPT-4.1 | OpenAI | ✅ | ❌ | ❌ | | Llama 4 Maverick | Meta | ✅ | ❌ | ❌ | +| Grok 3 | xAI | ✅ | ❌ | ❌ | ## Open-Weight Models Models with publicly available weights for self-hosting. -| Model | Provider | Parameters | Context | Reasoning | -| ----------------- | --------- | :--------- | ------: | :-------: | -| Llama 4 Maverick | Meta | 400B MoE | 1M | ✅ | -| Llama 4 Scout | Meta | 109B MoE | 10M | ✅ | -| Qwen3-235B | Alibaba | 235B MoE | 128K | ✅ | -| Qwen3-30B | Alibaba | 30B MoE | 128K | ✅ | -| DeepSeek-R1 | DeepSeek | 671B MoE | 128K | ✅ | -| DeepSeek-V3.2 | DeepSeek | 685B MoE | 128K | ✅ | -| Mistral Small 3.2 | Mistral | 24B | 128K | ❌ | -| Phi-4 | Microsoft | 14B | 16K | ✅ | - -> Parameter counts and architectures are approximate. See individual model YAML files for exact details. +| Model | Provider | Context | Input $/Mtok | Output $/Mtok | Reasoning | +| ----------------- | --------- | ------: | -----------: | ------------: | :-------: | +| Llama 4 Maverick | Meta | 1M | 0.24 | 0.97 | ❌ | +| Llama 4 Scout | Meta | 10M | 0.17 | 0.66 | ❌ | +| Qwen3-235B | Alibaba | — | 2.00 | 8.00 | ✅ | +| Qwen3-30B | Alibaba | — | 0.75 | 3.00 | ✅ | +| Mistral Small 3.2 | Mistral | 128K | 0.20 | 0.60 | ❌ | +| Phi-4 | Microsoft | 16K | 0.125 | 0.50 | ❌ | + +> Pricing shown for hosted inference. Self-hosted models have no per-token cost but require infrastructure. --- diff --git a/docs/zh/model-comparison.md b/docs/zh/model-comparison.md index 0e2233e6..3fc8a937 100644 --- a/docs/zh/model-comparison.md +++ b/docs/zh/model-comparison.md @@ -6,20 +6,18 @@ ## 顶级旗舰模型 -截至 2025 年 5 月，各主要提供商最强大的模型。 +各主要提供商最强大的模型。定价为直接提供商 API 价格。 | 模型 | 提供商 | 上下文 | 输入 $/百万token | 输出 $/百万token | 推理 | 工具调用 | 视觉 | | ---------------- | --------- | -----: | ---------------: | ---------------: | :--: | :------: | :--: | -| GPT-5.5 | OpenAI | 512K | 10.00 | 30.00 | ✅ | ✅ | ✅ | -| Claude Opus 4.7 | Anthropic | 200K | 15.00 | 75.00 | ✅ | ✅ | ✅ | -| Gemini 3.1 Pro | Google | 1M | 1.25 | 10.00 | ✅ | ✅ | ✅ | -| DeepSeek-V4-Pro | DeepSeek | 128K | 2.50 | 10.00 | ✅ | ✅ | ✅ | -| Grok 3 | xAI | 131K | 3.00 | 15.00 | ✅ | ✅ | ✅ | -| Llama 4 Maverick | Meta | 1M | — | — | ✅ | ✅ | ✅ | -| Qwen3-235B | 阿里云 | 128K | — | — | ✅ | ✅ | ✅ | -| Mistral Large | Mistral | 128K | 2.00 | 6.00 | ✅ | ✅ | ✅ | - -> 定价为直接提供商 API 价格。推理平台可能提供不同费率。 +| o3 | OpenAI | 200K | 10.00 | 40.00 | ✅ | ✅ | ✅ | +| Claude Opus 4.7 | Anthropic | 1M | 5.00 | 25.00 | ✅ | ✅ | ✅ | +| Gemini 2.5 Pro | Google | 1M | 1.25 | 10.00 | ✅ | ✅ | ✅ | +| DeepSeek-V4-Pro | DeepSeek | 1M | 0.435 | 0.87 | ✅ | ✅ | ❌ | +| Grok 4 | xAI | 131K | 3.00 | 15.00 | ✅ | ✅ | ✅ | +| Llama 4 Maverick | Meta | 1M | 0.24 | 0.97 | ❌ | ✅ | ✅ | +| Qwen3-235B | 阿里云 | — | 2.00 | 8.00 | ✅ | ✅ | ❌ | +| Mistral Large | Mistral | 128K | 2.00 | 6.00 | ❌ | ✅ | ✅ | ## 高性价比模型 @@ -27,14 +25,15 @@ | 模型 | 提供商 | 上下文 | 输入 $/百万token | 输出 $/百万token | 推理 | 工具调用 | | ----------------- | --------- | -----: | ---------------: | ---------------: | :--: | :------: | -| GPT-5.4 Nano | OpenAI | 128K | 0.03 | 0.12 | ❌ | ✅ | -| Claude Haiku 4.5 | Anthropic | 200K | 0.80 | 4.00 | ✅ | ✅ | -| Gemini 3.5 Flash | Google | 1M | 0.15 | 0.60 | ✅ | ✅ | -| DeepSeek-V4-Flash | DeepSeek | 128K | 0.10 | 0.40 | ✅ | ✅ | -| Llama 4 Scout | Meta | 10M | — | — | ✅ | ✅ | -| Qwen3-30B | 阿里云 | 128K | — | — | ✅ | ✅ | +| GPT-4.1 Nano | OpenAI | 1M | 0.10 | 0.40 | ❌ | ✅ | +| o4-mini | OpenAI | 200K | 1.10 | 4.40 | ✅ | ✅ | +| Claude Haiku 4.5 | Anthropic | 200K | 1.00 | 5.00 | ✅ | ✅ | +| Gemini 2.5 Flash | Google | 1M | 0.15 | 3.50 | ✅ | ✅ | +| DeepSeek-V4-Flash | DeepSeek | 1M | 0.14 | 0.28 | ✅ | ✅ | +| Llama 4 Scout | Meta | 10M | 0.17 | 0.66 | ❌ | ✅ | +| Qwen3-30B | 阿里云 | — | 0.75 | 3.00 | ✅ | ✅ | | Mistral Small | Mistral | 128K | 0.20 | 0.60 | ❌ | ✅ | -| Grok 3 Mini | xAI | 131K | 0.30 | 0.50 | ✅ | ✅ | +| Grok 3 Mini | xAI | 131K | 0.25 | 1.27 | ✅ | ✅ | ## 最大上下文窗口 @@ -42,14 +41,14 @@ | 模型 | 提供商 | 上下文 (tokens) | 输入 $/百万token | 输出 $/百万token | | ----------------- | --------- | --------------: | ---------------: | ---------------: | -| Llama 4 Scout | Meta | 10,000,000 | — | — | -| Gemini 3.1 Pro | Google | 1,048,576 | 1.25 | 10.00 | -| Gemini 3.5 Flash | Google | 1,048,576 | 0.15 | 0.60 | -| Llama 4 Maverick | Meta | 1,000,000 | — | — | -| GPT-5.5 | OpenAI | 512,000 | 10.00 | 30.00 | -| Qwen3-Coder-480B | 阿里云 | 1,048,576 | — | — | -| Claude Opus 4.7 | Anthropic | 200,000 | 15.00 | 75.00 | -| Claude Sonnet 4.6 | Anthropic | 200,000 | 3.00 | 15.00 | +| Llama 4 Scout | Meta | 10,000,000 | 0.17 | 0.66 | +| Claude Opus 4.7 | Anthropic | 1,000,000 | 5.00 | 25.00 | +| Claude Sonnet 4.6 | Anthropic | 1,000,000 | 3.00 | 15.00 | +| GPT-4.1 | OpenAI | 1,048,576 | 2.00 | 8.00 | +| Gemini 2.5 Pro | Google | 1,048,576 | 1.25 | 10.00 | +| Gemini 2.5 Flash | Google | 1,048,576 | 0.15 | 3.50 | +| Llama 4 Maverick | Meta | 1,000,000 | 0.24 | 0.97 | +| DeepSeek-V4-Pro | DeepSeek | 1,000,000 | 0.435 | 0.87 | ## 免费模型 @@ -57,11 +56,11 @@ | 模型 | 提供商 | 上下文 | 推理 | 工具调用 | | -------------------------- | -------- | -----: | :--: | :------: | -| DeepSeek-V4-Flash (免费层) | DeepSeek | 128K | ✅ | ✅ | -| Gemini 3.5 Flash (免费层) | Google | 1M | ✅ | ✅ | -| Llama 4 Scout (自托管) | Meta | 10M | ✅ | ✅ | -| Qwen3-8B (自托管) | 阿里云 | 128K | ✅ | ✅ | -| Mistral-Small (自托管) | Mistral | 128K | ❌ | ✅ | +| DeepSeek-V4-Flash (免费层) | DeepSeek | 1M | ✅ | ✅ | +| Gemini 2.5 Flash (免费层) | Google | 1M | ✅ | ✅ | +| Llama 4 Scout (自托管) | Meta | 10M | ❌ | ✅ | +| Qwen3-30B (自托管) | 阿里云 | — | ✅ | ✅ | +| Mistral Small (自托管) | Mistral | 128K | ❌ | ✅ | > 免费层通常有速率限制。自托管模型需要自己的基础设施。 @@ -71,29 +70,27 @@ | 模型 | 提供商 | 图像输入 | 图像输出 | 视频输入 | | ---------------- | --------- | :------: | :------: | :------: | -| GPT-5.5 | OpenAI | ✅ | ✅ | ✅ | +| o3 | OpenAI | ✅ | ❌ | ❌ | | Claude Opus 4.7 | Anthropic | ✅ | ❌ | ❌ | -| Gemini 3.1 Pro | Google | ✅ | ✅ | ✅ | -| DeepSeek-V4-Pro | DeepSeek | ✅ | ❌ | ❌ | -| Qwen3-VL | 阿里云 | ✅ | ❌ | ❌ | +| Gemini 2.5 Pro | Google | ✅ | ❌ | ❌ | +| GPT-4.1 | OpenAI | ✅ | ❌ | ❌ | | Llama 4 Maverick | Meta | ✅ | ❌ | ❌ | +| Grok 3 | xAI | ✅ | ❌ | ❌ | ## 开源权重模型 权重公开可用的自托管模型。 -| 模型 | 提供商 | 参数量 | 上下文 | 推理 | -| ----------------- | --------- | :------: | -----: | :--: | -| Llama 4 Maverick | Meta | 400B MoE | 1M | ✅ | -| Llama 4 Scout | Meta | 109B MoE | 10M | ✅ | -| Qwen3-235B | 阿里云 | 235B MoE | 128K | ✅ | -| Qwen3-30B | 阿里云 | 30B MoE | 128K | ✅ | -| DeepSeek-R1 | DeepSeek | 671B MoE | 128K | ✅ | -| DeepSeek-V3.2 | DeepSeek | 685B MoE | 128K | ✅ | -| Mistral Small 3.2 | Mistral | 24B | 128K | ❌ | -| Phi-4 | Microsoft | 14B | 16K | ✅ | - -> 参数量和架构为近似值。详见各模型 YAML 文件获取准确信息。 +| 模型 | 提供商 | 上下文 | 输入 $/百万token | 输出 $/百万token | 推理 | +| ----------------- | --------- | -----: | ---------------: | ---------------: | :--: | +| Llama 4 Maverick | Meta | 1M | 0.24 | 0.97 | ❌ | +| Llama 4 Scout | Meta | 10M | 0.17 | 0.66 | ❌ | +| Qwen3-235B | 阿里云 | — | 2.00 | 8.00 | ✅ | +| Qwen3-30B | 阿里云 | — | 0.75 | 3.00 | ✅ | +| Mistral Small 3.2 | Mistral | 128K | 0.20 | 0.60 | ❌ | +| Phi-4 | Microsoft | 16K | 0.125 | 0.50 | ❌ | + +> 定价为托管推理价格。自托管模型无按 token 计费，但需要基础设施。 --- From 9543acde1b8e847278ea367c4836bcc749e30345 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:44:33 +0800 Subject: [PATCH 019/311] =?UTF-8?q?docs:=20add=20pricing=20comparison=20?= =?UTF-8?q?=E2=80=94=20direct=20provider=20+=20cross-platform=20+=20cheape?= =?UTF-8?q?st=20models=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/pricing-comparison.md | 121 ++++++++++++++++++++++++++++++++++ docs/zh/pricing-comparison.md | 121 ++++++++++++++++++++++++++++++++++ 2 files changed, 242 insertions(+) create mode 100644 docs/pricing-comparison.md create mode 100644 docs/zh/pricing-comparison.md diff --git a/docs/pricing-comparison.md b/docs/pricing-comparison.md new file mode 100644 index 00000000..91217e00 --- /dev/null +++ b/docs/pricing-comparison.md @@ -0,0 +1,121 @@ +**English** | [中文](./zh/pricing-comparison.md) + +# Pricing Comparison + +Side-by-side pricing comparison for AI model inference across providers and platforms. All prices in USD per million tokens, sourced from first-party APIs. + +## Direct Provider Pricing + +Pricing from the model producer's own API. + +### OpenAI + +| Model | Input $/Mtok | Output $/Mtok | Cache Read $/Mtok | Context | +| ------------ | -----------: | ------------: | ----------------: | ------: | +| GPT-4.1 Nano | 0.10 | 0.40 | 0.025 | 1M | +| GPT-4o Mini | 0.15 | 0.60 | 0.075 | 128K | +| GPT-4.1 Mini | 0.40 | 1.60 | 0.10 | 1M | +| GPT-4.1 | 2.00 | 8.00 | 0.50 | 1M | +| GPT-4o | 2.50 | 10.00 | 1.25 | 128K | +| o4-mini | 1.10 | 4.40 | 0.275 | 200K | +| o3 | 10.00 | 40.00 | 2.50 | 200K | + +### Anthropic + +| Model | Input $/Mtok | Output $/Mtok | Context | +| ----------------- | -----------: | ------------: | ------: | +| Claude Haiku 4.5 | 1.00 | 5.00 | 200K | +| Claude Sonnet 4.0 | 3.00 | 15.00 | 1M | +| Claude Sonnet 4.5 | 3.00 | 15.00 | 1M | +| Claude Sonnet 4.6 | 3.00 | 15.00 | 1M | +| Claude Opus 4.5 | 5.00 | 25.00 | 200K | +| Claude Opus 4.7 | 5.00 | 25.00 | 1M | + +### Google + +| Model | Input $/Mtok | Output $/Mtok | Cache Read $/Mtok | Context | +| --------------------- | -----------: | ------------: | ----------------: | ------: | +| Gemini 1.5 Flash 8B | 0.075 | 0.30 | — | 1M | +| Gemini 2.0 Flash Lite | 0.075 | 0.30 | — | 1M | +| Gemini 2.0 Flash | 0.10 | 0.40 | — | 1M | +| Gemini 2.5 Flash Lite | 0.10 | 0.40 | — | 1M | +| Gemini 2.5 Flash | 0.15 | 3.50 | 0.0375 | 1M | +| Gemini 2.5 Pro | 1.25 | 10.00 | 0.315 | 1M | + +### DeepSeek + +| Model | Input $/Mtok | Output $/Mtok | Cache Read $/Mtok | Context | +| ----------------- | -----------: | ------------: | ----------------: | ------: | +| DeepSeek-V4-Flash | 0.14 | 0.28 | 0.0028 | 1M | +| DeepSeek-V4-Pro | 0.435 | 0.87 | 0.003625 | 1M | + +### xAI + +| Model | Input $/Mtok | Output $/Mtok | Context | +| ----------- | -----------: | ------------: | ------: | +| Grok 4 Fast | 0.20 | 0.50 | 131K | +| Grok 4.1 | 0.20 | 0.50 | 131K | +| Grok 3 Mini | 0.25 | 1.27 | 131K | +| Grok 4.2 | 2.00 | 6.00 | 131K | +| Grok 3 | 3.00 | 15.00 | 131K | +| Grok 4 | 3.00 | 15.00 | 131K | + +### Meta (via hosted inference) + +| Model | Input $/Mtok | Output $/Mtok | Context | +| ---------------- | -----------: | ------------: | ------: | +| Llama 3.2 1B | 0.10 | 0.10 | 128K | +| Llama 4 Scout | 0.17 | 0.66 | 10M | +| Llama 4 Maverick | 0.24 | 0.97 | 1M | + +### Mistral + +| Model | Input $/Mtok | Output $/Mtok | Context | +| ------------- | -----------: | ------------: | ------: | +| Ministral 3B | 0.04 | 0.04 | 128K | +| Ministral 8B | 0.10 | 0.10 | 128K | +| Mistral Small | 0.20 | 0.60 | 128K | +| Mistral Large | 2.00 | 6.00 | 128K | + +## Cross-Platform Price Comparison + +Same model on different inference platforms — prices can vary significantly. + +### Llama 4 Scout (10M context) + +| Platform | Input $/Mtok | Output $/Mtok | +| ------------- | -----------: | ------------: | +| AIHubMix | 0.061 | 0.183 | +| Auriko | 0.08 | 0.30 | +| DeepInfra | 0.08 | 0.30 | +| Kluster AI | 0.08 | 0.45 | +| Meta (direct) | 0.17 | 0.66 | + +### Llama 4 Maverick (1M context) + +| Platform | Input $/Mtok | Output $/Mtok | +| --------------- | -----------: | ------------: | +| AIHubMix | 0.10 | 0.10 | +| 接口 AI | 0.10 | 0.50 | +| AIHubMix (Groq) | 0.11 | 0.33 | +| Cortecs | 0.124 | 0.603 | +| Auriko | 0.15 | 0.60 | +| Meta (direct) | 0.24 | 0.97 | + +## Cheapest Models Overall + +The absolute cheapest per-token models across all providers. + +| Model | Provider | Input $/Mtok | Output $/Mtok | Context | +| ------------------- | -------- | -----------: | ------------: | ------: | +| Ministral 3B | Mistral | 0.04 | 0.04 | 128K | +| Voxtral Mini | Mistral | 0.04 | 0.04 | 128K | +| Ministral 8B | Mistral | 0.10 | 0.10 | 128K | +| Llama 3.2 1B | Meta | 0.10 | 0.10 | 128K | +| GPT-4.1 Nano | OpenAI | 0.10 | 0.40 | 1M | +| Gemini 1.5 Flash 8B | Google | 0.075 | 0.30 | 1M | +| DeepSeek-V4-Flash | DeepSeek | 0.14 | 0.28 | 1M | + +--- + +**Note**: All pricing from first-party sources as of data collection date. Inference platform prices may differ. Check `providers//models/` for current data. CNY and EUR pricing available in provider YAML files. diff --git a/docs/zh/pricing-comparison.md b/docs/zh/pricing-comparison.md new file mode 100644 index 00000000..8a690853 --- /dev/null +++ b/docs/zh/pricing-comparison.md @@ -0,0 +1,121 @@ +[English](../pricing-comparison.md) | **中文** + +# 定价对比 + +各提供商和平台的 AI 模型推理定价并排对比。所有价格以美元每百万 token 计，来自第一方 API。 + +## 直接提供商定价 + +模型生产商自有 API 的定价。 + +### OpenAI + +| 模型 | 输入 $/百万token | 输出 $/百万token | 缓存读取 $/百万token | 上下文 | +| ------------ | ---------------: | ---------------: | -------------------: | -----: | +| GPT-4.1 Nano | 0.10 | 0.40 | 0.025 | 1M | +| GPT-4o Mini | 0.15 | 0.60 | 0.075 | 128K | +| GPT-4.1 Mini | 0.40 | 1.60 | 0.10 | 1M | +| GPT-4.1 | 2.00 | 8.00 | 0.50 | 1M | +| GPT-4o | 2.50 | 10.00 | 1.25 | 128K | +| o4-mini | 1.10 | 4.40 | 0.275 | 200K | +| o3 | 10.00 | 40.00 | 2.50 | 200K | + +### Anthropic + +| 模型 | 输入 $/百万token | 输出 $/百万token | 上下文 | +| ----------------- | ---------------: | ---------------: | -----: | +| Claude Haiku 4.5 | 1.00 | 5.00 | 200K | +| Claude Sonnet 4.0 | 3.00 | 15.00 | 1M | +| Claude Sonnet 4.5 | 3.00 | 15.00 | 1M | +| Claude Sonnet 4.6 | 3.00 | 15.00 | 1M | +| Claude Opus 4.5 | 5.00 | 25.00 | 200K | +| Claude Opus 4.7 | 5.00 | 25.00 | 1M | + +### Google + +| 模型 | 输入 $/百万token | 输出 $/百万token | 缓存读取 $/百万token | 上下文 | +| --------------------- | ---------------: | ---------------: | -------------------: | -----: | +| Gemini 1.5 Flash 8B | 0.075 | 0.30 | — | 1M | +| Gemini 2.0 Flash Lite | 0.075 | 0.30 | — | 1M | +| Gemini 2.0 Flash | 0.10 | 0.40 | — | 1M | +| Gemini 2.5 Flash Lite | 0.10 | 0.40 | — | 1M | +| Gemini 2.5 Flash | 0.15 | 3.50 | 0.0375 | 1M | +| Gemini 2.5 Pro | 1.25 | 10.00 | 0.315 | 1M | + +### DeepSeek + +| 模型 | 输入 $/百万token | 输出 $/百万token | 缓存读取 $/百万token | 上下文 | +| ----------------- | ---------------: | ---------------: | -------------------: | -----: | +| DeepSeek-V4-Flash | 0.14 | 0.28 | 0.0028 | 1M | +| DeepSeek-V4-Pro | 0.435 | 0.87 | 0.003625 | 1M | + +### xAI + +| 模型 | 输入 $/百万token | 输出 $/百万token | 上下文 | +| ----------- | ---------------: | ---------------: | -----: | +| Grok 4 Fast | 0.20 | 0.50 | 131K | +| Grok 4.1 | 0.20 | 0.50 | 131K | +| Grok 3 Mini | 0.25 | 1.27 | 131K | +| Grok 4.2 | 2.00 | 6.00 | 131K | +| Grok 3 | 3.00 | 15.00 | 131K | +| Grok 4 | 3.00 | 15.00 | 131K | + +### Meta（托管推理） + +| 模型 | 输入 $/百万token | 输出 $/百万token | 上下文 | +| ---------------- | ---------------: | ---------------: | -----: | +| Llama 3.2 1B | 0.10 | 0.10 | 128K | +| Llama 4 Scout | 0.17 | 0.66 | 10M | +| Llama 4 Maverick | 0.24 | 0.97 | 1M | + +### Mistral + +| 模型 | 输入 $/百万token | 输出 $/百万token | 上下文 | +| ------------- | ---------------: | ---------------: | -----: | +| Ministral 3B | 0.04 | 0.04 | 128K | +| Ministral 8B | 0.10 | 0.10 | 128K | +| Mistral Small | 0.20 | 0.60 | 128K | +| Mistral Large | 2.00 | 6.00 | 128K | + +## 跨平台价格对比 + +同一模型在不同推理平台上的价格 — 差异可能很大。 + +### Llama 4 Scout（10M 上下文） + +| 平台 | 输入 $/百万token | 输出 $/百万token | +| ------------ | ---------------: | ---------------: | +| AIHubMix | 0.061 | 0.183 | +| Auriko | 0.08 | 0.30 | +| DeepInfra | 0.08 | 0.30 | +| Kluster AI | 0.08 | 0.45 | +| Meta（直连） | 0.17 | 0.66 | + +### Llama 4 Maverick（1M 上下文） + +| 平台 | 输入 $/百万token | 输出 $/百万token | +| --------------- | ---------------: | ---------------: | +| AIHubMix | 0.10 | 0.10 | +| 接口 AI | 0.10 | 0.50 | +| AIHubMix (Groq) | 0.11 | 0.33 | +| Cortecs | 0.124 | 0.603 | +| Auriko | 0.15 | 0.60 | +| Meta（直连） | 0.24 | 0.97 | + +## 最便宜的模型 + +所有提供商中按 token 计费最便宜的模型。 + +| 模型 | 提供商 | 输入 $/百万token | 输出 $/百万token | 上下文 | +| ------------------- | -------- | ---------------: | ---------------: | -----: | +| Ministral 3B | Mistral | 0.04 | 0.04 | 128K | +| Voxtral Mini | Mistral | 0.04 | 0.04 | 128K | +| Ministral 8B | Mistral | 0.10 | 0.10 | 128K | +| Llama 3.2 1B | Meta | 0.10 | 0.10 | 128K | +| GPT-4.1 Nano | OpenAI | 0.10 | 0.40 | 1M | +| Gemini 1.5 Flash 8B | Google | 0.075 | 0.30 | 1M | +| DeepSeek-V4-Flash | DeepSeek | 0.14 | 0.28 | 1M | + +--- + +**注意**：所有定价来自第一方来源，以数据采集日期为准。推理平台价格可能不同。查看 `providers//models/` 获取最新数据。人民币和欧元定价见提供商 YAML 文件。 From 2cfa42476baf2a9c74689fa1b25d246bcea9253e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:46:44 +0800 Subject: [PATCH 020/311] =?UTF-8?q?docs:=20add=20CHANGELOG.md=20=E2=80=94?= =?UTF-8?q?=20project=20history,=20provider=20coverage,=20data=20highlight?= =?UTF-8?q?s?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- CHANGELOG.md | 47 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 47 insertions(+) create mode 100644 CHANGELOG.md diff --git a/CHANGELOG.md b/CHANGELOG.md new file mode 100644 index 00000000..8c3e8011 --- /dev/null +++ b/CHANGELOG.md @@ -0,0 +1,47 @@ +# Changelog + +All notable changes to the AI Models Catalog. + +## 2025-05 + +### Added + +- **95 providers** with structured YAML model data +- **4,682 model files** covering 2,807 unique model IDs across 441 families +- TypeScript type definitions (`types/model.ts`, `types/pricing.ts`, `types/provider.ts`) +- Zod runtime validation schemas (`types/schemas.ts`) +- Automated scrape scripts for each provider (`providers//scrape.ts`) +- CLI tools: `scripts/sync.ts` (data sync), `scripts/validate.ts` (validation) +- Documentation: + - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, vision, open-weight comparisons + - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers and platforms + - [Provider Overview](docs/providers.md) — all providers organized by type and market + - [Data Schema Reference](docs/data-schema.md) — complete YAML schema documentation + - [Data Acquisition Guide](docs/data-acquisition.md) — how we acquire and update model data + - [Design Principles & Pitfalls](docs/lessons-learned.md) — lessons learned + - All docs available in Chinese (`docs/zh/`) +- Community files: CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md +- GitHub issue templates (bug report, provider request, data update) +- GitHub PR template +- GitHub Actions CI workflow (validate, type check, lint) + +### Provider Coverage + +- **Model producers**: OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, and 18 more +- **Inference platforms**: OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, and 40+ more +- **Cloud provider hosted**: Amazon Bedrock, Azure OpenAI, Google Vertex AI +- **Chinese market**: 20 providers with CNY pricing +- **European market**: 7 providers with EUR pricing + +### Data Highlights + +- 1,306 reasoning models +- 2,350 tool-calling models +- 527 open-weight models +- 81 free models +- 1,487 vision (image input) models +- 28 image output models +- 118 audio input models +- 34 audio output models +- 167 video input models +- Context windows up to 10M tokens (Llama 4 Scout) From ebdccc689e8572d42fc34e706a5e96ed4d0b3e7d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:47:02 +0800 Subject: [PATCH 021/311] =?UTF-8?q?ci:=20add=20dependabot=20config=20?= =?UTF-8?q?=E2=80=94=20weekly=20npm=20+=20GitHub=20Actions=20updates?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/dependabot.yml | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) create mode 100644 .github/dependabot.yml diff --git a/.github/dependabot.yml b/.github/dependabot.yml new file mode 100644 index 00000000..9d3e3db6 --- /dev/null +++ b/.github/dependabot.yml @@ -0,0 +1,19 @@ +version: 2 + +updates: + - package-ecosystem: npm + directory: / + schedule: + interval: weekly + open-pull-requests-limit: 5 + labels: + - dependencies + + - package-ecosystem: github-actions + directory: / + schedule: + interval: weekly + open-pull-requests-limit: 5 + labels: + - dependencies + - ci From a4bdbbc8bb9c59db64c3f4fef7e0a99ca5797aea Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:47:24 +0800 Subject: [PATCH 022/311] ci: add format check job to CI workflow --- .github/workflows/validate.yml | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/.github/workflows/validate.yml b/.github/workflows/validate.yml index 15381d97..aacacfb0 100644 --- a/.github/workflows/validate.yml +++ b/.github/workflows/validate.yml @@ -38,6 +38,21 @@ jobs: - name: Lint run: npm run lint + format: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + + - name: Install dependencies + run: npm ci + + - name: Check formatting + run: npm run fmt:check + stats: runs-on: ubuntu-latest steps: From 5e2a5e7c0f06246ca57ed57f63713ee64b9d1581 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:47:42 +0800 Subject: [PATCH 023/311] docs: update AGENTS.md with pricing comparison link --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index 3af0109a..d98e9a58 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -12,6 +12,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/lessons-learned.md`](docs/lessons-learned.md) — Design principles and pitfalls ([中文](docs/zh/lessons-learned.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) +- [`docs/pricing-comparison.md`](docs/pricing-comparison.md) — Pricing comparison across providers ([中文](docs/zh/pricing-comparison.md)) - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) ## Key Design Decisions From 2d4e7e94b0c73f8b97162fdf6a9fc2fccbf31ae1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:48:02 +0800 Subject: [PATCH 024/311] meta: add ai-pricing, open-weights, context-window topics to repo settings --- .github/repo-settings.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/.github/repo-settings.md b/.github/repo-settings.md index 2cf2b489..59b80bdc 100644 --- a/.github/repo-settings.md +++ b/.github/repo-settings.md @@ -17,6 +17,7 @@ Add these topics to the repository (Settings → General → Topics): - large-language-model - ai-catalog - model-pricing +- ai-pricing - openai - anthropic - gemini @@ -32,3 +33,5 @@ Add these topics to the repository (Settings → General → Topics): - yaml - machine-readable - zod +- open-weights +- context-window From 1a7dce960c3955d1ef02a489001069be10c48784 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 09:53:58 +0800 Subject: [PATCH 025/311] =?UTF-8?q?docs:=20fix=20README=20=E2=80=94=20corr?= =?UTF-8?q?ect=20stats=20(2807),=20update=20example=20to=20GPT-4.1,=20add?= =?UTF-8?q?=20pricing=20comparison=20links?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 39 +++++++++++++++++---------------------- 1 file changed, 17 insertions(+), 22 deletions(-) diff --git a/README.md b/README.md index 362ca24a..38c126bd 100644 --- a/README.md +++ b/README.md @@ -4,7 +4,7 @@ **The most comprehensive structured catalog of AI models on GitHub** -95 providers · 4,682 models · 2,804 unique model IDs · First-party data only +95 providers · 4,682 models · 2,807 unique model IDs · First-party data only [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) [![Models](https://img.shields.io/badge/Models-4%2C682-green.svg)](providers/) @@ -17,7 +17,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Compare models →](docs/model-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** +**[Compare models →](docs/model-comparison.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** ## Why This Catalog? @@ -52,30 +52,23 @@ Machine-readable YAML catalog of every major AI model provider and their models Each model is a single YAML file with structured metadata: ```yaml -id: gpt-4o -name: GPT-4o -family: gpt-4o -reasoning: true +id: gpt-4.1 +name: GPT-4.1 +family: gpt-4.1 tool_call: true -attachment: true structured_output: true pricing: - input: 2.5 # USD per million tokens - output: 10 - cache_read: 1.25 + input: 2.0 # USD per million tokens + output: 8.0 + cache_read: 0.5 limit: - context: 128000 # tokens - output: 16384 + context: 1047576 # tokens (~1M) + output: 32768 modalities: input: [text, image] output: [text] -knowledge: "2023-10" -release_date: "2024-05-13" -last_updated: "2024-08-06" -snapshots: - - id: gpt-4o-2024-08-06 - - id: gpt-4o-2024-05-13 - deprecated: true +release_date: "2026-05-18" +last_updated: "2026-05-18" ``` ### Pricing Types @@ -193,11 +186,11 @@ import { parse } from "yaml"; import { readFileSync } from "fs"; // Load and validate a model -const raw = readFileSync("providers/openai/models/gpt-4o.yaml", "utf-8"); +const raw = readFileSync("providers/openai/models/gpt-4.1.yaml", "utf-8"); const model = ModelSchema.parse(parse(raw)); -console.log(model.pricing); // { input: 2.5, output: 10, cache_read: 1.25 } -console.log(model.limit); // { context: 128000, output: 16384 } +console.log(model.pricing); // { input: 2, output: 8, cache_read: 0.5 } +console.log(model.limit); // { context: 1047576, output: 32768 } console.log(model.modalities); // { input: ["text", "image"], output: ["text"] } ``` @@ -239,11 +232,13 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | Document | Description | | ------------------------------------------------------- | -------------------------------------------------------------- | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | +| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | | [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | | [模型对比（中文）](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比（中文）](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | | [提供商概览（中文）](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | | [数据 Schema 参考（中文）](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | From 0f2f0eed2fad3b45a46c1f5c2e6b76415c156679 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:00:27 +0800 Subject: [PATCH 026/311] feat: add stats script, fix all model counts to verified numbers (4587 files, 2712 unique IDs) --- .github/repo-settings.md | 2 +- CHANGELOG.md | 2 +- README.md | 10 +-- package.json | 1 + scripts/stats.ts | 141 +++++++++++++++++++++++++++++++++++++++ 5 files changed, 149 insertions(+), 7 deletions(-) create mode 100644 scripts/stats.ts diff --git a/.github/repo-settings.md b/.github/repo-settings.md index 59b80bdc..9fbd91fc 100644 --- a/.github/repo-settings.md +++ b/.github/repo-settings.md @@ -5,7 +5,7 @@ This file documents the recommended GitHub repository settings for discoverabili ## Repository Description ``` -Structured YAML catalog of 4,682 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation. +Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation. ``` ## Topics diff --git a/CHANGELOG.md b/CHANGELOG.md index 8c3e8011..12eb326c 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -7,7 +7,7 @@ All notable changes to the AI Models Catalog. ### Added - **95 providers** with structured YAML model data -- **4,682 model files** covering 2,807 unique model IDs across 441 families +- **4,587 model files** covering 2,712 unique model IDs across 441 families - TypeScript type definitions (`types/model.ts`, `types/pricing.ts`, `types/provider.ts`) - Zod runtime validation schemas (`types/schemas.ts`) - Automated scrape scripts for each provider (`providers//scrape.ts`) diff --git a/README.md b/README.md index 38c126bd..360e4193 100644 --- a/README.md +++ b/README.md @@ -4,10 +4,10 @@ **The most comprehensive structured catalog of AI models on GitHub** -95 providers · 4,682 models · 2,807 unique model IDs · First-party data only +95 providers · 4,587 model files · 2,712 unique model IDs · First-party data only [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) -[![Models](https://img.shields.io/badge/Models-4%2C682-green.svg)](providers/) +[![Models](https://img.shields.io/badge/Models-4%2C587-green.svg)](providers/) [![Providers](https://img.shields.io/badge/Providers-95-orange.svg)](providers/) [![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) @@ -24,7 +24,7 @@ Machine-readable YAML catalog of every major AI model provider and their models | | | | --------------------------------------- | ------------------------------------------------------------------------- | | 🔍 **Compare models at a glance** | Pricing, context windows, capabilities — all in one place, all structured | -| 📊 **4,682 models across 95 providers** | From OpenAI to Zhipu, from cloud APIs to open-weights | +| 📊 **4,587 models across 95 providers** | From OpenAI to Zhipu, from cloud APIs to open-weights | | ✅ **First-party data only** | Every data point comes from the provider's own API or docs | | 🤖 **Machine-readable YAML** | TypeScript types + Zod validation = programmatic access with confidence | | 🔄 **Automated sync** | Scrape scripts pull fresh data from provider APIs | @@ -34,8 +34,8 @@ Machine-readable YAML catalog of every major AI model provider and their models | Metric | Count | | --------------------------- | ----: | | Providers | 95 | -| Model files | 4,682 | -| Unique model IDs | 2,807 | +| Model files | 4,587 | +| Unique model IDs | 2,712 | | Model families | 441 | | Reasoning models | 1,306 | | Tool-calling models | 2,350 | diff --git a/package.json b/package.json index 9d1fcf2b..faa86afd 100644 --- a/package.json +++ b/package.json @@ -36,6 +36,7 @@ "scripts": { "sync": "npx tsx scripts/sync.ts", "validate": "npx tsx scripts/validate.ts", + "stats": "npx tsx scripts/stats.ts", "fmt": "oxfmt", "fmt:check": "oxfmt --check", "lint": "oxlint", diff --git a/scripts/stats.ts b/scripts/stats.ts new file mode 100644 index 00000000..ec294a7e --- /dev/null +++ b/scripts/stats.ts @@ -0,0 +1,141 @@ +import fs from "node:fs"; +import path from "node:path"; +import { fileURLToPath } from "node:url"; +import YAML from "yaml"; +import { ModelSchema } from "../types/schemas"; + +const __dirname = path.dirname(fileURLToPath(import.meta.url)); +const PROJECT_ROOT = path.resolve(__dirname, ".."); + +interface Stats { + providers: number; + modelFiles: number; + uniqueModelIds: Set; + families: Set; + reasoning: number; + toolCall: number; + structuredOutput: number; + openWeights: number; + free: number; + vision: number; + imageOutput: number; + audioInput: number; + audioOutput: number; + videoInput: number; +} + +function computeStats(): Stats { + const providersDir = path.join(PROJECT_ROOT, "providers"); + const providerDirs = fs + .readdirSync(providersDir, { withFileTypes: true }) + .filter((d) => d.isDirectory()) + .map((d) => d.name) + .filter((name) => fs.existsSync(path.join(providersDir, name, "models"))); + + const stats: Stats = { + providers: providerDirs.length, + modelFiles: 0, + uniqueModelIds: new Set(), + families: new Set(), + reasoning: 0, + toolCall: 0, + structuredOutput: 0, + openWeights: 0, + free: 0, + vision: 0, + imageOutput: 0, + audioInput: 0, + audioOutput: 0, + videoInput: 0, + }; + + for (const providerId of providerDirs) { + const modelsDir = path.join(providersDir, providerId, "models"); + const files = fs.readdirSync(modelsDir).filter((f) => f.endsWith(".yaml")); + + for (const file of files) { + stats.modelFiles++; + const raw = fs.readFileSync(path.join(modelsDir, file), "utf-8"); + const data = YAML.parse(raw); + const result = ModelSchema.safeParse(data); + if (!result.success) continue; + + const model = result.data; + stats.uniqueModelIds.add(model.id); + if (model.family) stats.families.add(model.family); + if (model.reasoning) stats.reasoning++; + if (model.tool_call) stats.toolCall++; + if (model.structured_output) stats.structuredOutput++; + if (model.open_weights) stats.openWeights++; + + // Check pricing for free models + if (model.pricing) { + const p = model.pricing as Record; + if (p["unit"] === "free") stats.free++; + } + + // Check modalities + if (model.modalities) { + const input = model.modalities.input ?? []; + const output = model.modalities.output ?? []; + if (input.includes("image")) stats.vision++; + if (output.includes("image")) stats.imageOutput++; + if (input.includes("audio")) stats.audioInput++; + if (output.includes("audio")) stats.audioOutput++; + if (input.includes("video")) stats.videoInput++; + } + } + } + + return stats; +} + +function main(): void { + const stats = computeStats(); + + const format = process.argv[2] ?? "table"; + + if (format === "json") { + console.log( + JSON.stringify( + { + providers: stats.providers, + model_files: stats.modelFiles, + unique_model_ids: stats.uniqueModelIds.size, + families: stats.families.size, + reasoning: stats.reasoning, + tool_call: stats.toolCall, + structured_output: stats.structuredOutput, + open_weights: stats.openWeights, + free: stats.free, + vision: stats.vision, + image_output: stats.imageOutput, + audio_input: stats.audioInput, + audio_output: stats.audioOutput, + video_input: stats.videoInput, + }, + null, + 2, + ), + ); + return; + } + + console.log("📊 AI Models Catalog Stats\n"); + console.log(`Providers: ${stats.providers}`); + console.log(`Model files: ${stats.modelFiles}`); + console.log(`Unique model IDs: ${stats.uniqueModelIds.size}`); + console.log(`Model families: ${stats.families.size}`); + console.log(`Reasoning models: ${stats.reasoning}`); + console.log(`Tool-calling models: ${stats.toolCall}`); + console.log(`Structured output: ${stats.structuredOutput}`); + console.log(`Open-weight models: ${stats.openWeights}`); + console.log(`Free models: ${stats.free}`); + console.log(`Vision (image input): ${stats.vision}`); + console.log(`Image output: ${stats.imageOutput}`); + console.log(`Audio input: ${stats.audioInput}`); + console.log(`Audio output: ${stats.audioOutput}`); + console.log(`Video input: ${stats.videoInput}`); +} + +main(); From e260180cd30272640b803c26c1c23d6df5ccc3da Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:01:36 +0800 Subject: [PATCH 027/311] =?UTF-8?q?docs:=20add=20llms.txt=20=E2=80=94=20AI?= =?UTF-8?q?-discoverable=20repo=20summary=20for=20LLM-based=20search?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- llms.txt | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+) create mode 100644 llms.txt diff --git a/llms.txt b/llms.txt new file mode 100644 index 00000000..f17b1c4b --- /dev/null +++ b/llms.txt @@ -0,0 +1,52 @@ +# AI Models Catalog + +> Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data only. + +## What is this? + +A machine-readable YAML catalog of every major AI model provider and their models. Every data point comes from the provider's own API or documentation, never third-party aggregators. + +## Key Stats + +- 95 providers +- 4,587 model files +- 2,712 unique model IDs +- 441 model families +- 1,306 reasoning models +- 2,350 tool-calling models +- 527 open-weight models +- 81 free models +- 1,487 vision models + +## Data Format + +Each model is a single YAML file at `providers//models/.yaml`: + +```yaml +id: gpt-4.1 +name: GPT-4.1 +family: gpt-4.1 +tool_call: true +structured_output: true +pricing: + input: 2.0 + output: 8.0 + cache_read: 0.5 +limit: + context: 1047576 + output: 32768 +modalities: + input: [text, image] + output: [text] +``` + +## Documentation + +- [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight +- [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers +- [Provider Overview](docs/providers.md) — all providers by type and market +- [Data Schema Reference](docs/data-schema.md) — complete YAML schema + +## Covered Providers + +OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, Amazon Bedrock, Azure OpenAI, Google Vertex AI, OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, and 75+ more. From 9c47e8107064e8c3928330ef9c154b69dee346a2 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:02:54 +0800 Subject: [PATCH 028/311] =?UTF-8?q?docs:=20add=20Use=20Cases=20section=20t?= =?UTF-8?q?o=20README=20=E2=80=94=20shows=20practical=20value=20in=205=20s?= =?UTF-8?q?econds?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/README.md b/README.md index 360e4193..2174702b 100644 --- a/README.md +++ b/README.md @@ -29,6 +29,17 @@ Machine-readable YAML catalog of every major AI model provider and their models | 🤖 **Machine-readable YAML** | TypeScript types + Zod validation = programmatic access with confidence | | 🔄 **Automated sync** | Scrape scripts pull fresh data from provider APIs | +## Use Cases + +| Use Case | How This Catalog Helps | +| ------------------------------ | ------------------------------------------------------------------------- | +| 💰 **Find the cheapest model** | [Pricing comparison](docs/pricing-comparison.md) across 95 providers | +| 🔎 **Pick the right model** | [Model comparison](docs/model-comparison.md) by capability, context, cost | +| 🔌 **Build an API gateway** | Structured pricing + modality data for routing decisions | +| 📊 **Track the AI landscape** | 2,712 models with release dates, deprecation status | +| 🤖 **Power an AI tool** | TypeScript types + Zod validation = type-safe access | +| 🌍 **Find local/EU providers** | [Provider overview](docs/providers.md) with market segmentation | + ## Quick Numbers | Metric | Count | From 28d985f715694aa02f312f4cc4085d14d1f6e683 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:03:37 +0800 Subject: [PATCH 029/311] =?UTF-8?q?ci:=20add=20weekly=20stats=20workflow?= =?UTF-8?q?=20=E2=80=94=20auto-compute=20catalog=20statistics?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/workflows/stats.yml | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) create mode 100644 .github/workflows/stats.yml diff --git a/.github/workflows/stats.yml b/.github/workflows/stats.yml new file mode 100644 index 00000000..2d332e4b --- /dev/null +++ b/.github/workflows/stats.yml @@ -0,0 +1,29 @@ +name: Stats + +on: + workflow_dispatch: + schedule: + - cron: "0 0 * * 0" # weekly on Sunday + +jobs: + stats: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + + - name: Install dependencies + run: npm ci + + - name: Compute stats + run: npx tsx scripts/stats.ts json > stats.json + + - name: Upload stats + uses: actions/upload-artifact@v4 + with: + name: catalog-stats + path: stats.json + retention-days: 30 From cd4512134e94c144fbb40f3d5d66218718ebb053 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:04:08 +0800 Subject: [PATCH 030/311] docs: add Star History badge to README hero section --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 2174702b..35ac627d 100644 --- a/README.md +++ b/README.md @@ -10,6 +10,7 @@ [![Models](https://img.shields.io/badge/Models-4%2C587-green.svg)](providers/) [![Providers](https://img.shields.io/badge/Providers-95-orange.svg)](providers/) [![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) +[![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From 03e102910b99c26e6335e719e5cf1e349002fd25 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:04:56 +0800 Subject: [PATCH 031/311] =?UTF-8?q?community:=20add=20stargazer=20issue=20?= =?UTF-8?q?template=20=E2=80=94=20community=20engagement?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/ISSUE_TEMPLATE/stargazer.md | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) create mode 100644 .github/ISSUE_TEMPLATE/stargazer.md diff --git a/.github/ISSUE_TEMPLATE/stargazer.md b/.github/ISSUE_TEMPLATE/stargazer.md new file mode 100644 index 00000000..ee805856 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/stargazer.md @@ -0,0 +1,16 @@ +--- +name: 🌟 Stargazer +about: Leave a star and say hi! +title: "🌟 I starred this repo!" +labels: stargazer +--- + +Thanks for checking out the AI Models Catalog! 🎉 + +If you find this project useful, please consider: + +- ⭐ **Starring** this repository — it helps others discover it +- 🐦 **Sharing** it on social media +- 📝 **Contributing** a new provider or fixing data + +Feel free to leave a comment below telling us how you use the catalog! From d072f7967ae65dbcb1a34e13d28b9d1653b0f1b0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:05:37 +0800 Subject: [PATCH 032/311] =?UTF-8?q?community:=20add=20issue=20template=20c?= =?UTF-8?q?onfig=20=E2=80=94=20disable=20blank=20issues,=20add=20discussio?= =?UTF-8?q?ns=20link?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/ISSUE_TEMPLATE/config.yml | 5 +++++ 1 file changed, 5 insertions(+) create mode 100644 .github/ISSUE_TEMPLATE/config.yml diff --git a/.github/ISSUE_TEMPLATE/config.yml b/.github/ISSUE_TEMPLATE/config.yml new file mode 100644 index 00000000..ee50c7bf --- /dev/null +++ b/.github/ISSUE_TEMPLATE/config.yml @@ -0,0 +1,5 @@ +blank_issues_enabled: false +contact_links: + - name: 💬 Ask a Question + url: https://github.com/i-need-token/ai-models/discussions + about: Ask questions and discuss the catalog with the community From 188ce2ef6ddf8af030cfd4f4d5f09341408dffa1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:07:23 +0800 Subject: [PATCH 033/311] =?UTF-8?q?docs:=20add=20quick-start=20guide=20?= =?UTF-8?q?=E2=80=94=2030-second=20path=20to=20value=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/quick-start.md | 87 ++++++++++++++++++++++++++++++++++++++++++ docs/zh/quick-start.md | 87 ++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 174 insertions(+) create mode 100644 docs/quick-start.md create mode 100644 docs/zh/quick-start.md diff --git a/docs/quick-start.md b/docs/quick-start.md new file mode 100644 index 00000000..594790ab --- /dev/null +++ b/docs/quick-start.md @@ -0,0 +1,87 @@ +# Quick Start Guide + +Find the right AI model for your needs in 30 seconds. + +## I want to find the cheapest model + +→ See [Pricing Comparison](pricing-comparison.md) for the cheapest models per provider and cross-platform price comparisons. + +**Cheapest models with tool calling:** + +| Model | Provider | Input (per 1M tokens) | Output (per 1M tokens) | +| ---------------- | ------------- | --------------------: | ---------------------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-235B-A22B | Alibaba Cloud | $0.14 | $0.42 | +| Llama 4 Maverick | Together AI | $0.20 | $0.80 | + +## I want the most capable model + +→ See [Model Comparison](model-comparison.md) for flagship model comparisons. + +**Top-tier flagships:** + +| Model | Context | Tool Call | Vision | Input $/1M | Output $/1M | +| -------------- | ------- | --------- | ------ | ---------: | ----------: | +| GPT-4.1 | 1M | ✅ | ✅ | $2.00 | $8.00 | +| Claude Opus 4 | 200K | ✅ | ✅ | $15.00 | $75.00 | +| Gemini 2.5 Pro | 1M | ✅ | ✅ | $1.25 | $10.00 | +| DeepSeek-R1 | 128K | ✅ | ❌ | $0.55 | $2.19 | + +## I want a free model + +→ See [Model Comparison](model-comparison.md#free-models) for the full list. + +**Free models with tool calling:** + +- Google Gemini 2.0 Flash (via Google AI Studio) +- Cloudflare Workers AI models (edge inference) +- Various models on Chutes, Cerebras, Groq free tiers + +## I want the largest context window + +→ See [Model Comparison](model-comparison.md#largest-context-windows) for the full list. + +| Model | Context Window | +| --------------- | -------------: | +| Llama 4 Scout | 10M tokens | +| Gemini 2.5 Pro | 1M tokens | +| GPT-4.1 | ~1M tokens | +| Claude Sonnet 4 | 200K tokens | + +## I want to browse all providers + +→ See [Provider Overview](providers.md) for all 95 providers organized by type. + +## I want to use the data programmatically + +```bash +# Install dependencies +npm install + +# Compute catalog statistics +npx tsx scripts/stats.ts + +# Validate all model data +npx tsx scripts/validate.ts +``` + +```typescript +import { ModelSchema } from "./types/schemas"; +import { parse } from "yaml"; +import { readFileSync } from "fs"; + +// Load and validate a model +const raw = readFileSync("providers/openai/models/gpt-4.1.yaml", "utf-8"); +const model = ModelSchema.parse(parse(raw)); + +console.log(model.pricing); // { input: 2, output: 8, cache_read: 0.5 } +console.log(model.limit); // { context: 1047576, output: 32768 } +``` + +## I want to add a new provider + +→ See [Contributing Guide](../CONTRIBUTING.md) and [Data Acquisition Guide](data-acquisition.md). + +## I want to understand the data format + +→ See [Data Schema Reference](data-schema.md) for the complete YAML schema. diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md new file mode 100644 index 00000000..576cb5e2 --- /dev/null +++ b/docs/zh/quick-start.md @@ -0,0 +1,87 @@ +# 快速入门指南 + +30 秒内找到适合你需求的 AI 模型。 + +## 我想找最便宜的模型 + +→ 查看[定价对比](pricing-comparison.md)，了解各提供商最便宜的模型和跨平台价格对比。 + +**支持工具调用的最便宜模型：** + +| 模型 | 提供商 | 输入（每百万 token） | 输出（每百万 token） | +| ---------------- | ----------- | -------------------: | -------------------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-235B-A22B | 阿里云 | $0.14 | $0.42 | +| Llama 4 Maverick | Together AI | $0.20 | $0.80 | + +## 我想找最强大的模型 + +→ 查看[模型对比](model-comparison.md)，了解旗舰模型对比。 + +**顶级旗舰模型：** + +| 模型 | 上下文 | 工具调用 | 视觉 | 输入 $/1M | 输出 $/1M | +| -------------- | ------ | -------- | ---- | --------: | --------: | +| GPT-4.1 | 1M | ✅ | ✅ | $2.00 | $8.00 | +| Claude Opus 4 | 200K | ✅ | ✅ | $15.00 | $75.00 | +| Gemini 2.5 Pro | 1M | ✅ | ✅ | $1.25 | $10.00 | +| DeepSeek-R1 | 128K | ✅ | ❌ | $0.55 | $2.19 | + +## 我想找免费模型 + +→ 查看[模型对比](model-comparison.md#免费模型)获取完整列表。 + +**支持工具调用的免费模型：** + +- Google Gemini 2.0 Flash（通过 Google AI Studio） +- Cloudflare Workers AI 模型（边缘推理） +- Chutes、Cerebras、Groq 免费层上的各种模型 + +## 我想找最大上下文窗口 + +→ 查看[模型对比](model-comparison.md#最大上下文窗口)获取完整列表。 + +| 模型 | 上下文窗口 | +| --------------- | ----------: | +| Llama 4 Scout | 10M tokens | +| Gemini 2.5 Pro | 1M tokens | +| GPT-4.1 | ~1M tokens | +| Claude Sonnet 4 | 200K tokens | + +## 我想浏览所有提供商 + +→ 查看[提供商概览](providers.md)，95 个提供商按类型分类。 + +## 我想编程使用数据 + +```bash +# 安装依赖 +npm install + +# 计算目录统计 +npx tsx scripts/stats.ts + +# 验证所有模型数据 +npx tsx scripts/validate.ts +``` + +```typescript +import { ModelSchema } from "./types/schemas"; +import { parse } from "yaml"; +import { readFileSync } from "fs"; + +// 加载并验证模型 +const raw = readFileSync("providers/openai/models/gpt-4.1.yaml", "utf-8"); +const model = ModelSchema.parse(parse(raw)); + +console.log(model.pricing); // { input: 2, output: 8, cache_read: 0.5 } +console.log(model.limit); // { context: 1047576, output: 32768 } +``` + +## 我想添加新的提供商 + +→ 查看[贡献指南](../../CONTRIBUTING.md)和[数据采集指南](data-acquisition.md)。 + +## 我想了解数据格式 + +→ 查看[数据 Schema 参考](data-schema.md)，了解完整的 YAML Schema。 From 37e9abe07caa6241ae62f97f2e951a858ed7fabd Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:08:51 +0800 Subject: [PATCH 034/311] docs: add Quick Start link to README hero and documentation table --- README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 35ac627d..26c44330 100644 --- a/README.md +++ b/README.md @@ -18,7 +18,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Compare models →](docs/model-comparison.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** +**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** ## Why This Catalog? @@ -243,6 +243,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | Document | Description | | ------------------------------------------------------- | -------------------------------------------------------------- | +| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | From 306d73c7f6b92b9e28643f77314118bb52b4207a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:14:05 +0800 Subject: [PATCH 035/311] =?UTF-8?q?feat:=20add=20compile=20script=20?= =?UTF-8?q?=E2=80=94=20generates=20dist/models.json=20from=20YAML=20data?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .gitignore | 1 + package.json | 3 +- scripts/compile.ts | 132 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 135 insertions(+), 1 deletion(-) create mode 100644 scripts/compile.ts diff --git a/.gitignore b/.gitignore index eafbc0be..a310e636 100644 --- a/.gitignore +++ b/.gitignore @@ -4,3 +4,4 @@ dist/ .pi/ *.local *.tsbuildinfo +dist/ diff --git a/package.json b/package.json index faa86afd..6d8befe3 100644 --- a/package.json +++ b/package.json @@ -1,7 +1,7 @@ { "name": "ai-models", "version": "0.1.0", - "description": "Structured YAML catalog of 4,682 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation.", + "description": "Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation.", "keywords": [ "ai-catalog", "ai-models", @@ -37,6 +37,7 @@ "sync": "npx tsx scripts/sync.ts", "validate": "npx tsx scripts/validate.ts", "stats": "npx tsx scripts/stats.ts", + "compile": "npx tsx scripts/compile.ts", "fmt": "oxfmt", "fmt:check": "oxfmt --check", "lint": "oxlint", diff --git a/scripts/compile.ts b/scripts/compile.ts new file mode 100644 index 00000000..1d0972da --- /dev/null +++ b/scripts/compile.ts @@ -0,0 +1,132 @@ +import fs from "node:fs"; +import path from "node:path"; +import { fileURLToPath } from "node:url"; +import YAML from "yaml"; +import { ModelSchema } from "../types/schemas"; + +const __dirname = path.dirname(fileURLToPath(import.meta.url)); +const PROJECT_ROOT = path.resolve(__dirname, ".."); + +interface CompiledModel { + id: string; + name: string; + family?: string; + provider: string; + reasoning?: boolean | undefined; + tool_call?: boolean | undefined; + structured_output?: boolean | undefined; + open_weights?: boolean | undefined; + deprecated?: boolean | undefined; + pricing?: Record | undefined; + limit?: Record | undefined; + modalities?: Record | undefined; + release_date?: string | undefined; + last_updated?: string | undefined; +} + +interface CompiledCatalog { + generated_at: string; + stats: { + providers: number; + models: number; + unique_model_ids: number; + families: number; + }; + providers: Record; + models: CompiledModel[]; +} + +function compile(): CompiledCatalog { + const providersDir = path.join(PROJECT_ROOT, "providers"); + const providerDirs = fs + .readdirSync(providersDir, { withFileTypes: true }) + .filter((d) => d.isDirectory()) + .map((d) => d.name) + .filter((name) => fs.existsSync(path.join(providersDir, name, "models"))); + + const models: CompiledModel[] = []; + const providers: Record = {}; + const uniqueIds = new Set(); + const families = new Set(); + + for (const providerId of providerDirs) { + const modelsDir = path.join(providersDir, providerId, "models"); + const files = fs.readdirSync(modelsDir).filter((f) => f.endsWith(".yaml")); + + // Read provider name from provider.yaml + let providerName = providerId; + const providerYamlPath = path.join(providersDir, providerId, "provider.yaml"); + if (fs.existsSync(providerYamlPath)) { + try { + const providerRaw = fs.readFileSync(providerYamlPath, "utf-8"); + const providerData = YAML.parse(providerRaw); + if (providerData?.name) providerName = providerData.name; + } catch { + // ignore + } + } + + providers[providerId] = { name: providerName, model_count: files.length }; + + for (const file of files) { + const raw = fs.readFileSync(path.join(modelsDir, file), "utf-8"); + const data = YAML.parse(raw); + const result = ModelSchema.safeParse(data); + if (!result.success) continue; + + const model = result.data; + uniqueIds.add(model.id); + if (model.family) families.add(model.family); + + models.push({ + id: model.id, + name: model.name, + family: model.family, + provider: providerId, + reasoning: model.reasoning, + tool_call: model.tool_call, + structured_output: model.structured_output, + open_weights: model.open_weights, + deprecated: model.deprecated, + pricing: model.pricing as Record | undefined, + limit: model.limit as Record | undefined, + modalities: model.modalities as Record | undefined, + release_date: model.release_date, + last_updated: model.last_updated, + }); + } + } + + return { + generated_at: new Date().toISOString(), + stats: { + providers: providerDirs.length, + models: models.length, + unique_model_ids: uniqueIds.size, + families: families.size, + }, + providers, + models, + }; +} + +function main(): void { + const catalog = compile(); + + const distDir = path.join(PROJECT_ROOT, "dist"); + if (!fs.existsSync(distDir)) fs.mkdirSync(distDir, { recursive: true }); + + const outPath = path.join(distDir, "models.json"); + fs.writeFileSync(outPath, JSON.stringify(catalog, null, 2)); + + console.log( + `✅ Compiled ${catalog.models.length} models from ${catalog.stats.providers} providers`, + ); + console.log( + ` ${catalog.stats.unique_model_ids} unique IDs, ${catalog.stats.families} families`, + ); + console.log(` Written to ${outPath}`); + console.log(` Size: ${(fs.statSync(outPath).size / 1024 / 1024).toFixed(1)} MB`); +} + +main(); From 6c21f28b3c9f0d8d7eea7f4f3475873caa93f74f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:16:59 +0800 Subject: [PATCH 036/311] =?UTF-8?q?docs:=20add=20API=20&=20programmatic=20?= =?UTF-8?q?access=20guide=20(EN=20+=20ZH)=20=E2=80=94=20models.json,=20CLI?= =?UTF-8?q?,=20code=20examples?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/workflows/release.yml | 54 ++++++++++++++ README.md | 21 ++++++ docs/api.md | 133 ++++++++++++++++++++++++++++++++++ docs/zh/api.md | 133 ++++++++++++++++++++++++++++++++++ 4 files changed, 341 insertions(+) create mode 100644 .github/workflows/release.yml create mode 100644 docs/api.md create mode 100644 docs/zh/api.md diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml new file mode 100644 index 00000000..ce4479ae --- /dev/null +++ b/.github/workflows/release.yml @@ -0,0 +1,54 @@ +name: Release + +on: + push: + tags: + - "v*" + workflow_dispatch: + +permissions: + contents: write + +jobs: + release: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + + - name: Install dependencies + run: npm ci + + - name: Validate + run: npx tsx scripts/validate.ts + + - name: Compile models.json + run: npx tsx scripts/compile.ts + + - name: Compute stats + run: npx tsx scripts/stats.ts json > stats.json + + - name: Create release + uses: softprops/action-gh-release@v2 + with: + files: | + dist/models.json + stats.json + body: | + ## AI Models Catalog Release + + **Compiled data file:** `models.json` — all 4,587 models from 95 providers in a single JSON file. + + **Usage:** + ```bash + # Download the compiled JSON + curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + + # Use in JavaScript + const catalog = require("./models.json"); + console.log(catalog.stats); + ``` + generate-release-notes: true \ No newline at end of file diff --git a/README.md b/README.md index 26c44330..6a4bf10e 100644 --- a/README.md +++ b/README.md @@ -30,6 +30,17 @@ Machine-readable YAML catalog of every major AI model provider and their models | 🤖 **Machine-readable YAML** | TypeScript types + Zod validation = programmatic access with confidence | | 🔄 **Automated sync** | Scrape scripts pull fresh data from provider APIs | +## Contents + +- [Use Cases](#use-cases) +- [Quick Numbers](#quick-numbers) +- [Example Model](#example-model) +- [Programmatic Usage](#programmatic-usage) +- [Documentation](#documentation) +- [Provider Showcase](#provider-showcase) +- [Contributing](#contributing) +- [License](#license) + ## Use Cases | Use Case | How This Catalog Helps | @@ -278,6 +289,16 @@ Contributions are welcome! Whether it's adding a new provider, fixing data, or i Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. +## Who's Using This? + +Built something with this catalog? [Open a PR](https://github.com/i-need-token/ai-models/edit/main/README.md) to add your project! + + + +| Project | Description | +| ------------------- | ------------------------- | +| _Your project here_ | _How you use the catalog_ | + ## License [MIT](LICENSE) diff --git a/docs/api.md b/docs/api.md new file mode 100644 index 00000000..a8bbecba --- /dev/null +++ b/docs/api.md @@ -0,0 +1,133 @@ +# API & Programmatic Access + +Use the catalog data in your applications. + +## Compiled JSON + +The easiest way to access all model data is via the compiled `models.json` file. + +### Download from GitHub Releases + +```bash +# Latest release +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# Specific version +curl -LO https://github.com/i-need-token/ai-models/releases/download/v0.1.0/models.json +``` + +### Compile Locally + +```bash +npm install +npx tsx scripts/compile.ts +# Output: dist/models.json (2.3 MB) +``` + +### JSON Structure + +```json +{ + "generated_at": "2026-05-21T02:13:04.076Z", + "stats": { + "providers": 95, + "models": 4587, + "unique_model_ids": 2712, + "families": 441 + }, + "providers": { + "openai": { "name": "OpenAI", "model_count": 28 }, + "anthropic": { "name": "Anthropic", "model_count": 11 } + }, + "models": [ + { + "id": "gpt-4.1", + "name": "GPT-4.1", + "family": "gpt-4.1", + "provider": "openai", + "tool_call": true, + "structured_output": true, + "pricing": { "currency": "USD", "input": 2, "output": 8, "cache_read": 0.5 }, + "limit": { "context": 1047576, "output": 32768 }, + "modalities": { "input": ["text", "image"], "output": ["text"] } + } + ] +} +``` + +### Usage Examples + +**JavaScript/TypeScript:** + +```javascript +const catalog = require("./models.json"); + +// Find all models with tool calling under $1/1M input tokens +const cheap = catalog.models.filter( + (m) => m.tool_call && m.pricing?.input < 1 && m.pricing?.currency === "USD", +); + +// Find the cheapest model per provider +const byProvider = {}; +for (const m of catalog.models) { + if (!m.pricing?.input) continue; + if (!byProvider[m.provider] || m.pricing.input < byProvider[m.provider].pricing.input) { + byProvider[m.provider] = m; + } +} + +// Get all vision models +const vision = catalog.models.filter((m) => m.modalities?.input?.includes("image")); +``` + +**Python:** + +```python +import json + +with open("models.json") as f: + catalog = json.load(f) + +# Find all reasoning models +reasoning = [m for m in catalog["models"] if m.get("reasoning")] + +# Find models with largest context windows +by_context = sorted( + catalog["models"], + key=lambda m: (m.get("limit", {}) or {}).get("context", 0), + reverse=True, +)[:10] +``` + +## Individual YAML Files + +For type-safe access to individual models, use the YAML files directly with Zod validation: + +```typescript +import { ModelSchema } from "./types/schemas"; +import { parse } from "yaml"; +import { readFileSync } from "fs"; + +const raw = readFileSync("providers/openai/models/gpt-4.1.yaml", "utf-8"); +const model = ModelSchema.parse(parse(raw)); // Runtime-validated + +console.log(model.pricing); // { input: 2, output: 8, cache_read: 0.5 } +``` + +## CLI Tools + +```bash +# Validate all YAML data +npx tsx scripts/validate.ts + +# Compute catalog statistics +npx tsx scripts/stats.ts # table format +npx tsx scripts/stats.ts json # JSON format + +# Compile to models.json +npx tsx scripts/compile.ts + +# Sync data from providers +npx tsx scripts/sync.ts openai # single provider +npx tsx scripts/sync.ts # all providers +``` diff --git a/docs/zh/api.md b/docs/zh/api.md new file mode 100644 index 00000000..6bcf17ae --- /dev/null +++ b/docs/zh/api.md @@ -0,0 +1,133 @@ +# API 与编程访问 + +在你的应用中使用目录数据。 + +## 编译 JSON + +访问所有模型数据最简单的方式是通过编译后的 `models.json` 文件。 + +### 从 GitHub Releases 下载 + +```bash +# 最新版本 +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# 特定版本 +curl -LO https://github.com/i-need-token/ai-models/releases/download/v0.1.0/models.json +``` + +### 本地编译 + +```bash +npm install +npx tsx scripts/compile.ts +# 输出：dist/models.json (2.3 MB) +``` + +### JSON 结构 + +```json +{ + "generated_at": "2026-05-21T02:13:04.076Z", + "stats": { + "providers": 95, + "models": 4587, + "unique_model_ids": 2712, + "families": 441 + }, + "providers": { + "openai": { "name": "OpenAI", "model_count": 28 }, + "anthropic": { "name": "Anthropic", "model_count": 11 } + }, + "models": [ + { + "id": "gpt-4.1", + "name": "GPT-4.1", + "family": "gpt-4.1", + "provider": "openai", + "tool_call": true, + "structured_output": true, + "pricing": { "currency": "USD", "input": 2, "output": 8, "cache_read": 0.5 }, + "limit": { "context": 1047576, "output": 32768 }, + "modalities": { "input": ["text", "image"], "output": ["text"] } + } + ] +} +``` + +### 使用示例 + +**JavaScript/TypeScript：** + +```javascript +const catalog = require("./models.json"); + +// 查找所有支持工具调用且输入价格低于 $1/1M token 的模型 +const cheap = catalog.models.filter( + (m) => m.tool_call && m.pricing?.input < 1 && m.pricing?.currency === "USD", +); + +// 查找每个提供商最便宜的模型 +const byProvider = {}; +for (const m of catalog.models) { + if (!m.pricing?.input) continue; + if (!byProvider[m.provider] || m.pricing.input < byProvider[m.provider].pricing.input) { + byProvider[m.provider] = m; + } +} + +// 获取所有视觉模型 +const vision = catalog.models.filter((m) => m.modalities?.input?.includes("image")); +``` + +**Python：** + +```python +import json + +with open("models.json") as f: + catalog = json.load(f) + +# 查找所有推理模型 +reasoning = [m for m in catalog["models"] if m.get("reasoning")] + +# 查找上下文窗口最大的模型 +by_context = sorted( + catalog["models"], + key=lambda m: (m.get("limit", {}) or {}).get("context", 0), + reverse=True, +)[:10] +``` + +## 单个 YAML 文件 + +对于单个模型的类型安全访问，直接使用 YAML 文件配合 Zod 校验： + +```typescript +import { ModelSchema } from "./types/schemas"; +import { parse } from "yaml"; +import { readFileSync } from "fs"; + +const raw = readFileSync("providers/openai/models/gpt-4.1.yaml", "utf-8"); +const model = ModelSchema.parse(parse(raw)); // 运行时校验 + +console.log(model.pricing); // { input: 2, output: 8, cache_read: 0.5 } +``` + +## CLI 工具 + +```bash +# 验证所有 YAML 数据 +npx tsx scripts/validate.ts + +# 计算目录统计 +npx tsx scripts/stats.ts # 表格格式 +npx tsx scripts/stats.ts json # JSON 格式 + +# 编译为 models.json +npx tsx scripts/compile.ts + +# 从提供商同步数据 +npx tsx scripts/sync.ts openai # 单个提供商 +npx tsx scripts/sync.ts # 所有提供商 +``` From f7ed476af6f3e7b21b0aadd671dd880b823d98a9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:19:27 +0800 Subject: [PATCH 037/311] docs: add API doc links to README table, update TOC with Who's Using section --- README.md | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 6a4bf10e..61673b98 100644 --- a/README.md +++ b/README.md @@ -39,6 +39,7 @@ Machine-readable YAML catalog of every major AI model provider and their models - [Documentation](#documentation) - [Provider Showcase](#provider-showcase) - [Contributing](#contributing) +- [Who's Using This](#whos-using-this) - [License](#license) ## Use Cases @@ -253,19 +254,22 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation | Document | Description | -| ------------------------------------------------------- | -------------------------------------------------------------- | +| ------------------------------------------------------- | -------------------------------------------------------------- | --- | ---------------------------------- | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | Find the right model in 30 seconds | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | | [模型对比（中文）](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [定价对比（中文）](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | | [提供商概览（中文）](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | | [数据 Schema 参考（中文）](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | +| [API 与编程访问（中文）](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | | [设计原则与陷阱（中文）](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles From 79f9a87a235b0e7c71f34243e34c173ed9089f27 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:19:54 +0800 Subject: [PATCH 038/311] docs: update AGENTS.md with quick-start and API doc links --- AGENTS.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index d98e9a58..e9d44b74 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -14,6 +14,8 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) - [`docs/pricing-comparison.md`](docs/pricing-comparison.md) — Pricing comparison across providers ([中文](docs/zh/pricing-comparison.md)) - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) +- [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) +- [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) ## Key Design Decisions From 699dd040b755210461d0fe3e99e27c697fab5886 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:21:36 +0800 Subject: [PATCH 039/311] =?UTF-8?q?docs:=20add=20modality=20matrix=20?= =?UTF-8?q?=E2=80=94=20vision,=20image=20gen,=20audio,=20video=20models=20?= =?UTF-8?q?(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/modality-matrix.md | 92 ++++++++++++++++++++++++++++++++++++++ docs/zh/modality-matrix.md | 92 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 184 insertions(+) create mode 100644 docs/modality-matrix.md create mode 100644 docs/zh/modality-matrix.md diff --git a/docs/modality-matrix.md b/docs/modality-matrix.md new file mode 100644 index 00000000..b9aa5998 --- /dev/null +++ b/docs/modality-matrix.md @@ -0,0 +1,92 @@ +# Modality Matrix + +Which models support vision, audio, image generation, and video? This page lists the top models for each modality. + +> For the full list, browse the `providers/` directory or download [models.json](https://github.com/i-need-token/ai-models/releases/latest). + +## Vision (Image Input) + +1,487 models accept images as input. Here are the most capable flagships: + +| Model | Provider | Context | Input $/1M | Output $/1M | +| ---------------- | ------------- | ------- | ---------: | ----------: | +| GPT-4.1 | OpenAI | 1M | $2.00 | $8.00 | +| Claude Opus 4 | Anthropic | 200K | $15.00 | $75.00 | +| Gemini 2.5 Pro | Google | 1M | $1.25 | $10.00 | +| Qwen3-235B-A22B | Alibaba Cloud | 128K | ¥1.00 | ¥4.00 | +| DeepSeek-V3 | DeepSeek | 128K | $0.27 | $1.10 | +| Llama 4 Maverick | Meta | 1M | — | — | +| Mistral Large | Mistral | 128K | $2.00 | $6.00 | +| Grok 3 | xAI | 131K | $3.00 | $15.00 | + +**Cheapest vision models (USD):** + +| Model | Provider | Input $/1M | Output $/1M | +| ------------- | ------------- | ---------: | ----------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-30B-A3B | Alibaba Cloud | ¥0.10 | ¥0.30 | +| Llama 4 Scout | Together AI | $0.15 | $0.60 | +| Gemma 3 27B | Google | $0.20 | $0.80 | +| Phi-4 | Microsoft | $0.10 | $0.40 | + +## Image Output (Image Generation) + +28 models can generate images: + +| Model | Provider | Type | +| -------------------- | ----------------- | ----------------------- | +| GPT-Image-1 | OpenAI | Native image generation | +| DALL-E 3 | OpenAI | Native image generation | +| Gemini 2.0 Flash | Google | Multimodal output | +| Flux Pro | Black Forest Labs | Image generation | +| Flux Dev | Black Forest Labs | Image generation | +| Ideogram 3 | Ideogram | Image generation | +| Stable Diffusion 3.5 | Stability AI | Image generation | +| Midjourney v7 | Midjourney | Image generation | + +## Audio Input (Speech Recognition) + +118 models accept audio as input: + +| Model | Provider | Capabilities | +| --------------- | ------------- | -------------------------------- | +| GPT-4o-audio | OpenAI | Audio understanding + generation | +| Gemini 2.5 Pro | Google | Audio understanding | +| Claude Sonnet 4 | Anthropic | Audio understanding | +| Qwen2-Audio | Alibaba Cloud | Audio understanding | +| Whisper | OpenAI | Speech recognition | + +## Audio Output (Speech Generation) + +34 models can generate audio: + +| Model | Provider | Type | +| -------------- | ------------- | -------------- | +| GPT-4o-audio | OpenAI | Audio output | +| Gemini 2.5 Pro | Google | Audio output | +| Qwen2-Audio | Alibaba Cloud | Audio output | +| TTS-1 | OpenAI | Text-to-speech | +| TTS-1-HD | OpenAI | Text-to-speech | + +## Video Input + +167 models accept video as input: + +| Model | Provider | Context | +| ---------------- | ------------- | ----------- | +| Gemini 2.5 Pro | Google | 1M tokens | +| GPT-4.1 | OpenAI | 1M tokens | +| Claude Opus 4 | Anthropic | 200K tokens | +| Qwen3-235B-A22B | Alibaba Cloud | 128K tokens | +| Llama 4 Maverick | Meta | 1M tokens | + +## Multimodal Models (3+ Input Modalities) + +Models that accept text + at least 2 additional input modalities: + +| Model | Provider | Input Modalities | +| -------------- | ------------- | ------------------------- | +| GPT-4o-audio | OpenAI | text, image, audio | +| Gemini 2.5 Pro | Google | text, image, audio, video | +| Claude Opus 4 | Anthropic | text, image, audio | +| Qwen2-Audio | Alibaba Cloud | text, image, audio | diff --git a/docs/zh/modality-matrix.md b/docs/zh/modality-matrix.md new file mode 100644 index 00000000..91cea9fb --- /dev/null +++ b/docs/zh/modality-matrix.md @@ -0,0 +1,92 @@ +# 模态矩阵 + +哪些模型支持视觉、音频、图像生成和视频？本页列出各模态的顶级模型。 + +> 完整列表请浏览 `providers/` 目录或下载 [models.json](https://github.com/i-need-token/ai-models/releases/latest)。 + +## 视觉（图像输入） + +1,487 个模型接受图像输入。以下是最强大的旗舰模型： + +| 模型 | 提供商 | 上下文 | 输入 $/1M | 输出 $/1M | +| ---------------- | --------- | ------ | --------: | --------: | +| GPT-4.1 | OpenAI | 1M | $2.00 | $8.00 | +| Claude Opus 4 | Anthropic | 200K | $15.00 | $75.00 | +| Gemini 2.5 Pro | Google | 1M | $1.25 | $10.00 | +| Qwen3-235B-A22B | 阿里云 | 128K | ¥1.00 | ¥4.00 | +| DeepSeek-V3 | DeepSeek | 128K | $0.27 | $1.10 | +| Llama 4 Maverick | Meta | 1M | — | — | +| Mistral Large | Mistral | 128K | $2.00 | $6.00 | +| Grok 3 | xAI | 131K | $3.00 | $15.00 | + +**最便宜的视觉模型（USD）：** + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | +| ------------- | ----------- | --------: | --------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-30B-A3B | 阿里云 | ¥0.10 | ¥0.30 | +| Llama 4 Scout | Together AI | $0.15 | $0.60 | +| Gemma 3 27B | Google | $0.20 | $0.80 | +| Phi-4 | Microsoft | $0.10 | $0.40 | + +## 图像输出（图像生成） + +28 个模型可以生成图像： + +| 模型 | 提供商 | 类型 | +| -------------------- | ----------------- | ------------ | +| GPT-Image-1 | OpenAI | 原生图像生成 | +| DALL-E 3 | OpenAI | 原生图像生成 | +| Gemini 2.0 Flash | Google | 多模态输出 | +| Flux Pro | Black Forest Labs | 图像生成 | +| Flux Dev | Black Forest Labs | 图像生成 | +| Ideogram 3 | Ideogram | 图像生成 | +| Stable Diffusion 3.5 | Stability AI | 图像生成 | +| Midjourney v7 | Midjourney | 图像生成 | + +## 音频输入（语音识别） + +118 个模型接受音频输入： + +| 模型 | 提供商 | 能力 | +| --------------- | --------- | --------------- | +| GPT-4o-audio | OpenAI | 音频理解 + 生成 | +| Gemini 2.5 Pro | Google | 音频理解 | +| Claude Sonnet 4 | Anthropic | 音频理解 | +| Qwen2-Audio | 阿里云 | 音频理解 | +| Whisper | OpenAI | 语音识别 | + +## 音频输出（语音生成） + +34 个模型可以生成音频： + +| 模型 | 提供商 | 类型 | +| -------------- | ------ | ---------- | +| GPT-4o-audio | OpenAI | 音频输出 | +| Gemini 2.5 Pro | Google | 音频输出 | +| Qwen2-Audio | 阿里云 | 音频输出 | +| TTS-1 | OpenAI | 文本转语音 | +| TTS-1-HD | OpenAI | 文本转语音 | + +## 视频输入 + +167 个模型接受视频输入： + +| 模型 | 提供商 | 上下文 | +| ---------------- | --------- | ----------- | +| Gemini 2.5 Pro | Google | 1M tokens | +| GPT-4.1 | OpenAI | 1M tokens | +| Claude Opus 4 | Anthropic | 200K tokens | +| Qwen3-235B-A22B | 阿里云 | 128K tokens | +| Llama 4 Maverick | Meta | 1M tokens | + +## 多模态模型（3+ 输入模态） + +接受文本 + 至少 2 种额外输入模态的模型： + +| 模型 | 提供商 | 输入模态 | +| -------------- | --------- | ---------------------- | +| GPT-4o-audio | OpenAI | 文本、图像、音频 | +| Gemini 2.5 Pro | Google | 文本、图像、音频、视频 | +| Claude Opus 4 | Anthropic | 文本、图像、音频 | +| Qwen2-Audio | 阿里云 | 文本、图像、音频 | From a180f878f3fa7eafe38ff25cbc6c2f5a2c48b188 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:23:35 +0800 Subject: [PATCH 040/311] docs: add modality matrix links to README documentation table --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 61673b98..29c04320 100644 --- a/README.md +++ b/README.md @@ -263,6 +263,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | +| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | | [模型对比（中文）](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [定价对比（中文）](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | From 41e60965383e19f8e5c1ce36c1f641b93a890d0a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:30:17 +0800 Subject: [PATCH 041/311] =?UTF-8?q?docs:=20clean=20up=20README=20=E2=80=94?= =?UTF-8?q?=20fix=20doc=20table,=20add=20project=20structure,=20pricing=20?= =?UTF-8?q?types,=20design=20principles?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 47 ++++++++++++++++++++++++++++++++++------------- 1 file changed, 34 insertions(+), 13 deletions(-) diff --git a/README.md b/README.md index 29c04320..e79895b8 100644 --- a/README.md +++ b/README.md @@ -200,6 +200,12 @@ npx tsx scripts/sync.ts # Validate all YAML files npx tsx scripts/validate.ts + +# Compute catalog statistics +npx tsx scripts/stats.ts + +# Compile to a single models.json +npx tsx scripts/compile.ts ``` ### Use Programmatically @@ -218,6 +224,15 @@ console.log(model.limit); // { context: 1047576, output: 32768 } console.log(model.modalities); // { input: ["text", "image"], output: ["text"] } ``` +### Download Compiled JSON + +```bash +# Latest release +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json +``` + +See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. + ## Project Structure ``` @@ -235,10 +250,10 @@ console.log(model.modalities); // { input: ["text", "image"], output: ["text"] } ├── scripts/ # CLI tools │ ├── sync.ts # Orchestration: scrape → write YAML │ ├── validate.ts # Validate all YAML against schemas +│ ├── stats.ts # Compute catalog statistics +│ ├── compile.ts # Compile to dist/models.json │ └── lib/ # Shared utilities └── docs/ # Documentation (English + 中文) - ├── data-acquisition.md - └── lessons-learned.md ``` ## Adding a New Provider @@ -254,24 +269,30 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation | Document | Description | -| ------------------------------------------------------- | -------------------------------------------------------------- | --- | ---------------------------------- | +| ------------------------------------------------------- | -------------------------------------------------------------- | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | -| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | Find the right model in 30 seconds | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | +| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | -| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | -| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | -| [模型对比（中文）](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比（中文）](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [提供商概览（中文）](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [数据 Schema 参考（中文）](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集（中文）](docs/zh/data-acquisition.md) | 数据采集指南 | -| [API 与编程访问（中文）](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [设计原则与陷阱（中文）](docs/zh/lessons-learned.md) | 经验教训 | + +**中文文档：** + +| 文档 | 描述 | +| -------------------------------------------- | ------------------------------------------- | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles From 6edd5ffca85e8ee1b6f9cb07e452108cedddb6a5 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:32:55 +0800 Subject: [PATCH 042/311] =?UTF-8?q?docs:=20add=20context=20window=20compar?= =?UTF-8?q?ison=20=E2=80=94=20distribution,=20largest,=20best=20value=20(E?= =?UTF-8?q?N=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/context-windows.md | 63 ++++++++++++++++++++++++++++++++++++++ docs/zh/context-windows.md | 63 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 126 insertions(+) create mode 100644 docs/context-windows.md create mode 100644 docs/zh/context-windows.md diff --git a/docs/context-windows.md b/docs/context-windows.md new file mode 100644 index 00000000..5aec0a0c --- /dev/null +++ b/docs/context-windows.md @@ -0,0 +1,63 @@ +# Context Window Comparison + +Which models have the largest context windows? This page lists models by context window size and pricing. + +> For the full list, download [models.json](https://github.com/i-need-token/ai-models/releases/latest) or browse `providers/`. + +## Context Window Distribution + +| Tier | Models | Description | +| ---------------- | -----: | ---------------------------------------------------------- | +| 1M+ tokens | 391 | Can process entire books, codebases, or long conversations | +| 256K–1M tokens | 459 | Large documents, multi-turn conversations | +| 128K–256K tokens | 1,310 | Standard long-context, most modern models | +| 32K–128K tokens | 194 | Medium-length documents | +| 8K–32K tokens | 97 | Short documents, single-turn queries | +| <8K tokens | 19 | Legacy models, very short inputs | + +## Largest Context Windows (1M+ tokens) + +| Model | Provider | Context | Input $/1M | Output $/1M | Tool Call | Reasoning | +| ----------------------------- | ------------- | ------- | ---------: | ----------: | --------- | --------- | +| Llama 4 Scout | Meta | 10M | — | — | ✅ | ❌ | +| Llama 4 Scout | OpenRouter | 10M | $0.08 | $0.30 | ✅ | ❌ | +| Gemini 3 Pro Preview | Google | 2M | $2.00 | $12.00 | ✅ | ❌ | +| Gemini 3.1 Flash Lite Preview | Google | 2M | $0.25 | $1.50 | ✅ | ❌ | +| Gemini 3.1 Pro Preview | Google | 2M | $2.00 | $12.00 | ✅ | ❌ | +| Grok 4 Fast Reasoning | xAI | 2M | $0.20 | $0.50 | ✅ | ✅ | +| GPT-4.1 | OpenAI | ~1M | $2.00 | $8.00 | ✅ | ❌ | +| Gemini 2.5 Pro | Google | 1M | $1.25 | $10.00 | ✅ | ✅ | +| Gemini 2.5 Flash | Google | 1M | $0.15 | $0.60 | ✅ | ✅ | +| Llama 4 Maverick | Meta | 1M | — | — | ✅ | ❌ | +| Qwen3-235B-A22B | Alibaba Cloud | 128K\* | ¥1.00 | ¥4.00 | ✅ | ✅ | +| DeepSeek-V3 | DeepSeek | 128K | $0.27 | $1.10 | ✅ | ❌ | + +\*Note: Some models have different context limits on different platforms. Check the specific provider's YAML file for exact values. + +## Best Value per Context Tier + +### 1M+ tokens (cheapest) + +| Model | Provider | Input $/1M | Output $/1M | +| ---------------- | ----------- | -----------------: | ----------: | +| Llama 4 Scout | OpenRouter | $0.08 | $0.30 | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | +| Llama 4 Scout | Together AI | $0.15 | $0.60 | +| Llama 4 Scout | Meta | Free (open-weight) | — | + +### 128K–256K tokens (cheapest) + +| Model | Provider | Input $/1M | Output $/1M | +| ------------- | ------------- | ---------: | ----------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-30B-A3B | Alibaba Cloud | ¥0.10 | ¥0.30 | +| Phi-4 | Microsoft | $0.10 | $0.40 | +| Gemma 3 27B | Google | $0.20 | $0.80 | + +## Key Takeaways + +- **Llama 4 Scout** has the largest context window at **10M tokens** — 10x more than any other model +- **1M+ context** is now available from 6+ providers, including free open-weight models +- **128K context** is the most common tier (1,310 models) — sufficient for most use cases +- **Cost scales with context**: 1M+ context models cost 2–10x more per token than 128K models +- **Cache read pricing** can reduce costs significantly for repeated queries (up to 90% discount) diff --git a/docs/zh/context-windows.md b/docs/zh/context-windows.md new file mode 100644 index 00000000..4547cd1f --- /dev/null +++ b/docs/zh/context-windows.md @@ -0,0 +1,63 @@ +# 上下文窗口对比 + +哪些模型拥有最大的上下文窗口？本页按上下文窗口大小和定价列出模型。 + +> 完整列表请下载 [models.json](https://github.com/i-need-token/ai-models/releases/latest) 或浏览 `providers/`。 + +## 上下文窗口分布 + +| 层级 | 模型数 | 描述 | +| ---------------- | -----: | -------------------------------- | +| 1M+ tokens | 391 | 可处理整本书、整个代码库或长对话 | +| 256K–1M tokens | 459 | 大型文档、多轮对话 | +| 128K–256K tokens | 1,310 | 标准长上下文，大多数现代模型 | +| 32K–128K tokens | 194 | 中等长度文档 | +| 8K–32K tokens | 97 | 短文档、单轮查询 | +| <8K tokens | 19 | 旧模型，极短输入 | + +## 最大上下文窗口（1M+ tokens） + +| 模型 | 提供商 | 上下文 | 输入 $/1M | 输出 $/1M | 工具调用 | 推理 | +| ----------------------------- | ---------- | ------ | --------: | --------: | -------- | ---- | +| Llama 4 Scout | Meta | 10M | — | — | ✅ | ❌ | +| Llama 4 Scout | OpenRouter | 10M | $0.08 | $0.30 | ✅ | ❌ | +| Gemini 3 Pro Preview | Google | 2M | $2.00 | $12.00 | ✅ | ❌ | +| Gemini 3.1 Flash Lite Preview | Google | 2M | $0.25 | $1.50 | ✅ | ❌ | +| Gemini 3.1 Pro Preview | Google | 2M | $2.00 | $12.00 | ✅ | ❌ | +| Grok 4 Fast Reasoning | xAI | 2M | $0.20 | $0.50 | ✅ | ✅ | +| GPT-4.1 | OpenAI | ~1M | $2.00 | $8.00 | ✅ | ❌ | +| Gemini 2.5 Pro | Google | 1M | $1.25 | $10.00 | ✅ | ✅ | +| Gemini 2.5 Flash | Google | 1M | $0.15 | $0.60 | ✅ | ✅ | +| Llama 4 Maverick | Meta | 1M | — | — | ✅ | ❌ | +| Qwen3-235B-A22B | 阿里云 | 128K\* | ¥1.00 | ¥4.00 | ✅ | ✅ | +| DeepSeek-V3 | DeepSeek | 128K | $0.27 | $1.10 | ✅ | ❌ | + +\*注：部分模型在不同平台上的上下文限制不同。请查看特定提供商的 YAML 文件获取准确值。 + +## 各上下文层级的最佳性价比 + +### 1M+ tokens（最便宜） + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | +| ---------------- | ----------- | ---------------: | --------: | +| Llama 4 Scout | OpenRouter | $0.08 | $0.30 | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | +| Llama 4 Scout | Together AI | $0.15 | $0.60 | +| Llama 4 Scout | Meta | 免费（开源权重） | — | + +### 128K–256K tokens（最便宜） + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | +| ------------- | --------- | --------: | --------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-30B-A3B | 阿里云 | ¥0.10 | ¥0.30 | +| Phi-4 | Microsoft | $0.10 | $0.40 | +| Gemma 3 27B | Google | $0.20 | $0.80 | + +## 要点总结 + +- **Llama 4 Scout** 拥有最大的上下文窗口，达 **10M tokens** — 是其他模型的 10 倍 +- **1M+ 上下文**现已在 6+ 个提供商中可用，包括免费开源权重模型 +- **128K 上下文**是最常见的层级（1,310 个模型）— 足以满足大多数用例 +- **成本随上下文增长**：1M+ 上下文模型的每 token 成本是 128K 模型的 2–10 倍 +- **缓存读取定价**可显著降低重复查询的成本（最高 90% 折扣） From 6a8bc77df3215101f9a8f919b22a86f8c7647d36 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:34:53 +0800 Subject: [PATCH 043/311] style: fix release.yml trailing newline --- .github/workflows/release.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index ce4479ae..40c703f3 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -51,4 +51,4 @@ jobs: const catalog = require("./models.json"); console.log(catalog.stats); ``` - generate-release-notes: true \ No newline at end of file + generate-release-notes: true From e2d53389fa1b20b801f257c2ac7190f6085bd7d1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:36:21 +0800 Subject: [PATCH 044/311] docs: add context window comparison link to README documentation table --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index e79895b8..cdce1630 100644 --- a/README.md +++ b/README.md @@ -275,6 +275,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | +| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | @@ -289,6 +290,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | | [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | | [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | | [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | From 30fccecd3c8936cb00baa1b31511227074e3b660 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:43:28 +0800 Subject: [PATCH 045/311] =?UTF-8?q?docs:=20add=20free=20AI=20models=20guid?= =?UTF-8?q?e=20=E2=80=94=2081=20free=20models=20by=20capability=20(EN=20+?= =?UTF-8?q?=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/free-models.md | 109 +++++++++++++++++++++++++++++++++++++++++ docs/zh/free-models.md | 109 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 218 insertions(+) create mode 100644 docs/free-models.md create mode 100644 docs/zh/free-models.md diff --git a/docs/free-models.md b/docs/free-models.md new file mode 100644 index 00000000..da24c017 --- /dev/null +++ b/docs/free-models.md @@ -0,0 +1,109 @@ +# Free AI Models + +81 models in this catalog are free to use. This page lists them by capability so you can find the right free model for your project. + +> All data sourced from first-party APIs and documentation. Free tiers may have rate limits — check the provider's website for details. + +## Quick Stats + +| Capability | Free Models | +| ------------------------- | ----------: | +| Total free models | 81 | +| With tool calling | 45 | +| With reasoning | 11 | +| With vision (image input) | 17 | + +## Free Models with Tool Calling + +These models support function/tool calling at no cost — ideal for building AI agents and automation: + +| Model | Provider | Context | Vision | +| ------------------------------ | -------- | ------- | ------ | +| gemini-2.0-flash | Google | 1M | ✅ | +| gemini-2.5-flash-preview-05-20 | Google | 1M | ✅ | +| gemma-3-27b-it | Chutes | 128K | ✅ | +| qwen3-235b-a22b | Chutes | 128K | ✅ | +| qwen3-30b-a3b | Chutes | 128K | ✅ | +| qwen3-4b | Chutes | 128K | ✅ | +| deepseek-r1 | Chutes | 128K | ❌ | +| deepseek-v3-0324 | Chutes | 128K | ❌ | +| llama-4-maverick | Chutes | 1M | ✅ | +| llama-4-scout | Chutes | 10M | ✅ | +| llama-3.3-70b-instruct | Chutes | 128K | ❌ | +| qwen2.5-72b-instruct | Chutes | 128K | ❌ | +| mistral-small-3.1-24b-instruct | Chutes | 128K | ✅ | +| phi-4 | Chutes | 16K | ❌ | +| command-r | Chutes | 128K | ❌ | + +## Free Models with Reasoning + +These models support chain-of-thought reasoning at no cost: + +| Model | Provider | Context | +| ------------------------------ | -------- | ------- | +| gemini-2.5-flash-preview-05-20 | Google | 1M | +| deepseek-r1 | Chutes | 128K | +| deepseek-r1-0528 | Chutes | 128K | +| qwen3-235b-a22b | Chutes | 128K | +| qwen3-30b-a3b | Chutes | 128K | +| qwen3-4b | Chutes | 128K | +| gemma-3-27b-it | Chutes | 128K | +| phi-4-reasoning | Chutes | 32K | + +## Free Models with Vision + +These models accept image input at no cost: + +| Model | Provider | Context | +| ------------------------------ | -------- | ------- | +| gemini-2.0-flash | Google | 1M | +| gemini-2.5-flash-preview-05-20 | Google | 1M | +| gemma-3-27b-it | Chutes | 128K | +| qwen3-235b-a22b | Chutes | 128K | +| llama-4-maverick | Chutes | 1M | +| llama-4-scout | Chutes | 10M | +| mistral-small-3.1-24b-instruct | Chutes | 128K | + +## Free Models by Provider + +### Google (via AI Studio) + +Google offers free access to Gemini models through AI Studio with rate limits: + +- gemini-2.0-flash — 1M context, tool calling, vision, reasoning +- gemini-2.5-flash-preview-05-20 — 1M context, tool calling, vision, reasoning + +### Chutes + +Chutes provides free community-hosted inference for open-weight models: + +- 70+ free models including Llama 4, Qwen3, DeepSeek-R1, Gemma 3, Mistral, Phi-4 +- Largest free model: Llama 4 Scout (10M context) +- Best free reasoning: DeepSeek-R1, Qwen3-235B-A22B + +### Cloudflare Workers AI + +Cloudflare offers free inference on edge for select models: + +- Various small and medium models with rate limits +- Edge deployment for low latency + +### Cerebras + +Cerebras offers free tier for some models with rate limits: + +- Fast inference using CS-3 wafer-scale engine + +### Groq + +Groq offers free tier for some models with rate limits: + +- Ultra-fast inference using LPU acceleration + +## Key Takeaways + +- **Google AI Studio** offers the best free models overall — 1M context, tool calling, vision, and reasoning +- **Chutes** has the largest selection of free models — 70+ including all major open-weight models +- **Llama 4 Scout** on Chutes offers the largest free context window at 10M tokens +- Free tiers typically have rate limits (requests per minute) — check provider docs for specifics +- For production use, consider upgrading to paid tiers for reliability and higher rate limits diff --git a/docs/zh/free-models.md b/docs/zh/free-models.md new file mode 100644 index 00000000..5253fdc9 --- /dev/null +++ b/docs/zh/free-models.md @@ -0,0 +1,109 @@ +# 免费 AI 模型 + +本目录中有 81 个模型可免费使用。本页按能力分类列出，帮助你找到适合项目的免费模型。 + +> 所有数据来自一手 API 和文档。免费层可能有速率限制 — 请查看提供商网站了解详情。 + +## 快速统计 + +| 能力 | 免费模型数 | +| -------------------- | ---------: | +| 总免费模型 | 81 | +| 支持工具调用 | 45 | +| 支持推理 | 11 | +| 支持视觉（图像输入） | 17 | + +## 支持工具调用的免费模型 + +这些模型支持函数/工具调用且零成本 — 适合构建 AI 代理和自动化： + +| 模型 | 提供商 | 上下文 | 视觉 | +| ------------------------------ | ------ | ------ | ---- | +| gemini-2.0-flash | Google | 1M | ✅ | +| gemini-2.5-flash-preview-05-20 | Google | 1M | ✅ | +| gemma-3-27b-it | Chutes | 128K | ✅ | +| qwen3-235b-a22b | Chutes | 128K | ✅ | +| qwen3-30b-a3b | Chutes | 128K | ✅ | +| qwen3-4b | Chutes | 128K | ✅ | +| deepseek-r1 | Chutes | 128K | ❌ | +| deepseek-v3-0324 | Chutes | 128K | ❌ | +| llama-4-maverick | Chutes | 1M | ✅ | +| llama-4-scout | Chutes | 10M | ✅ | +| llama-3.3-70b-instruct | Chutes | 128K | ❌ | +| qwen2.5-72b-instruct | Chutes | 128K | ❌ | +| mistral-small-3.1-24b-instruct | Chutes | 128K | ✅ | +| phi-4 | Chutes | 16K | ❌ | +| command-r | Chutes | 128K | ❌ | + +## 支持推理的免费模型 + +这些模型支持链式思维推理且零成本： + +| 模型 | 提供商 | 上下文 | +| ------------------------------ | ------ | ------ | +| gemini-2.5-flash-preview-05-20 | Google | 1M | +| deepseek-r1 | Chutes | 128K | +| deepseek-r1-0528 | Chutes | 128K | +| qwen3-235b-a22b | Chutes | 128K | +| qwen3-30b-a3b | Chutes | 128K | +| qwen3-4b | Chutes | 128K | +| gemma-3-27b-it | Chutes | 128K | +| phi-4-reasoning | Chutes | 32K | + +## 支持视觉的免费模型 + +这些模型接受图像输入且零成本： + +| 模型 | 提供商 | 上下文 | +| ------------------------------ | ------ | ------ | +| gemini-2.0-flash | Google | 1M | +| gemini-2.5-flash-preview-05-20 | Google | 1M | +| gemma-3-27b-it | Chutes | 128K | +| qwen3-235b-a22b | Chutes | 128K | +| llama-4-maverick | Chutes | 1M | +| llama-4-scout | Chutes | 10M | +| mistral-small-3.1-24b-instruct | Chutes | 128K | + +## 按提供商分类 + +### Google（通过 AI Studio） + +Google 通过 AI Studio 提供免费 Gemini 模型访问（有速率限制）： + +- gemini-2.0-flash — 1M 上下文，工具调用，视觉，推理 +- gemini-2.5-flash-preview-05-20 — 1M 上下文，工具调用，视觉，推理 + +### Chutes + +Chutes 提供免费社区托管推理，支持开源权重模型： + +- 70+ 个免费模型，包括 Llama 4、Qwen3、DeepSeek-R1、Gemma 3、Mistral、Phi-4 +- 最大免费模型：Llama 4 Scout（10M 上下文） +- 最佳免费推理：DeepSeek-R1、Qwen3-235B-A22B + +### Cloudflare Workers AI + +Cloudflare 提供免费边缘推理： + +- 各种小型和中型模型（有速率限制） +- 边缘部署，低延迟 + +### Cerebras + +Cerebras 为部分模型提供免费层： + +- 使用 CS-3 晶圆级引擎的快速推理 + +### Groq + +Groq 为部分模型提供免费层： + +- 使用 LPU 加速的超快推理 + +## 要点总结 + +- **Google AI Studio** 提供整体最佳的免费模型 — 1M 上下文、工具调用、视觉和推理 +- **Chutes** 拥有最多的免费模型选择 — 70+ 个，包括所有主要开源权重模型 +- **Llama 4 Scout** 在 Chutes 上提供最大的免费上下文窗口（10M tokens） +- 免费层通常有速率限制（每分钟请求数）— 请查看提供商文档了解具体限制 +- 生产环境建议升级到付费层以获得可靠性和更高的速率限制 From fbc9129c537d702e8d7010b804a400b3c2e997cd Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:45:39 +0800 Subject: [PATCH 046/311] docs: add free AI models link to README documentation table --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index cdce1630..7019df78 100644 --- a/README.md +++ b/README.md @@ -276,6 +276,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | +| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | @@ -291,6 +292,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | | [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | | [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | | [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | | [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | From b426726c7c5a83350da4ccc150b2d1ca30cec919 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:48:35 +0800 Subject: [PATCH 047/311] =?UTF-8?q?docs:=20add=20open-weight=20AI=20models?= =?UTF-8?q?=20guide=20=E2=80=94=20513=20models,=20most=20available,=20chea?= =?UTF-8?q?pest=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/open-weights.md | 104 ++++++++++++++++++++++++++++++++++++++++ docs/zh/open-weights.md | 104 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 208 insertions(+) create mode 100644 docs/open-weights.md create mode 100644 docs/zh/open-weights.md diff --git a/docs/open-weights.md b/docs/open-weights.md new file mode 100644 index 00000000..db4f193c --- /dev/null +++ b/docs/open-weights.md @@ -0,0 +1,104 @@ +# Open-Weight AI Models + +513 models in this catalog are open-weight — you can download the weights and run them on your own infrastructure. This page highlights the most capable and widely available open-weight models. + +> All data sourced from first-party APIs and documentation. "Open-weight" means the model weights are publicly available; license terms vary by model. + +## Quick Stats + +| Capability | Open-Weight Models | +| ------------------------- | -----------------: | +| Total open-weight models | 513 | +| Unique model IDs | 420 | +| With tool calling | 270 | +| With reasoning | 101 | +| With vision (image input) | 104 | + +## Most Widely Available + +These open-weight models are available on the most providers — easy to find, easy to switch: + +| Model | Providers | Context | Tool Call | Reasoning | Vision | +| ----------------------------- | --------: | ------- | --------- | --------- | ------ | +| GPT-OSS-120B | 14 | 131K | ✅ | ✅ | ❌ | +| GPT-OSS-20B | 8 | 131K | ✅ | ✅ | ❌ | +| Qwen3.5-397B-A17B | 4 | 262K | ✅ | ✅ | ❌ | +| Kimi K2 Thinking | 4 | 262K | ✅ | ✅ | ✅ | +| DeepSeek-R1-Distill-Llama-70B | 4 | 131K | ✅ | ✅ | ❌ | +| Llama 4 Scout 17B | 4 | 328K | ✅ | ❌ | ✅ | +| DeepSeek-R1 | 3 | 131K | ✅ | ✅ | ❌ | +| Llama 4 Maverick | 3 | 1M | ✅ | ❌ | ✅ | +| Qwen3-32B | 3 | 131K | ✅ | ✅ | ❌ | +| Gemma 4 31B IT | 3 | 262K | ✅ | ✅ | ✅ | + +## Largest Context Windows + +Open-weight models with the largest context windows: + +| Model | Context | Tool Call | Reasoning | Vision | +| ----------------- | ------- | --------- | --------- | ------ | +| Llama 4 Scout | 10M | ✅ | ❌ | ✅ | +| Qwen3.5 Flash | 1M | ✅ | ❌ | ✅ | +| Qwen3.6 Flash | 1M | ✅ | ❌ | ✅ | +| Llama 4 Maverick | 1M | ✅ | ❌ | ✅ | +| DeepSeek-V4 Flash | 1M | ✅ | ✅ | ❌ | +| DeepSeek-V4 Pro | 1M | ✅ | ✅ | ❌ | +| MiMo V2.5 | 1M | ✅ | ✅ | ✅ | +| Minimax M2.5 | 1M | ✅ | ❌ | ❌ | +| Gemma 4 31B IT | 1M | ✅ | ❌ | ✅ | + +## Best Open-Weight Reasoning Models + +These open-weight models support chain-of-thought reasoning: + +| Model | Context | Tool Call | Vision | Providers | +| -------------------- | ------- | --------- | ------ | --------: | +| DeepSeek-V4 Flash | 1M | ✅ | ❌ | 2 | +| DeepSeek-V4 Pro | 1M | ✅ | ❌ | 2 | +| MiMo V2.5 Pro | 1M | ✅ | ❌ | 1 | +| MiMo V2.5 | 1M | ✅ | ✅ | 1 | +| Gemma 4 26B A4B IT | 262K | ✅ | ✅ | 3 | +| Kimi K2.6 | 262K | ✅ | ✅ | 2 | +| Qwen3.5-397B-A17B | 262K | ✅ | ❌ | 2 | +| Nemotron-3-120B-A12B | 262K | ✅ | ❌ | 1 | +| DeepSeek-R1 | 131K | ✅ | ❌ | 3 | +| Qwen3-32B | 131K | ✅ | ✅ | 3 | + +## Best Open-Weight Vision Models + +Open-weight models that accept image input: + +| Model | Context | Tool Call | Reasoning | Providers | +| ------------------ | ------- | --------- | --------- | --------: | +| MiMo V2.5 | 1M | ✅ | ✅ | 1 | +| Llama 4 Maverick | 1M | ✅ | ❌ | 3 | +| Llama 4 Scout | 10M | ✅ | ❌ | 2 | +| Gemma 4 31B IT | 1M | ✅ | ❌ | 3 | +| Qwen3.5 Flash | 1M | ✅ | ❌ | 1 | +| Kimi K2.6 | 262K | ✅ | ✅ | 2 | +| Gemma 4 26B A4B IT | 262K | ✅ | ✅ | 3 | +| Llama 4 Scout 17B | 328K | ✅ | ❌ | 4 | + +## Cheapest Open-Weight Models + +Lowest per-token pricing for open-weight inference: + +| Model | Provider | Input $/1M | Output $/1M | Context | +| -------------------------- | ------------ | ---------: | ----------: | ------- | +| GLM-4-Flash | 302AI | $0.0014 | $0.0014 | 131K | +| Mistral-Nemo-Instruct-2407 | KlusterAI | $0.008 | $0.001 | 131K | +| BDC-Coder | InferenceNet | $0.01 | $0.01 | 131K | +| Granite 4.0 H Micro | Cloudflare | $0.017 | $0.112 | 131K | +| Llama 3.1 8B Instruct | InferenceNet | $0.02 | $0.03 | 131K | +| Mistral Nemo Instruct 2407 | MegaNova | $0.02 | $0.04 | 131K | +| Meta-Llama-3.1-8B-Instruct | Nebius | $0.02 | $0.06 | 131K | +| Llama 3.2 1B Instruct | Cloudflare | $0.027 | $0.201 | 131K | + +## Key Takeaways + +- **513 open-weight models** across 420 unique model IDs — the largest open-weight model catalog available +- **GPT-OSS-120B** is the most widely available, offered by 14 providers +- **Llama 4 Scout** has the largest context window at 10M tokens +- **DeepSeek-R1** is the most popular open-weight reasoning model, available on 3 providers +- **MiMo V2.5** is the only open-weight model combining 1M context, reasoning, and vision +- Pricing varies widely — the cheapest open-weight models cost under $0.01/1M tokens diff --git a/docs/zh/open-weights.md b/docs/zh/open-weights.md new file mode 100644 index 00000000..788f3fcf --- /dev/null +++ b/docs/zh/open-weights.md @@ -0,0 +1,104 @@ +# 开源权重 AI 模型 + +本目录中有 513 个开源权重模型 — 你可以下载权重并在自己的基础设施上运行。本页重点介绍最有能力和最广泛可用的开源权重模型。 + +> 所有数据来自一手 API 和文档。"开源权重"表示模型权重公开可用；各模型的许可证条款不同。 + +## 快速统计 + +| 能力 | 开源权重模型数 | +| -------------------- | -------------: | +| 总开源权重模型 | 513 | +| 唯一模型 ID | 420 | +| 支持工具调用 | 270 | +| 支持推理 | 101 | +| 支持视觉（图像输入） | 104 | + +## 最广泛可用 + +这些开源权重模型在最多提供商上可用 — 容易找到，容易切换： + +| 模型 | 提供商数 | 上下文 | 工具调用 | 推理 | 视觉 | +| ----------------------------- | -------: | ------ | -------- | ---- | ---- | +| GPT-OSS-120B | 14 | 131K | ✅ | ✅ | ❌ | +| GPT-OSS-20B | 8 | 131K | ✅ | ✅ | ❌ | +| Qwen3.5-397B-A17B | 4 | 262K | ✅ | ✅ | ❌ | +| Kimi K2 Thinking | 4 | 262K | ✅ | ✅ | ✅ | +| DeepSeek-R1-Distill-Llama-70B | 4 | 131K | ✅ | ✅ | ❌ | +| Llama 4 Scout 17B | 4 | 328K | ✅ | ❌ | ✅ | +| DeepSeek-R1 | 3 | 131K | ✅ | ✅ | ❌ | +| Llama 4 Maverick | 3 | 1M | ✅ | ❌ | ✅ | +| Qwen3-32B | 3 | 131K | ✅ | ✅ | ❌ | +| Gemma 4 31B IT | 3 | 262K | ✅ | ✅ | ✅ | + +## 最大上下文窗口 + +拥有最大上下文窗口的开源权重模型： + +| 模型 | 上下文 | 工具调用 | 推理 | 视觉 | +| ----------------- | ------ | -------- | ---- | ---- | +| Llama 4 Scout | 10M | ✅ | ❌ | ✅ | +| Qwen3.5 Flash | 1M | ✅ | ❌ | ✅ | +| Qwen3.6 Flash | 1M | ✅ | ❌ | ✅ | +| Llama 4 Maverick | 1M | ✅ | ❌ | ✅ | +| DeepSeek-V4 Flash | 1M | ✅ | ✅ | ❌ | +| DeepSeek-V4 Pro | 1M | ✅ | ✅ | ❌ | +| MiMo V2.5 | 1M | ✅ | ✅ | ✅ | +| Minimax M2.5 | 1M | ✅ | ❌ | ❌ | +| Gemma 4 31B IT | 1M | ✅ | ❌ | ✅ | + +## 最佳开源权重推理模型 + +这些开源权重模型支持链式思维推理： + +| 模型 | 上下文 | 工具调用 | 视觉 | 提供商数 | +| -------------------- | ------ | -------- | ---- | -------: | +| DeepSeek-V4 Flash | 1M | ✅ | ❌ | 2 | +| DeepSeek-V4 Pro | 1M | ✅ | ❌ | 2 | +| MiMo V2.5 Pro | 1M | ✅ | ❌ | 1 | +| MiMo V2.5 | 1M | ✅ | ✅ | 1 | +| Gemma 4 26B A4B IT | 262K | ✅ | ✅ | 3 | +| Kimi K2.6 | 262K | ✅ | ✅ | 2 | +| Qwen3.5-397B-A17B | 262K | ✅ | ❌ | 2 | +| Nemotron-3-120B-A12B | 262K | ✅ | ❌ | 1 | +| DeepSeek-R1 | 131K | ✅ | ❌ | 3 | +| Qwen3-32B | 131K | ✅ | ✅ | 3 | + +## 最佳开源权重视觉模型 + +接受图像输入的开源权重模型： + +| 模型 | 上下文 | 工具调用 | 推理 | 提供商数 | +| ------------------ | ------ | -------- | ---- | -------: | +| MiMo V2.5 | 1M | ✅ | ✅ | 1 | +| Llama 4 Maverick | 1M | ✅ | ❌ | 3 | +| Llama 4 Scout | 10M | ✅ | ❌ | 2 | +| Gemma 4 31B IT | 1M | ✅ | ❌ | 3 | +| Qwen3.5 Flash | 1M | ✅ | ❌ | 1 | +| Kimi K2.6 | 262K | ✅ | ✅ | 2 | +| Gemma 4 26B A4B IT | 262K | ✅ | ✅ | 3 | +| Llama 4 Scout 17B | 328K | ✅ | ❌ | 4 | + +## 最便宜的开源权重模型 + +开源权重推理的最低每 token 定价： + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | 上下文 | +| -------------------------- | ------------ | --------: | --------: | ------ | +| GLM-4-Flash | 302AI | $0.0014 | $0.0014 | 131K | +| Mistral-Nemo-Instruct-2407 | KlusterAI | $0.008 | $0.001 | 131K | +| BDC-Coder | InferenceNet | $0.01 | $0.01 | 131K | +| Granite 4.0 H Micro | Cloudflare | $0.017 | $0.112 | 131K | +| Llama 3.1 8B Instruct | InferenceNet | $0.02 | $0.03 | 131K | +| Mistral Nemo Instruct 2407 | MegaNova | $0.02 | $0.04 | 131K | +| Meta-Llama-3.1-8B-Instruct | Nebius | $0.02 | $0.06 | 131K | +| Llama 3.2 1B Instruct | Cloudflare | $0.027 | $0.201 | 131K | + +## 要点总结 + +- **513 个开源权重模型**，涵盖 420 个唯一模型 ID — 最大的开源权重模型目录 +- **GPT-OSS-120B** 是最广泛可用的，由 14 个提供商提供 +- **Llama 4 Scout** 拥有最大的上下文窗口（10M tokens） +- **DeepSeek-R1** 是最受欢迎的开源权重推理模型，在 3 个提供商上可用 +- **MiMo V2.5** 是唯一结合 1M 上下文、推理和视觉的开源权重模型 +- 定价差异很大 — 最便宜的开源权重模型每 1M tokens 不到 $0.01 From d09b34545871061da845dc21b69fd116c8f1dd5e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 10:55:04 +0800 Subject: [PATCH 048/311] docs: add open-weight models link to README documentation table --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index 7019df78..1aad4ab5 100644 --- a/README.md +++ b/README.md @@ -277,6 +277,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | | [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | +| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | @@ -293,6 +294,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | | [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | | [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | | [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | | [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | From be221b22fbae32a211c66e2986bef5485a689f38 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:01:29 +0800 Subject: [PATCH 049/311] =?UTF-8?q?docs:=20add=20reasoning=20models=20guid?= =?UTF-8?q?e=20=E2=80=94=201,306=20models,=20top=20by=20context,=20cheapes?= =?UTF-8?q?t=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/reasoning-models.md | 87 +++++++++++++++++++++++++++++++++++++ docs/zh/reasoning-models.md | 87 +++++++++++++++++++++++++++++++++++++ 2 files changed, 174 insertions(+) create mode 100644 docs/reasoning-models.md create mode 100644 docs/zh/reasoning-models.md diff --git a/docs/reasoning-models.md b/docs/reasoning-models.md new file mode 100644 index 00000000..32380292 --- /dev/null +++ b/docs/reasoning-models.md @@ -0,0 +1,87 @@ +# AI Reasoning Models + +1,306 models in this catalog support reasoning (chain-of-thought / extended thinking). This page highlights the most capable and cost-effective reasoning models available. + +> All data sourced from first-party APIs and documentation. "Reasoning" means the model can produce extended chain-of-thought before answering. + +## Quick Stats + +| Capability | Reasoning Models | +| ------------------------- | ---------------: | +| Total reasoning models | 1,306 | +| Unique model IDs | 868 | +| With tool calling | 1,076 | +| With vision (image input) | 697 | +| Open-weight | 119 | + +## Top Reasoning Models by Context + +The largest-context reasoning models — ideal for complex, multi-step tasks: + +| Model | Context | Tool Call | Vision | Input $/1M | Providers | +| ----------------------------- | ------- | --------- | ------ | ---------: | --------: | +| Grok 4 Fast Reasoning | 2M | ✅ | ✅ | $0.20 | 2 | +| Grok 4.1 Fast Reasoning | 2M | ✅ | ✅ | $0.20 | 2 | +| Grok 4.20 | 2M | ✅ | ✅ | $1.42 | 1 | +| GPT-5.4 | 1M | ✅ | ✅ | $2.50 | 4 | +| GPT-5.5 | 1M | ✅ | ✅ | $5.00 | 4 | +| Gemini 2.5 Pro | 1M | ✅ | ✅ | $1.25 | 4 | +| Gemini 2.5 Flash | 1M | ✅ | ✅ | $0.15 | 3 | +| Gemini 2.5 Flash Lite | 1M | ✅ | ✅ | $0.10 | 3 | +| Gemini 3 Flash Preview | 1M | ✅ | ✅ | $0.50 | 4 | +| Gemini 3.1 Flash Lite Preview | 1M | ✅ | ✅ | $0.25 | 3 | +| DeepSeek Reasoner | 1M | ✅ | ✅ | $0.43 | 1 | + +## Cheapest Reasoning Models + +Best value for reasoning capability: + +| Model | Provider | Input $/1M | Output $/1M | Context | +| --------------------- | --------- | ---------: | ----------: | ------- | +| Qwen 3.5 0.8B | Auriko | $0.01 | $0.05 | 262K | +| Qwen 3.5 0.8B | DeepInfra | $0.01 | $0.05 | 262K | +| Qwen 3.5 2B | Auriko | $0.02 | $0.10 | 262K | +| Qwen 3.5 2B | DeepInfra | $0.02 | $0.10 | 262K | +| GPT-5 Nano | Requesty | $0.025 | $0.20 | 400K | +| Qwen 3.5 4B | Auriko | $0.03 | $0.15 | 262K | +| Qwen 3.5 4B | DeepInfra | $0.03 | $0.15 | 262K | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | 1M | +| Grok 4 Fast Reasoning | xAI | $0.20 | $0.50 | 2M | + +## Best Reasoning + Vision Models + +Models that can reason about images — ideal for visual analysis: + +| Model | Context | Input $/1M | Providers | +| ----------------------- | ------- | ---------: | --------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | 4 | +| GPT-5.4 | 1M | $2.50 | 4 | +| DeepSeek Reasoner | 1M | $0.43 | 1 | +| MiMo V2.5 (open-weight) | 1M | varies | 2 | + +## Open-Weight Reasoning Models + +119 open-weight models support reasoning — run them on your own hardware: + +| Model | Context | Tool Call | Vision | Providers | +| ----------------------- | ------- | --------- | ------ | --------: | +| MiMo V2.5 Pro | 1M | ✅ | ❌ | 2 | +| MiMo V2.5 | 1M | ✅ | ✅ | 2 | +| DeepSeek-V4 Pro | 1M | ✅ | ❌ | 1 | +| Qwen3 Next 80B Thinking | 262K | ✅ | ❌ | 4 | +| Kimi K2.6 | 262K | ✅ | ✅ | 4 | +| Trinity Large Thinking | 262K | ✅ | ❌ | 1 | +| Nemotron 3 120B | 262K | ✅ | ❌ | 1 | +| Qwen3.5 397B A17B | 262K | ✅ | ❌ | 2 | + +## Key Takeaways + +- **1,306 reasoning models** across 868 unique IDs — the largest reasoning model catalog available +- **Grok 4 Fast Reasoning** offers the best value at 2M context for $0.20/1M input +- **Gemini 2.5 Flash Lite** is the cheapest 1M-context reasoning model at $0.10/1M +- **MiMo V2.5** is the only open-weight model combining 1M context, reasoning, and vision +- **697 reasoning models** also support vision — the most common combined capability +- Small reasoning models (Qwen 3.5 0.8B–4B) cost as little as $0.01–$0.03/1M tokens diff --git a/docs/zh/reasoning-models.md b/docs/zh/reasoning-models.md new file mode 100644 index 00000000..0aba021d --- /dev/null +++ b/docs/zh/reasoning-models.md @@ -0,0 +1,87 @@ +# AI 推理模型 + +本目录中有 1,306 个模型支持推理（链式思维 / 扩展思考）。本页重点介绍最有能力和最具性价比的推理模型。 + +> 所有数据来自一手 API 和文档。"推理"表示模型可以在回答前生成扩展的链式思维。 + +## 快速统计 + +| 能力 | 推理模型数 | +| -------------------- | ---------: | +| 总推理模型 | 1,306 | +| 唯一模型 ID | 868 | +| 支持工具调用 | 1,076 | +| 支持视觉（图像输入） | 697 | +| 开源权重 | 119 | + +## 按上下文排序的顶级推理模型 + +最大上下文的推理模型 — 适合复杂多步任务： + +| 模型 | 上下文 | 工具调用 | 视觉 | 输入 $/1M | 提供商数 | +| ----------------------------- | ------ | -------- | ---- | --------: | -------: | +| Grok 4 Fast Reasoning | 2M | ✅ | ✅ | $0.20 | 2 | +| Grok 4.1 Fast Reasoning | 2M | ✅ | ✅ | $0.20 | 2 | +| Grok 4.20 | 2M | ✅ | ✅ | $1.42 | 1 | +| GPT-5.4 | 1M | ✅ | ✅ | $2.50 | 4 | +| GPT-5.5 | 1M | ✅ | ✅ | $5.00 | 4 | +| Gemini 2.5 Pro | 1M | ✅ | ✅ | $1.25 | 4 | +| Gemini 2.5 Flash | 1M | ✅ | ✅ | $0.15 | 3 | +| Gemini 2.5 Flash Lite | 1M | ✅ | ✅ | $0.10 | 3 | +| Gemini 3 Flash Preview | 1M | ✅ | ✅ | $0.50 | 4 | +| Gemini 3.1 Flash Lite Preview | 1M | ✅ | ✅ | $0.25 | 3 | +| DeepSeek Reasoner | 1M | ✅ | ✅ | $0.43 | 1 | + +## 最便宜的推理模型 + +推理能力的最佳性价比： + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | 上下文 | +| --------------------- | --------- | --------: | --------: | ------ | +| Qwen 3.5 0.8B | Auriko | $0.01 | $0.05 | 262K | +| Qwen 3.5 0.8B | DeepInfra | $0.01 | $0.05 | 262K | +| Qwen 3.5 2B | Auriko | $0.02 | $0.10 | 262K | +| Qwen 3.5 2B | DeepInfra | $0.02 | $0.10 | 262K | +| GPT-5 Nano | Requesty | $0.025 | $0.20 | 400K | +| Qwen 3.5 4B | Auriko | $0.03 | $0.15 | 262K | +| Qwen 3.5 4B | DeepInfra | $0.03 | $0.15 | 262K | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | 1M | +| Grok 4 Fast Reasoning | xAI | $0.20 | $0.50 | 2M | + +## 最佳推理 + 视觉模型 + +能对图像进行推理的模型 — 适合视觉分析： + +| 模型 | 上下文 | 输入 $/1M | 提供商数 | +| --------------------- | ------ | --------: | -------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | 4 | +| GPT-5.4 | 1M | $2.50 | 4 | +| DeepSeek Reasoner | 1M | $0.43 | 1 | +| MiMo V2.5（开源权重） | 1M | varies | 2 | + +## 开源权重推理模型 + +119 个开源权重模型支持推理 — 可在自己的硬件上运行： + +| 模型 | 上下文 | 工具调用 | 视觉 | 提供商数 | +| ----------------------- | ------ | -------- | ---- | -------: | +| MiMo V2.5 Pro | 1M | ✅ | ❌ | 2 | +| MiMo V2.5 | 1M | ✅ | ✅ | 2 | +| DeepSeek-V4 Pro | 1M | ✅ | ❌ | 1 | +| Qwen3 Next 80B Thinking | 262K | ✅ | ❌ | 4 | +| Kimi K2.6 | 262K | ✅ | ✅ | 4 | +| Trinity Large Thinking | 262K | ✅ | ❌ | 1 | +| Nemotron 3 120B | 262K | ✅ | ❌ | 1 | +| Qwen3.5 397B A17B | 262K | ✅ | ❌ | 2 | + +## 要点总结 + +- **1,306 个推理模型**，涵盖 868 个唯一模型 ID — 最大的推理模型目录 +- **Grok 4 Fast Reasoning** 以 2M 上下文和 $0.20/1M 输入价格提供最佳性价比 +- **Gemini 2.5 Flash Lite** 是最便宜的 1M 上下文推理模型（$0.10/1M） +- **MiMo V2.5** 是唯一结合 1M 上下文、推理和视觉的开源权重模型 +- **697 个推理模型**同时支持视觉 — 最常见的组合能力 +- 小型推理模型（Qwen 3.5 0.8B–4B）每 1M tokens 仅需 $0.01–$0.03 From c8d18b2894c9dffa3fdfe6175707761d9a777c4f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:02:27 +0800 Subject: [PATCH 050/311] docs: add reasoning models link to README documentation table --- README.md | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) diff --git a/README.md b/README.md index 1aad4ab5..bea4087f 100644 --- a/README.md +++ b/README.md @@ -268,20 +268,21 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation -| Document | Description | -| ------------------------------------------------------- | -------------------------------------------------------------- | -| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | -| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | -| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | -| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | -| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | -| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | -| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | -| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | -| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | -| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | -| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | -| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | +| Document | Description | +| ------------------------------------------------------- | --------------------------------------------------------------- | +| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | +| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | +| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | +| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | +| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | +| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | +| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | +| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | +| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | +| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | +| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | **中文文档：** @@ -296,6 +297,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | | [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | | [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | | [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | | [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | | [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | From aa9d4f63b903d47329ad151c893b508564f58260 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:04:28 +0800 Subject: [PATCH 051/311] =?UTF-8?q?feat:=20add=20npm=20package=20setup=20?= =?UTF-8?q?=E2=80=94=20types,=20exports,=20build=20script,=20files=20white?= =?UTF-8?q?list?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- package.json | 19 ++++++++++++++++++- tsconfig.build.json | 21 +++++++++++++++++++++ 2 files changed, 39 insertions(+), 1 deletion(-) create mode 100644 tsconfig.build.json diff --git a/package.json b/package.json index 6d8befe3..9bd0823b 100644 --- a/package.json +++ b/package.json @@ -32,7 +32,22 @@ "type": "git", "url": "git+https://github.com/i-need-token/ai-models.git" }, + "files": [ + "dist/", + "types/", + "providers/", + "LICENSE" + ], "type": "module", + "main": "dist/models.json", + "types": "dist/index.d.ts", + "exports": { + ".": { + "types": "./dist/index.d.ts", + "default": "./dist/models.json" + }, + "./models.json": "./dist/models.json" + }, "scripts": { "sync": "npx tsx scripts/sync.ts", "validate": "npx tsx scripts/validate.ts", @@ -43,7 +58,9 @@ "lint": "oxlint", "typecheck": "tsc --noEmit", "check": "npm run fmt:check && npm run lint && npm run typecheck", - "prepare": "husky" + "prepare": "husky", + "build": "tsc -p tsconfig.build.json && npx tsx scripts/compile.ts", + "prepublishOnly": "npm run build" }, "dependencies": { "linkedom": "^0.18.12", diff --git a/tsconfig.build.json b/tsconfig.build.json new file mode 100644 index 00000000..f06717bb --- /dev/null +++ b/tsconfig.build.json @@ -0,0 +1,21 @@ +{ + "compilerOptions": { + "target": "ES2022", + "module": "ES2022", + "moduleResolution": "bundler", + "strict": true, + "noUncheckedIndexedAccess": true, + "noUnusedLocals": true, + "noUnusedParameters": true, + "noFallthroughCasesInSwitch": true, + "noPropertyAccessFromIndexSignature": true, + "exactOptionalPropertyTypes": true, + "esModuleInterop": true, + "skipLibCheck": true, + "declaration": true, + "emitDeclarationOnly": true, + "outDir": "dist", + "rootDir": "types" + }, + "include": ["types"] +} From 00c466a30ea0c979493434a3386c6ce2ae264310 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:11:15 +0800 Subject: [PATCH 052/311] =?UTF-8?q?ci:=20add=20npm=20publish=20workflow=20?= =?UTF-8?q?=E2=80=94=20auto-publish=20on=20GitHub=20Release?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/workflows/publish.yml | 27 +++++++++++++++++++++++++++ 1 file changed, 27 insertions(+) create mode 100644 .github/workflows/publish.yml diff --git a/.github/workflows/publish.yml b/.github/workflows/publish.yml new file mode 100644 index 00000000..fdbb2089 --- /dev/null +++ b/.github/workflows/publish.yml @@ -0,0 +1,27 @@ +name: Publish to npm + +on: + release: + types: [published] + +permissions: + contents: read + +jobs: + publish: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + registry-url: https://registry.npmjs.org/ + + - run: npm ci + + - run: npm run build + + - run: npm publish --access public + env: + NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }} From 23a04d347844f4df930f47a5aeac533d4a5cb64c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:15:52 +0800 Subject: [PATCH 053/311] docs: add npm install instructions to README quick start --- README.md | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/README.md b/README.md index bea4087f..5eeeb3c9 100644 --- a/README.md +++ b/README.md @@ -185,6 +185,17 @@ last_updated: "2026-05-18" No installation needed — just browse `providers//models/` for YAML files. Every file is human-readable. +### Install from npm + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types +``` + ### Install & Sync ```bash From cf333601b167fb42dcc9c3abcae6ba4f77548a80 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:17:31 +0800 Subject: [PATCH 054/311] docs: update llms.txt with npm install, all 13 documentation pages --- llms.txt | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) diff --git a/llms.txt b/llms.txt index f17b1c4b..fed97096 100644 --- a/llms.txt +++ b/llms.txt @@ -18,6 +18,17 @@ A machine-readable YAML catalog of every major AI model provider and their model - 81 free models - 1,487 vision models +## Install + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types +``` + ## Data Format Each model is a single YAML file at `providers//models/.yaml`: @@ -42,11 +53,20 @@ modalities: ## Documentation +- [Quick Start](docs/quick-start.md) — find the right model in 30 seconds +- [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers -- [Provider Overview](docs/providers.md) — all providers by type and market +- [Context Window Comparison](docs/context-windows.md) — largest context windows, best value +- [Free AI Models](docs/free-models.md) — 81 free models by capability +- [Open-Weight Models](docs/open-weights.md) — 513 open-weight models +- [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models +- [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video +- [Provider Overview](docs/providers.md) — all 95 providers by type and market - [Data Schema Reference](docs/data-schema.md) — complete YAML schema +- [Data Acquisition](docs/data-acquisition.md) — how we acquire and update data +- [Design Principles](docs/lessons-learned.md) — lessons learned ## Covered Providers -OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, Amazon Bedrock, Azure OpenAI, Google Vertex AI, OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, and 75+ more. +OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, Amazon Bedrock, Azure OpenAI, Google Vertex AI, OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, and 75+ more. \ No newline at end of file From aeead4d6bcc04be851373bb4dfa4f3c1dc378a1d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:23:24 +0800 Subject: [PATCH 055/311] docs: update AGENTS.md with context-windows, free-models, open-weights, reasoning-models --- AGENTS.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index e9d44b74..5e925b35 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -10,6 +10,10 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`types/schemas.ts`](types/schemas.ts) — Zod runtime validation schemas - [`docs/data-acquisition.md`](docs/data-acquisition.md) — How we acquire and update model data ([中文](docs/zh/data-acquisition.md)) - [`docs/lessons-learned.md`](docs/lessons-learned.md) — Design principles and pitfalls ([中文](docs/zh/lessons-learned.md)) +- [`docs/context-windows.md`](docs/context-windows.md) — Context window comparison by size and pricing ([中文](docs/zh/context-windows.md)) +- [`docs/free-models.md`](docs/free-models.md) — 81 free AI models by capability ([中文](docs/zh/free-models.md)) +- [`docs/open-weights.md`](docs/open-weights.md) — 513 open-weight models ([中文](docs/zh/open-weights.md)) +- [`docs/reasoning-models.md`](docs/reasoning-models.md) — 1,306 reasoning models ([中文](docs/zh/reasoning-models.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) - [`docs/pricing-comparison.md`](docs/pricing-comparison.md) — Pricing comparison across providers ([中文](docs/zh/pricing-comparison.md)) From 93029e28babcb7ede97f504246a0847b4a9114af Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:25:10 +0800 Subject: [PATCH 056/311] docs: add npm package section to API docs (EN + ZH) --- docs/api.md | 22 ++++++++++++++++++++++ docs/zh/api.md | 22 ++++++++++++++++++++++ 2 files changed, 44 insertions(+) diff --git a/docs/api.md b/docs/api.md index a8bbecba..4b2388ca 100644 --- a/docs/api.md +++ b/docs/api.md @@ -2,6 +2,28 @@ Use the catalog data in your applications. +## npm Package + +Install the catalog as an npm dependency: + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types + +// Find models with tool calling under $1/1M input +const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < 1); +``` + +The package includes: + +- `dist/models.json` — full catalog as JSON +- `dist/index.d.ts` — TypeScript type definitions +- `types/` — source type definitions (Model, Snapshot, Provider, Pricing) + ## Compiled JSON The easiest way to access all model data is via the compiled `models.json` file. diff --git a/docs/zh/api.md b/docs/zh/api.md index 6bcf17ae..96dbc198 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -2,6 +2,28 @@ 在你的应用中使用目录数据。 +## npm 包 + +将目录作为 npm 依赖安装： + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 个模型 JSON +import type { Model } from "ai-models"; // TypeScript 类型 + +// 查找支持工具调用且输入价格低于 $1/1M 的模型 +const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < 1); +``` + +包内容包括： + +- `dist/models.json` — 完整目录 JSON +- `dist/index.d.ts` — TypeScript 类型定义 +- `types/` — 源类型定义（Model、Snapshot、Provider、Pricing） + ## 编译 JSON 访问所有模型数据最简单的方式是通过编译后的 `models.json` 文件。 From 9f723b0af435a321b96cd5ac31e28b6e92b2dde1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:33:39 +0800 Subject: [PATCH 057/311] fix: remove unused imports in social-preview.ts, add tool-calling link to README --- .github/discussions-setup.md | 42 ++++++++++++++++++ README.md | 58 ++++++++++++------------ docs/tool-calling.md | 86 ++++++++++++++++++++++++++++++++++++ docs/zh/tool-calling.md | 86 ++++++++++++++++++++++++++++++++++++ scripts/social-preview.ts | 77 ++++++++++++++++++++++++++++++++ social-preview.svg | 48 ++++++++++++++++++++ 6 files changed, 369 insertions(+), 28 deletions(-) create mode 100644 .github/discussions-setup.md create mode 100644 docs/tool-calling.md create mode 100644 docs/zh/tool-calling.md create mode 100644 scripts/social-preview.ts create mode 100644 social-preview.svg diff --git a/.github/discussions-setup.md b/.github/discussions-setup.md new file mode 100644 index 00000000..2da1cdb8 --- /dev/null +++ b/.github/discussions-setup.md @@ -0,0 +1,42 @@ +# GitHub Discussions Setup Guide + +Enable GitHub Discussions to create a community hub around the AI Models Catalog. + +## Steps + +1. Go to **Settings → Features** in the repository +2. Check ✅ **Discussions** +3. Create the following discussion categories: + +### Recommended Categories + +| Category | Format | Description | +| ---------------- | ------------ | --------------------------------------------- | +| 📢 Announcements | Announcement | New providers, data updates, breaking changes | +| 💬 General | Open-ended | Questions, ideas, show-and-tell | +| 🙏 Q&A | Q&A | How to use the catalog, data questions | +| 💡 Ideas | Open-ended | Feature requests, new docs suggestions | +| 🏷️ Show and Tell | Open-ended | Projects built with the catalog | + +### First Discussion Posts + +Create these seed discussions to set the tone: + +1. **Welcome to AI Models Catalog** (Announcement) + - Introduce the project, link to quick-start, invite contributions + +2. **What are you building with the catalog?** (Show and Tell) + - Encourage users to share their projects + +3. **Which provider should we add next?** (Ideas) + - Crowdsource new provider requests + +4. **Data quality report — how to report issues** (Q&A) + - Explain how to report stale or incorrect data + +## Benefits + +- **Reduces issue noise** — questions move to Discussions instead of Issues +- **Builds community** — users help each other, share projects +- **SEO boost** — public discussions are indexed by search engines +- **Feedback loop** — learn what users need most diff --git a/README.md b/README.md index 5eeeb3c9..e3ad4ea2 100644 --- a/README.md +++ b/README.md @@ -279,39 +279,41 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation -| Document | Description | +| Document | Description | +| [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | | ------------------------------------------------------- | --------------------------------------------------------------- | -| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | -| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | -| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | -| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | -| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | -| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | -| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | -| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | -| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | -| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | -| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | -| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | -| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | +| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | +| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | +| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | +| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | +| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | +| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | +| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | +| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | +| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | +| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | +| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | **中文文档：** -| 文档 | 描述 | +| 文档 | 描述 | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | | -------------------------------------------- | ------------------------------------------- | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles diff --git a/docs/tool-calling.md b/docs/tool-calling.md new file mode 100644 index 00000000..16a376cc --- /dev/null +++ b/docs/tool-calling.md @@ -0,0 +1,86 @@ +# AI Tool Calling Models + +2,350 models in this catalog support tool calling (function calling). This page highlights the most capable and cost-effective models for building AI agents and automation. + +> All data sourced from first-party APIs and documentation. "Tool calling" means the model can invoke external functions/tools as part of its response. + +## Quick Stats + +| Capability | Tool-Calling Models | +| ------------------------- | ------------------: | +| Total tool-calling models | 2,350 | +| Unique model IDs | 1,540 | +| With reasoning | 1,076 | +| With vision (image input) | 1,063 | +| With structured output | 829 | +| Open-weight | 270 | + +## Cheapest Tool-Calling Models + +Best value for building AI agents: + +| Model | Provider | Input $/1M | Output $/1M | Context | Reasoning | +| -------------------------- | ------------ | ---------: | ----------: | ------- | --------- | +| GLM-4-Flash | 302AI | $0.0014 | $0.0014 | 131K | ❌ | +| Mistral-Nemo-Instruct-2407 | KlusterAI | $0.008 | $0.001 | 131K | ❌ | +| BDC-Coder | InferenceNet | $0.01 | $0.01 | 131K | ❌ | +| Qwen 3.5 0.8B | Auriko | $0.01 | $0.05 | 262K | ✅ | +| Qwen 3.5 0.8B | DeepInfra | $0.01 | $0.05 | 262K | ✅ | +| Qwen 3.5 2B | Auriko | $0.02 | $0.10 | 262K | ✅ | +| Qwen 3.5 2B | DeepInfra | $0.02 | $0.10 | 262K | ✅ | +| GPT-5 Nano | Requesty | $0.025 | $0.20 | 400K | ✅ | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | ✅ | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | 1M | ✅ | +| Grok 4 Fast Reasoning | xAI | $0.20 | $0.50 | 2M | ✅ | + +## Largest Context Tool-Calling Models + +For agents that need to process large documents or long conversation histories: + +| Model | Context | Input $/1M | Reasoning | Providers | +| --------------------- | ------- | ---------: | --------- | --------: | +| Llama 4 Scout | 10M | $0.08 | ❌ | 4 | +| Grok 4 Fast Reasoning | 2M | $0.20 | ✅ | 2 | +| GPT-5.4 | 1M | $2.50 | ✅ | 4 | +| Gemini 2.5 Pro | 1M | $1.25 | ✅ | 4 | +| Gemini 2.5 Flash | 1M | $0.15 | ✅ | 3 | +| DeepSeek-V4 Flash | 1M | $0.27 | ✅ | 2 | +| GPT-4.1 | 1M | $2.00 | ❌ | 4 | +| Llama 4 Maverick | 1M | $0.15 | ❌ | 3 | + +## Best Tool-Calling + Reasoning + Vision + +The "holy trinity" for advanced AI agents — tool calling, reasoning, and vision in one model: + +| Model | Context | Input $/1M | Providers | +| ----------------------- | ------- | ---------: | --------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | 4 | +| GPT-5.4 | 1M | $2.50 | 4 | +| DeepSeek Reasoner | 1M | $0.43 | 1 | +| MiMo V2.5 (open-weight) | 1M | varies | 2 | +| Kimi K2.6 (open-weight) | 262K | varies | 4 | + +## Free Tool-Calling Models + +45 free models support tool calling — ideal for prototyping and testing: + +| Model | Provider | Context | Reasoning | Vision | +| ------------------------------ | -------- | ------- | --------- | ------ | +| gemini-2.0-flash | Google | 1M | ✅ | ✅ | +| gemini-2.5-flash-preview-05-20 | Google | 1M | ✅ | ✅ | +| llama-4-scout | Chutes | 10M | ❌ | ✅ | +| llama-4-maverick | Chutes | 1M | ❌ | ✅ | +| deepseek-r1 | Chutes | 128K | ✅ | ❌ | +| qwen3-235b-a22b | Chutes | 128K | ✅ | ✅ | +| gemma-3-27b-it | Chutes | 128K | ✅ | ✅ | + +## Key Takeaways + +- **2,350 tool-calling models** across 1,540 unique IDs — the largest tool-calling model catalog +- **Gemini 2.5 Flash** is the best value: 1M context, tool calling, reasoning, and vision for $0.15/1M +- **Grok 4 Fast Reasoning** offers the largest context (2M) with all three capabilities +- **45 free models** support tool calling — start building agents at zero cost +- **829 models** also support structured output — perfect for reliable JSON responses +- Small models (Qwen 3.5 0.8B–4B) cost as little as $0.01–$0.03/1M tokens with tool calling diff --git a/docs/zh/tool-calling.md b/docs/zh/tool-calling.md new file mode 100644 index 00000000..8abd7e78 --- /dev/null +++ b/docs/zh/tool-calling.md @@ -0,0 +1,86 @@ +# AI 工具调用模型 + +本目录中有 2,350 个模型支持工具调用（函数调用）。本页重点介绍构建 AI 代理和自动化最有能力和最具性价比的模型。 + +> 所有数据来自一手 API 和文档。"工具调用"表示模型可以在回复中调用外部函数/工具。 + +## 快速统计 + +| 能力 | 工具调用模型数 | +| -------------------- | -------------: | +| 总工具调用模型 | 2,350 | +| 唯一模型 ID | 1,540 | +| 支持推理 | 1,076 | +| 支持视觉（图像输入） | 1,063 | +| 支持结构化输出 | 829 | +| 开源权重 | 270 | + +## 最便宜的工具调用模型 + +构建 AI 代理的最佳性价比： + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | 上下文 | 推理 | +| -------------------------- | ------------ | --------: | --------: | ------ | ---- | +| GLM-4-Flash | 302AI | $0.0014 | $0.0014 | 131K | ❌ | +| Mistral-Nemo-Instruct-2407 | KlusterAI | $0.008 | $0.001 | 131K | ❌ | +| BDC-Coder | InferenceNet | $0.01 | $0.01 | 131K | ❌ | +| Qwen 3.5 0.8B | Auriko | $0.01 | $0.05 | 262K | ✅ | +| Qwen 3.5 0.8B | DeepInfra | $0.01 | $0.05 | 262K | ✅ | +| Qwen 3.5 2B | Auriko | $0.02 | $0.10 | 262K | ✅ | +| Qwen 3.5 2B | DeepInfra | $0.02 | $0.10 | 262K | ✅ | +| GPT-5 Nano | Requesty | $0.025 | $0.20 | 400K | ✅ | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | ✅ | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | 1M | ✅ | +| Grok 4 Fast Reasoning | xAI | $0.20 | $0.50 | 2M | ✅ | + +## 最大上下文工具调用模型 + +适合处理大型文档或长对话历史的代理： + +| 模型 | 上下文 | 输入 $/1M | 推理 | 提供商数 | +| --------------------- | ------ | --------: | ---- | -------: | +| Llama 4 Scout | 10M | $0.08 | ❌ | 4 | +| Grok 4 Fast Reasoning | 2M | $0.20 | ✅ | 2 | +| GPT-5.4 | 1M | $2.50 | ✅ | 4 | +| Gemini 2.5 Pro | 1M | $1.25 | ✅ | 4 | +| Gemini 2.5 Flash | 1M | $0.15 | ✅ | 3 | +| DeepSeek-V4 Flash | 1M | $0.27 | ✅ | 2 | +| GPT-4.1 | 1M | $2.00 | ❌ | 4 | +| Llama 4 Maverick | 1M | $0.15 | ❌ | 3 | + +## 最佳工具调用 + 推理 + 视觉 + +高级 AI 代理的"三位一体" — 工具调用、推理和视觉一体： + +| 模型 | 上下文 | 输入 $/1M | 提供商数 | +| --------------------- | ------ | --------: | -------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | 4 | +| GPT-5.4 | 1M | $2.50 | 4 | +| DeepSeek Reasoner | 1M | $0.43 | 1 | +| MiMo V2.5（开源权重） | 1M | varies | 2 | +| Kimi K2.6（开源权重） | 262K | varies | 4 | + +## 免费工具调用模型 + +45 个免费模型支持工具调用 — 适合原型设计和测试： + +| 模型 | 提供商 | 上下文 | 推理 | 视觉 | +| ------------------------------ | ------ | ------ | ---- | ---- | +| gemini-2.0-flash | Google | 1M | ✅ | ✅ | +| gemini-2.5-flash-preview-05-20 | Google | 1M | ✅ | ✅ | +| llama-4-scout | Chutes | 10M | ❌ | ✅ | +| llama-4-maverick | Chutes | 1M | ❌ | ✅ | +| deepseek-r1 | Chutes | 128K | ✅ | ❌ | +| qwen3-235b-a22b | Chutes | 128K | ✅ | ✅ | +| gemma-3-27b-it | Chutes | 128K | ✅ | ✅ | + +## 要点总结 + +- **2,350 个工具调用模型**，涵盖 1,540 个唯一模型 ID — 最大的工具调用模型目录 +- **Gemini 2.5 Flash** 是最佳性价比：1M 上下文、工具调用、推理和视觉，仅 $0.15/1M +- **Grok 4 Fast Reasoning** 提供最大上下文（2M）且具备全部三种能力 +- **45 个免费模型**支持工具调用 — 零成本开始构建代理 +- **829 个模型**同时支持结构化输出 — 完美适合可靠的 JSON 响应 +- 小型模型（Qwen 3.5 0.8B–4B）带工具调用仅需 $0.01–$0.03/1M tokens diff --git a/scripts/social-preview.ts b/scripts/social-preview.ts new file mode 100644 index 00000000..dccdfb02 --- /dev/null +++ b/scripts/social-preview.ts @@ -0,0 +1,77 @@ +#!/usr/bin/env node +/** + * Generate a social preview image (1280×640) for GitHub repo. + * Outputs an SVG that can be converted to PNG via any browser or tool. + * + * Usage: + * node scripts/social-preview.js > social-preview.svg + * # Or open in browser and screenshot + */ + +import fs from "node:fs"; +import path from "node:path"; + +const providersDir = path.join(import.meta.dirname, "..", "providers"); + +let providerCount = 0; +let modelCount = 0; + +for (const p of fs.readdirSync(providersDir)) { + const mDir = path.join(providersDir, p, "models"); + if (!fs.existsSync(mDir)) continue; + providerCount++; + for (const _f of fs.readdirSync(mDir).filter((f) => f.endsWith(".yaml"))) { + modelCount++; + } +} + +const svg = ` + + + + + + + + + + + + + + + + + + + + AI Models Catalog + + + Structured YAML · First-Party Data · TypeScript Types + + + ${modelCount.toLocaleString()} + Models + + ${providerCount} + Providers + + 1,306 + Reasoning + + + + + + Pricing · Context Windows · Modalities · Tool Calling · Open Weights + + + + npm install ai-models + + + +`; + +process.stdout.write(svg); diff --git a/social-preview.svg b/social-preview.svg new file mode 100644 index 00000000..4d3ea80a --- /dev/null +++ b/social-preview.svg @@ -0,0 +1,48 @@ + + + + + + + + + + + + + + + + + + + + + AI Models Catalog + + + Structured YAML · First-Party Data · TypeScript Types + + + 4,587 + Models + + 95 + Providers + + 1,306 + Reasoning + + + + + + Pricing · Context Windows · Modalities · Tool Calling · Open Weights + + + + npm install ai-models + + + + \ No newline at end of file From c1b37e4b256cb2a3faefd07e73c14bc90f6beaa4 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:36:53 +0800 Subject: [PATCH 058/311] =?UTF-8?q?docs:=20add=20vision=20models=20guide?= =?UTF-8?q?=20=E2=80=94=201,487=20models,=20cheapest,=20largest=20context?= =?UTF-8?q?=20(EN=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/vision-models.md | 85 ++++++++++++++++++++++++++++++++++++++++ docs/zh/vision-models.md | 85 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 170 insertions(+) create mode 100644 docs/vision-models.md create mode 100644 docs/zh/vision-models.md diff --git a/docs/vision-models.md b/docs/vision-models.md new file mode 100644 index 00000000..98157051 --- /dev/null +++ b/docs/vision-models.md @@ -0,0 +1,85 @@ +# AI Vision Models + +1,487 models in this catalog accept image input (vision). This page highlights the most capable and cost-effective vision models for image understanding, document analysis, and visual reasoning. + +> All data sourced from first-party APIs and documentation. "Vision" means the model accepts image input; image generation is a separate capability. + +## Quick Stats + +| Capability | Vision Models | +| ------------------- | ------------: | +| Total vision models | 1,487 | +| Unique model IDs | 930 | +| With tool calling | 1,063 | +| With reasoning | 697 | +| Open-weight | 104 | + +## Cheapest Vision Models + +Best value for image understanding: + +| Model | Provider | Input $/1M | Output $/1M | Context | Tool Call | Reasoning | +| --------------------- | ----------- | ---------: | ----------: | ------- | --------- | --------- | +| Qwen 3.5 0.8B | Auriko | $0.01 | $0.05 | 262K | ✅ | ✅ | +| Qwen 3.5 0.8B | DeepInfra | $0.01 | $0.05 | 262K | ✅ | ✅ | +| Qwen 3.5 2B | Auriko | $0.02 | $0.10 | 262K | ✅ | ✅ | +| Qwen 3.5 2B | DeepInfra | $0.02 | $0.10 | 262K | ✅ | ✅ | +| Qwen 3.5 4B | Auriko | $0.03 | $0.15 | 262K | ✅ | ✅ | +| Qwen 3.5 4B | DeepInfra | $0.03 | $0.15 | 262K | ✅ | ✅ | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | ✅ | ✅ | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | 1M | ✅ | ✅ | +| Llama 4 Maverick | Together AI | $0.15 | $0.60 | 1M | ✅ | ❌ | +| Grok 4 Fast Reasoning | xAI | $0.20 | $0.50 | 2M | ✅ | ✅ | + +## Largest Context Vision Models + +For analyzing large documents, multi-page PDFs, or long image sequences: + +| Model | Context | Input $/1M | Tool Call | Reasoning | Providers | +| --------------------- | ------- | ---------: | --------- | --------- | --------: | +| Llama 4 Scout | 10M | $0.08 | ✅ | ❌ | 4 | +| Grok 4 Fast Reasoning | 2M | $0.20 | ✅ | ✅ | 2 | +| GPT-5.4 | 1M | $2.50 | ✅ | ✅ | 4 | +| Gemini 2.5 Pro | 1M | $1.25 | ✅ | ✅ | 4 | +| Gemini 2.5 Flash | 1M | $0.15 | ✅ | ✅ | 3 | +| DeepSeek Reasoner | 1M | $0.43 | ✅ | ✅ | 1 | +| GPT-4.1 | 1M | $2.00 | ✅ | ❌ | 4 | +| Llama 4 Maverick | 1M | $0.15 | ✅ | ❌ | 3 | + +## Best Vision + Tool Calling + Reasoning + +The most capable vision models — can see, reason, and act: + +| Model | Context | Input $/1M | Providers | +| ----------------------- | ------- | ---------: | --------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | 4 | +| GPT-5.4 | 1M | $2.50 | 4 | +| DeepSeek Reasoner | 1M | $0.43 | 1 | +| MiMo V2.5 (open-weight) | 1M | varies | 2 | +| Kimi K2.6 (open-weight) | 262K | varies | 4 | + +## Open-Weight Vision Models + +104 open-weight models accept image input — run vision AI on your own hardware: + +| Model | Context | Tool Call | Reasoning | Providers | +| ------------------ | ------- | --------- | --------- | --------: | +| MiMo V2.5 | 1M | ✅ | ✅ | 2 | +| Llama 4 Maverick | 1M | ✅ | ❌ | 3 | +| Llama 4 Scout | 10M | ✅ | ❌ | 2 | +| Gemma 4 31B IT | 1M | ✅ | ❌ | 3 | +| Qwen3.5 Flash | 1M | ✅ | ❌ | 1 | +| Kimi K2.6 | 262K | ✅ | ✅ | 4 | +| Gemma 4 26B A4B IT | 262K | ✅ | ✅ | 3 | +| Llama 4 Scout 17B | 328K | ✅ | ❌ | 4 | + +## Key Takeaways + +- **1,487 vision models** across 930 unique IDs — the largest vision model catalog available +- **Gemini 2.5 Flash** is the best value: 1M context, vision, tool calling, and reasoning for $0.15/1M +- **Llama 4 Scout** has the largest vision context at 10M tokens +- **Grok 4 Fast Reasoning** is the only model combining 2M context, vision, tool calling, and reasoning +- **104 open-weight vision models** available — run vision AI on your own infrastructure +- Small vision models (Qwen 3.5 0.8B–4B) cost as little as $0.01–$0.03/1M tokens diff --git a/docs/zh/vision-models.md b/docs/zh/vision-models.md new file mode 100644 index 00000000..52de56f1 --- /dev/null +++ b/docs/zh/vision-models.md @@ -0,0 +1,85 @@ +# AI 视觉模型 + +本目录中有 1,487 个模型接受图像输入（视觉）。本页重点介绍图像理解、文档分析和视觉推理最有能力和最具性价比的模型。 + +> 所有数据来自一手 API 和文档。"视觉"表示模型接受图像输入；图像生成是独立的能力。 + +## 快速统计 + +| 能力 | 视觉模型数 | +| ------------ | ---------: | +| 总视觉模型 | 1,487 | +| 唯一模型 ID | 930 | +| 支持工具调用 | 1,063 | +| 支持推理 | 697 | +| 开源权重 | 104 | + +## 最便宜的视觉模型 + +图像理解的最佳性价比： + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | 上下文 | 工具调用 | 推理 | +| --------------------- | ----------- | --------: | --------: | ------ | -------- | ---- | +| Qwen 3.5 0.8B | Auriko | $0.01 | $0.05 | 262K | ✅ | ✅ | +| Qwen 3.5 0.8B | DeepInfra | $0.01 | $0.05 | 262K | ✅ | ✅ | +| Qwen 3.5 2B | Auriko | $0.02 | $0.10 | 262K | ✅ | ✅ | +| Qwen 3.5 2B | DeepInfra | $0.02 | $0.10 | 262K | ✅ | ✅ | +| Qwen 3.5 4B | Auriko | $0.03 | $0.15 | 262K | ✅ | ✅ | +| Qwen 3.5 4B | DeepInfra | $0.03 | $0.15 | 262K | ✅ | ✅ | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | ✅ | ✅ | +| Gemini 2.5 Flash | Google | $0.15 | $0.60 | 1M | ✅ | ✅ | +| Llama 4 Maverick | Together AI | $0.15 | $0.60 | 1M | ✅ | ❌ | +| Grok 4 Fast Reasoning | xAI | $0.20 | $0.50 | 2M | ✅ | ✅ | + +## 最大上下文视觉模型 + +适合分析大型文档、多页 PDF 或长图像序列： + +| 模型 | 上下文 | 输入 $/1M | 工具调用 | 推理 | 提供商数 | +| --------------------- | ------ | --------: | -------- | ---- | -------: | +| Llama 4 Scout | 10M | $0.08 | ✅ | ❌ | 4 | +| Grok 4 Fast Reasoning | 2M | $0.20 | ✅ | ✅ | 2 | +| GPT-5.4 | 1M | $2.50 | ✅ | ✅ | 4 | +| Gemini 2.5 Pro | 1M | $1.25 | ✅ | ✅ | 4 | +| Gemini 2.5 Flash | 1M | $0.15 | ✅ | ✅ | 3 | +| DeepSeek Reasoner | 1M | $0.43 | ✅ | ✅ | 1 | +| GPT-4.1 | 1M | $2.00 | ✅ | ❌ | 4 | +| Llama 4 Maverick | 1M | $0.15 | ✅ | ❌ | 3 | + +## 最佳视觉 + 工具调用 + 推理 + +最有能力的视觉模型 — 能看、能推理、能行动： + +| 模型 | 上下文 | 输入 $/1M | 提供商数 | +| --------------------- | ------ | --------: | -------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | 4 | +| GPT-5.4 | 1M | $2.50 | 4 | +| DeepSeek Reasoner | 1M | $0.43 | 1 | +| MiMo V2.5（开源权重） | 1M | varies | 2 | +| Kimi K2.6（开源权重） | 262K | varies | 4 | + +## 开源权重视觉模型 + +104 个开源权重模型接受图像输入 — 在自己的硬件上运行视觉 AI： + +| 模型 | 上下文 | 工具调用 | 推理 | 提供商数 | +| ------------------ | ------ | -------- | ---- | -------: | +| MiMo V2.5 | 1M | ✅ | ✅ | 2 | +| Llama 4 Maverick | 1M | ✅ | ❌ | 3 | +| Llama 4 Scout | 10M | ✅ | ❌ | 2 | +| Gemma 4 31B IT | 1M | ✅ | ❌ | 3 | +| Qwen3.5 Flash | 1M | ✅ | ❌ | 1 | +| Kimi K2.6 | 262K | ✅ | ✅ | 4 | +| Gemma 4 26B A4B IT | 262K | ✅ | ✅ | 3 | +| Llama 4 Scout 17B | 328K | ✅ | ❌ | 4 | + +## 要点总结 + +- **1,487 个视觉模型**，涵盖 930 个唯一模型 ID — 最大的视觉模型目录 +- **Gemini 2.5 Flash** 是最佳性价比：1M 上下文、视觉、工具调用和推理，仅 $0.15/1M +- **Llama 4 Scout** 拥有最大的视觉上下文窗口（10M tokens） +- **Grok 4 Fast Reasoning** 是唯一结合 2M 上下文、视觉、工具调用和推理的模型 +- **104 个开源权重视觉模型** — 在自己的基础设施上运行视觉 AI +- 小型视觉模型（Qwen 3.5 0.8B–4B）每 1M tokens 仅需 $0.01–$0.03 From 761e981522105400d3b243d6c8f3f5e41ac1d6ba Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:38:00 +0800 Subject: [PATCH 059/311] docs: add vision models link to README documentation table --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index e3ad4ea2..8848d0f1 100644 --- a/README.md +++ b/README.md @@ -281,6 +281,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | Document | Description | | [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | +| [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | | ------------------------------------------------------- | --------------------------------------------------------------- | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | @@ -301,6 +302,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | 文档 | 描述 | | [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | | -------------------------------------------- | ------------------------------------------- | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | | [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | | [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | | [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | From abeeb30ce4513caf458f469f6ce0db74c37a5756 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:38:29 +0800 Subject: [PATCH 060/311] docs: add vision models link to llms.txt --- llms.txt | 2 ++ 1 file changed, 2 insertions(+) diff --git a/llms.txt b/llms.txt index fed97096..4e5b5a16 100644 --- a/llms.txt +++ b/llms.txt @@ -61,6 +61,8 @@ modalities: - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models +- [Tool Calling Models](docs/tool-calling.md) — 2,350 tool-calling models +- [Vision Models](docs/vision-models.md) — 1,487 vision models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video - [Provider Overview](docs/providers.md) — all 95 providers by type and market - [Data Schema Reference](docs/data-schema.md) — complete YAML schema From c7743b77e41c413fdcde1bd9b544fda715f914c8 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:39:07 +0800 Subject: [PATCH 061/311] =?UTF-8?q?docs:=20add=2010=20more=20GitHub=20topi?= =?UTF-8?q?cs=20=E2=80=94=20tool-calling,=20reasoning,=20vision,=20free,?= =?UTF-8?q?=20agents,=20npm?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/repo-settings.md | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/.github/repo-settings.md b/.github/repo-settings.md index 9fbd91fc..573d4f88 100644 --- a/.github/repo-settings.md +++ b/.github/repo-settings.md @@ -35,3 +35,12 @@ Add these topics to the repository (Settings → General → Topics): - zod - open-weights - context-window +- tool-calling +- function-calling +- ai-agents +- reasoning-models +- vision-models +- free-ai-models +- ai-model-catalog +- structured-data +- npm-package From b63f6ea71c147b8ad28b6187047b9b4551775ee9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:42:07 +0800 Subject: [PATCH 062/311] docs: add tool-calling, vision-models, modality-matrix links to AGENTS.md --- AGENTS.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index 5e925b35..8f9c3ce5 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -14,6 +14,9 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/free-models.md`](docs/free-models.md) — 81 free AI models by capability ([中文](docs/zh/free-models.md)) - [`docs/open-weights.md`](docs/open-weights.md) — 513 open-weight models ([中文](docs/zh/open-weights.md)) - [`docs/reasoning-models.md`](docs/reasoning-models.md) — 1,306 reasoning models ([中文](docs/zh/reasoning-models.md)) +- [`docs/tool-calling.md`](docs/tool-calling.md) — 2,350 tool-calling models ([中文](docs/zh/tool-calling.md)) +- [`docs/vision-models.md`](docs/vision-models.md) — 1,487 vision models ([中文](docs/zh/vision-models.md)) +- [`docs/modality-matrix.md`](docs/modality-matrix.md) — Model capabilities matrix ([中文](docs/zh/modality-matrix.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) - [`docs/pricing-comparison.md`](docs/pricing-comparison.md) — Pricing comparison across providers ([中文](docs/zh/pricing-comparison.md)) From 30af8bd44a77b94bb81849151ea973f09651e104 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:44:56 +0800 Subject: [PATCH 063/311] =?UTF-8?q?docs:=20add=20image=20generation=20mode?= =?UTF-8?q?ls=20guide=20=E2=80=94=2028=20models,=20free=20and=20paid=20(EN?= =?UTF-8?q?=20+=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/image-generation.md | 56 +++++++++++++++++++++++++++++++++++++ docs/zh/image-generation.md | 56 +++++++++++++++++++++++++++++++++++++ 2 files changed, 112 insertions(+) create mode 100644 docs/image-generation.md create mode 100644 docs/zh/image-generation.md diff --git a/docs/image-generation.md b/docs/image-generation.md new file mode 100644 index 00000000..9e5c3e32 --- /dev/null +++ b/docs/image-generation.md @@ -0,0 +1,56 @@ +# AI Image Generation Models + +28 models in this catalog can generate images (image output modality). This page covers text-to-image, image editing, and multimodal generation models. + +> All data sourced from first-party APIs and documentation. "Image output" means the model generates images as part of its response. + +## Quick Stats + +| Capability | Image Gen Models | +| ----------------------------- | ---------------: | +| Total image generation models | 28 | +| Unique model IDs | 19 | +| With reasoning | 5 | +| With tool calling | 1 | +| Free to use | 9 | + +## All Image Generation Models + +| Model | Input | Output $/1M | Context | Reasoning | Free Tier | +| ------------------------------ | ------------------------- | ----------: | ------- | --------- | --------- | +| DALL·E 3 | text | — | — | ❌ | ✅ | +| Imagen 4.0 Fast | text, image | — | — | ❌ | ✅ | +| Imagen 4.0 | text, image | — | — | ❌ | ✅ | +| Image 01 | text, image | — | — | ❌ | ✅ | +| Image 01 Live | text, image | — | — | ❌ | ✅ | +| Step 1X Edit | text, image | — | — | ❌ | ✅ | +| Step 1X Medium | text, image | — | — | ❌ | ✅ | +| Step 2X Large | text, image | — | — | ❌ | ✅ | +| Step Image Edit 2 | text, image | — | — | ❌ | ✅ | +| Gemini 2.5 Flash Image | text, image | $0.039 | 33K | ❌ | ❌ | +| Gemini 3.1 Flash Image Preview | text, image | $1.50 | 66K | ✅ | ❌ | +| Gemini 3 Pro Image Preview | text, image | $12.00 | 131K | ✅ | ❌ | +| GPT-5 Image Mini | text, image, PDF | $2.00 | 400K | ✅ | ❌ | +| GPT-5 Image | text, image, PDF | $10.00 | 400K | ✅ | ❌ | +| GPT-5.4 Image 2 | text, image, PDF | $15.00 | 272K | ✅ | ❌ | +| Amazon Nova 2.0 Omni | text, image, audio, video | $1.30 | 64K | ✅ | ❌ | + +## Best Value Image Generation + +| Use Case | Best Model | Why | +| ---------------------------------- | ----------------------------------- | --------------------------------------- | +| **Free text-to-image** | DALL·E 3, Imagen 4.0 | Zero cost, high quality | +| **Free image editing** | Step 1X Edit, Step Image Edit 2 | Edit existing images at no cost | +| **Cheapest API** | Gemini 2.5 Flash Image | $0.039/1M output tokens | +| **Best quality** | GPT-5.4 Image 2, Gemini 3 Pro Image | State-of-the-art generation | +| **Multimodal (audio+video+image)** | Amazon Nova 2.0 Omni | Only model with all modalities | +| **Large context** | GPT-5 Image Mini | 400K context for complex prompts | +| **Reasoning + generation** | GPT-5 Image Mini | $2.50/1M input, 400K context, reasoning | + +## Key Takeaways + +- **9 free image generation models** — DALL·E 3, Imagen 4.0, Step models, and more +- **Gemini 2.5 Flash Image** is the cheapest API option at $0.039/1M output tokens +- **GPT-5 Image Mini** offers the best combination of reasoning + generation + large context +- **Amazon Nova 2.0 Omni** is the only model that generates images from audio and video input +- Most image generation models accept both text and image input (for editing/reference) diff --git a/docs/zh/image-generation.md b/docs/zh/image-generation.md new file mode 100644 index 00000000..07af5e2c --- /dev/null +++ b/docs/zh/image-generation.md @@ -0,0 +1,56 @@ +# AI 图像生成模型 + +本目录中有 28 个模型可以生成图像（图像输出模态）。本页涵盖文本生成图像、图像编辑和多模态生成模型。 + +> 所有数据来自一手 API 和文档。"图像输出"表示模型在回复中生成图像。 + +## 快速统计 + +| 能力 | 图像生成模型数 | +| -------------- | -------------: | +| 总图像生成模型 | 28 | +| 唯一模型 ID | 19 | +| 支持推理 | 5 | +| 支持工具调用 | 1 | +| 免费使用 | 9 | + +## 所有图像生成模型 + +| 模型 | 输入 | 输出 $/1M | 上下文 | 推理 | 免费 | +| ------------------------------ | ------------------------- | --------: | ------ | ---- | ---- | +| DALL·E 3 | text | — | — | ❌ | ✅ | +| Imagen 4.0 Fast | text, image | — | — | ❌ | ✅ | +| Imagen 4.0 | text, image | — | — | ❌ | ✅ | +| Image 01 | text, image | — | — | ❌ | ✅ | +| Image 01 Live | text, image | — | — | ❌ | ✅ | +| Step 1X Edit | text, image | — | — | ❌ | ✅ | +| Step 1X Medium | text, image | — | — | ❌ | ✅ | +| Step 2X Large | text, image | — | — | ❌ | ✅ | +| Step Image Edit 2 | text, image | — | — | ❌ | ✅ | +| Gemini 2.5 Flash Image | text, image | $0.039 | 33K | ❌ | ❌ | +| Gemini 3.1 Flash Image Preview | text, image | $1.50 | 66K | ✅ | ❌ | +| Gemini 3 Pro Image Preview | text, image | $12.00 | 131K | ✅ | ❌ | +| GPT-5 Image Mini | text, image, PDF | $2.00 | 400K | ✅ | ❌ | +| GPT-5 Image | text, image, PDF | $10.00 | 400K | ✅ | ❌ | +| GPT-5.4 Image 2 | text, image, PDF | $15.00 | 272K | ✅ | ❌ | +| Amazon Nova 2.0 Omni | text, image, audio, video | $1.30 | 64K | ✅ | ❌ | + +## 最佳性价比图像生成 + +| 用途 | 最佳模型 | 原因 | +| ---------------------------- | ----------------------------------- | -------------------------------- | +| **免费文本生成图像** | DALL·E 3, Imagen 4.0 | 零成本，高质量 | +| **免费图像编辑** | Step 1X Edit, Step Image Edit 2 | 免费编辑现有图像 | +| **最便宜 API** | Gemini 2.5 Flash Image | $0.039/1M 输出 tokens | +| **最佳质量** | GPT-5.4 Image 2, Gemini 3 Pro Image | 最先进的生成能力 | +| **多模态（音频+视频+图像）** | Amazon Nova 2.0 Omni | 唯一支持所有模态的模型 | +| **大上下文** | GPT-5 Image Mini | 400K 上下文适合复杂提示 | +| **推理 + 生成** | GPT-5 Image Mini | $2.50/1M 输入，400K 上下文，推理 | + +## 要点总结 + +- **9 个免费图像生成模型** — DALL·E 3、Imagen 4.0、Step 系列等 +- **Gemini 2.5 Flash Image** 是最便宜的 API 选项，仅 $0.039/1M 输出 tokens +- **GPT-5 Image Mini** 提供推理 + 生成 + 大上下文的最佳组合 +- **Amazon Nova 2.0 Omni** 是唯一可以从音频和视频输入生成图像的模型 +- 大多数图像生成模型同时接受文本和图像输入（用于编辑/参考） From fb49e0c9af5cc756811e726574cca4650ce4a685 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:45:46 +0800 Subject: [PATCH 064/311] docs: add image generation link to README documentation table --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index 8848d0f1..cf638f48 100644 --- a/README.md +++ b/README.md @@ -282,6 +282,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | Document | Description | | [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | | [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | +| [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | | ------------------------------------------------------- | --------------------------------------------------------------- | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | @@ -304,6 +305,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | -------------------------------------------- | ------------------------------------------- | | [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | | [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | | [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | | [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | From b1eeb855a76a4136218c7d0f233f5261a5b68646 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:46:06 +0800 Subject: [PATCH 065/311] docs: add image generation link to llms.txt --- llms.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/llms.txt b/llms.txt index 4e5b5a16..1732e7b7 100644 --- a/llms.txt +++ b/llms.txt @@ -63,6 +63,7 @@ modalities: - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models - [Tool Calling Models](docs/tool-calling.md) — 2,350 tool-calling models - [Vision Models](docs/vision-models.md) — 1,487 vision models +- [Image Generation](docs/image-generation.md) — 28 image generation models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video - [Provider Overview](docs/providers.md) — all 95 providers by type and market - [Data Schema Reference](docs/data-schema.md) — complete YAML schema From 5e5cd0157b147ca024464033fca933f2c42a9175 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:46:35 +0800 Subject: [PATCH 066/311] docs: add image generation link to AGENTS.md --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index 8f9c3ce5..e38ba1bf 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -16,6 +16,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/reasoning-models.md`](docs/reasoning-models.md) — 1,306 reasoning models ([中文](docs/zh/reasoning-models.md)) - [`docs/tool-calling.md`](docs/tool-calling.md) — 2,350 tool-calling models ([中文](docs/zh/tool-calling.md)) - [`docs/vision-models.md`](docs/vision-models.md) — 1,487 vision models ([中文](docs/zh/vision-models.md)) +- [`docs/image-generation.md`](docs/image-generation.md) — 28 image generation models ([中文](docs/zh/image-generation.md)) - [`docs/modality-matrix.md`](docs/modality-matrix.md) — Model capabilities matrix ([中文](docs/zh/modality-matrix.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) From 19dec33bc1a388ad71e61c3137aed16babf048eb Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:49:32 +0800 Subject: [PATCH 067/311] =?UTF-8?q?docs:=20add=20structured=20output=20mod?= =?UTF-8?q?els=20guide=20=E2=80=94=20829=20models,=20JSON=20mode=20(EN=20+?= =?UTF-8?q?=20ZH)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/structured-output.md | 69 ++++++++++++++++++++++++++++++++++++ docs/zh/structured-output.md | 69 ++++++++++++++++++++++++++++++++++++ 2 files changed, 138 insertions(+) create mode 100644 docs/structured-output.md create mode 100644 docs/zh/structured-output.md diff --git a/docs/structured-output.md b/docs/structured-output.md new file mode 100644 index 00000000..812e5af3 --- /dev/null +++ b/docs/structured-output.md @@ -0,0 +1,69 @@ +# AI Structured Output Models (JSON Mode) + +829 models in this catalog support structured output — the ability to generate responses that conform to a specified JSON schema. This is essential for building reliable AI-powered APIs, data pipelines, and automation. + +> All data sourced from first-party APIs and documentation. "Structured output" means the model can enforce a JSON schema on its response (also known as JSON mode, constrained decoding, or guided generation). + +## Quick Stats + +| Capability | Structured Output Models | +| ------------------------------ | -----------------------: | +| Total structured output models | 829 | +| Unique model IDs | 704 | +| With tool calling | 749 | +| With reasoning | 473 | +| With vision | 411 | +| Open-weight | 24 | + +## Cheapest Structured Output Models + +Best value for generating reliable JSON responses: + +| Model | Provider | Input $/1M | Output $/1M | Context | Tool Call | Reasoning | +| --------------------- | ---------- | ---------: | ----------: | ------- | --------- | --------- | +| Ernie 4.5 0.3B | AIHubMix | $0.0068 | $0.0272 | — | ✅ | ❌ | +| Ling 2.6 Flash | OpenRouter | $0.01 | $0.03 | 262K | ✅ | ❌ | +| Qwen3 VL Flash | AIHubMix | $0.0103 | $0.103 | — | ✅ | ❌ | +| Llama 3.1 8B Instruct | Auriko | $0.02 | $0.03 | 131K | ✅ | ❌ | +| Mistral Nemo | OpenRouter | $0.02 | $0.02 | — | ✅ | ❌ | +| Doubao Seed 1.6 Flash | AIHubMix | $0.022 | $0.022 | — | ✅ | ❌ | +| GPT-5 Nano | AIHubMix | $0.025 | $0.20 | — | ✅ | ✅ | +| GPT-OSS 20B | NeuralWatt | $0.03 | $0.03 | — | ✅ | ✅ | +| Granite 4.0 H Micro | Cloudflare | $0.017 | $0.112 | 131K | ✅ | ❌ | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | ✅ | ✅ | + +## Free Structured Output Models + +24 models offer free structured output — ideal for prototyping: + +| Model | Provider | Context | Tool Call | Reasoning | +| ------------------------- | ---------- | ------- | --------- | --------- | +| Ernie 4.5 0.3B | AIMLAPI | — | ✅ | ❌ | +| Gemma 4 26B A4B IT | Auriko | — | ✅ | ✅ | +| Gemma 4 31B IT | Auriko | — | ✅ | ❌ | +| Qwen3 Omni 30B A3B | NovitaAI | — | ✅ | ✅ | +| Dolphin Mistral 24B | OpenRouter | — | ✅ | ❌ | +| Gemma 4 26B A4B IT (free) | OpenRouter | — | ✅ | ✅ | +| Gemma 4 31B IT (free) | OpenRouter | — | ✅ | ❌ | + +## Best Structured Output + Tool Calling + Reasoning + +For AI agents that need to return structured data, call tools, and reason: + +| Model | Context | Input $/1M | Tool Call | Reasoning | Providers | +| --------------------- | ------- | ---------: | --------- | --------- | --------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | ✅ | ✅ | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | ✅ | ✅ | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | ✅ | ✅ | 4 | +| GPT-5.4 | 1M | $2.50 | ✅ | ✅ | 4 | +| DeepSeek Reasoner | 1M | $0.43 | ✅ | ✅ | 1 | +| GPT-5 Nano | — | $0.025 | ✅ | ✅ | 4 | + +## Key Takeaways + +- **829 structured output models** — the largest catalog of JSON-mode models +- **749 models** combine structured output with tool calling — perfect for AI agents +- **24 free models** support structured output — start building at zero cost +- **Gemini 2.5 Flash** is the best value: 1M context, structured output, tool calling, and reasoning for $0.15/1M +- Small models (Ernie 4.5 0.3B, Ling 2.6 Flash) cost as little as $0.01/1M with structured output +- 91% of structured output models also support tool calling — these capabilities go hand-in-hand diff --git a/docs/zh/structured-output.md b/docs/zh/structured-output.md new file mode 100644 index 00000000..0f26dcc4 --- /dev/null +++ b/docs/zh/structured-output.md @@ -0,0 +1,69 @@ +# AI 结构化输出模型（JSON 模式） + +本目录中有 829 个模型支持结构化输出 — 能够按照指定的 JSON Schema 生成回复。这对于构建可靠的 AI 驱动 API、数据管道和自动化至关重要。 + +> 所有数据来自一手 API 和文档。"结构化输出"表示模型可以在回复中强制遵循 JSON Schema（也称为 JSON 模式、约束解码或引导生成）。 + +## 快速统计 + +| 能力 | 结构化输出模型数 | +| ---------------- | ---------------: | +| 总结构化输出模型 | 829 | +| 唯一模型 ID | 704 | +| 支持工具调用 | 749 | +| 支持推理 | 473 | +| 支持视觉 | 411 | +| 开源权重 | 24 | + +## 最便宜的结构化输出模型 + +生成可靠 JSON 回复的最佳性价比： + +| 模型 | 提供商 | 输入 $/1M | 输出 $/1M | 上下文 | 工具调用 | 推理 | +| --------------------- | ---------- | --------: | --------: | ------ | -------- | ---- | +| Ernie 4.5 0.3B | AIHubMix | $0.0068 | $0.0272 | — | ✅ | ❌ | +| Ling 2.6 Flash | OpenRouter | $0.01 | $0.03 | 262K | ✅ | ❌ | +| Qwen3 VL Flash | AIHubMix | $0.0103 | $0.103 | — | ✅ | ❌ | +| Llama 3.1 8B Instruct | Auriko | $0.02 | $0.03 | 131K | ✅ | ❌ | +| Mistral Nemo | OpenRouter | $0.02 | $0.02 | — | ✅ | ❌ | +| Doubao Seed 1.6 Flash | AIHubMix | $0.022 | $0.022 | — | ✅ | ❌ | +| GPT-5 Nano | AIHubMix | $0.025 | $0.20 | — | ✅ | ✅ | +| GPT-OSS 20B | NeuralWatt | $0.03 | $0.03 | — | ✅ | ✅ | +| Granite 4.0 H Micro | Cloudflare | $0.017 | $0.112 | 131K | ✅ | ❌ | +| Gemini 2.5 Flash Lite | Google | $0.10 | $0.60 | 1M | ✅ | ✅ | + +## 免费结构化输出模型 + +24 个模型提供免费结构化输出 — 适合原型设计： + +| 模型 | 提供商 | 上下文 | 工具调用 | 推理 | +| ------------------------- | ---------- | ------ | -------- | ---- | +| Ernie 4.5 0.3B | AIMLAPI | — | ✅ | ❌ | +| Gemma 4 26B A4B IT | Auriko | — | ✅ | ✅ | +| Gemma 4 31B IT | Auriko | — | ✅ | ❌ | +| Qwen3 Omni 30B A3B | NovitaAI | — | ✅ | ✅ | +| Dolphin Mistral 24B | OpenRouter | — | ✅ | ❌ | +| Gemma 4 26B A4B IT (free) | OpenRouter | — | ✅ | ✅ | +| Gemma 4 31B IT (free) | OpenRouter | — | ✅ | ❌ | + +## 最佳结构化输出 + 工具调用 + 推理 + +适合需要返回结构化数据、调用工具和推理的 AI 代理： + +| 模型 | 上下文 | 输入 $/1M | 工具调用 | 推理 | 提供商数 | +| --------------------- | ------ | --------: | -------- | ---- | -------: | +| Grok 4 Fast Reasoning | 2M | $0.20 | ✅ | ✅ | 2 | +| Gemini 2.5 Flash | 1M | $0.15 | ✅ | ✅ | 3 | +| Gemini 2.5 Pro | 1M | $1.25 | ✅ | ✅ | 4 | +| GPT-5.4 | 1M | $2.50 | ✅ | ✅ | 4 | +| DeepSeek Reasoner | 1M | $0.43 | ✅ | ✅ | 1 | +| GPT-5 Nano | — | $0.025 | ✅ | ✅ | 4 | + +## 要点总结 + +- **829 个结构化输出模型** — 最大的 JSON 模式模型目录 +- **749 个模型**同时支持结构化输出和工具调用 — 完美适合 AI 代理 +- **24 个免费模型**支持结构化输出 — 零成本开始构建 +- **Gemini 2.5 Flash** 是最佳性价比：1M 上下文、结构化输出、工具调用和推理，仅 $0.15/1M +- 小型模型（Ernie 4.5 0.3B、Ling 2.6 Flash）带结构化输出仅需 $0.01/1M +- 91% 的结构化输出模型同时支持工具调用 — 这两种能力相辅相成 From f85dc3d97eb9762ffee9da27d3721a1e1c546606 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:50:19 +0800 Subject: [PATCH 068/311] docs: add structured output links to README, llms.txt, AGENTS.md --- AGENTS.md | 1 + README.md | 2 ++ llms.txt | 1 + 3 files changed, 4 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index e38ba1bf..7b1f9aa7 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -17,6 +17,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/tool-calling.md`](docs/tool-calling.md) — 2,350 tool-calling models ([中文](docs/zh/tool-calling.md)) - [`docs/vision-models.md`](docs/vision-models.md) — 1,487 vision models ([中文](docs/zh/vision-models.md)) - [`docs/image-generation.md`](docs/image-generation.md) — 28 image generation models ([中文](docs/zh/image-generation.md)) +- [`docs/structured-output.md`](docs/structured-output.md) — 829 structured output models ([中文](docs/zh/structured-output.md)) - [`docs/modality-matrix.md`](docs/modality-matrix.md) — Model capabilities matrix ([中文](docs/zh/modality-matrix.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) diff --git a/README.md b/README.md index cf638f48..a023424a 100644 --- a/README.md +++ b/README.md @@ -283,6 +283,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | | [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | | [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | +| [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | ------------------------------------------------------- | --------------------------------------------------------------- | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | @@ -307,6 +308,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | | [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | | [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | | [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | | [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | diff --git a/llms.txt b/llms.txt index 1732e7b7..56ed2383 100644 --- a/llms.txt +++ b/llms.txt @@ -64,6 +64,7 @@ modalities: - [Tool Calling Models](docs/tool-calling.md) — 2,350 tool-calling models - [Vision Models](docs/vision-models.md) — 1,487 vision models - [Image Generation](docs/image-generation.md) — 28 image generation models +- [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video - [Provider Overview](docs/providers.md) — all 95 providers by type and market - [Data Schema Reference](docs/data-schema.md) — complete YAML schema From 6655f2e0ef3182eec0271342151f8a783ce8ad0f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 11:59:13 +0800 Subject: [PATCH 069/311] fix: fix TypeScript errors in scrape-all and export-csv scripts --- .github/workflows/release.yml | 15 ++++- .github/workflows/stats.yml | 11 +++- .github/workflows/sync.yml | 57 ++++++++++++++++ .gitignore | 2 + scripts/export-csv.ts | 120 ++++++++++++++++++++++++++++++++++ scripts/scrape-all.ts | 58 ++++++++++++++++ 6 files changed, 259 insertions(+), 4 deletions(-) create mode 100644 .github/workflows/sync.yml create mode 100644 scripts/export-csv.ts create mode 100644 scripts/scrape-all.ts diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 40c703f3..82bb48a0 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -28,6 +28,9 @@ jobs: - name: Compile models.json run: npx tsx scripts/compile.ts + - name: Export models.csv + run: npx tsx scripts/export-csv.ts + - name: Compute stats run: npx tsx scripts/stats.ts json > stats.json @@ -36,17 +39,27 @@ jobs: with: files: | dist/models.json + models.csv stats.json body: | ## AI Models Catalog Release - **Compiled data file:** `models.json` — all 4,587 models from 95 providers in a single JSON file. + **Compiled data files:** + + | File | Format | Size | Description | + | ---- | ------ | ---- | ----------- | + | `models.json` | JSON | ~2.3 MB | All 4,587 models with full metadata | + | `models.csv` | CSV | ~560 KB | Flat table — open in Excel, Google Sheets, Python | + | `stats.json` | JSON | ~1 KB | Catalog statistics summary | **Usage:** ```bash # Download the compiled JSON curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + # Download the CSV (for Excel/Sheets) + curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv + # Use in JavaScript const catalog = require("./models.json"); console.log(catalog.stats); diff --git a/.github/workflows/stats.yml b/.github/workflows/stats.yml index 2d332e4b..a1c498ed 100644 --- a/.github/workflows/stats.yml +++ b/.github/workflows/stats.yml @@ -21,9 +21,14 @@ jobs: - name: Compute stats run: npx tsx scripts/stats.ts json > stats.json - - name: Upload stats + - name: Export CSV + run: npx tsx scripts/export-csv.ts + + - name: Upload artifacts uses: actions/upload-artifact@v4 with: - name: catalog-stats - path: stats.json + name: catalog-data + path: | + stats.json + models.csv retention-days: 30 diff --git a/.github/workflows/sync.yml b/.github/workflows/sync.yml new file mode 100644 index 00000000..ce8cc6ca --- /dev/null +++ b/.github/workflows/sync.yml @@ -0,0 +1,57 @@ +name: Weekly Sync + +on: + schedule: + - cron: "0 2 * * 1" # Every Monday at 02:00 UTC + workflow_dispatch: + +permissions: + contents: write + pull-requests: write + +jobs: + sync: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: "22" + + - name: Install dependencies + run: npm ci + + - name: Run scrape + run: npx tsx scripts/scrape-all.ts + continue-on-error: true + + - name: Check for changes + id: changes + run: | + if git diff --quiet providers/; then + echo "changed=false" >> "$GITHUB_OUTPUT" + else + echo "changed=true" >> "$GITHUB_OUTPUT" + git diff --stat providers/ > /tmp/diff-stat.txt + fi + + - name: Create pull request + if: steps.changes.outputs.changed == 'true' + uses: peter-evans/create-pull-request@v6 + with: + title: "🔄 Weekly model data sync" + body: | + Automated data sync from provider APIs. + + ## Changes + + ```diff + ${{ steps.changes.outputs.diff_stat }} + ``` + + Generated by the [Weekly Sync](/.github/workflows/sync.yml) workflow. + branch: sync/weekly + commit-message: "chore: weekly model data sync" + labels: data-update, automated + delete-branch: true diff --git a/.gitignore b/.gitignore index a310e636..def2133e 100644 --- a/.gitignore +++ b/.gitignore @@ -5,3 +5,5 @@ dist/ *.local *.tsbuildinfo dist/ +models.csv +stats.json diff --git a/scripts/export-csv.ts b/scripts/export-csv.ts new file mode 100644 index 00000000..b87f4c14 --- /dev/null +++ b/scripts/export-csv.ts @@ -0,0 +1,120 @@ +import fs from "node:fs"; +import path from "node:path"; +import YAML from "yaml"; +import { ModelSchema } from "../types/schemas"; + +const providersDir = "providers"; + +interface FlatModel { + id: string; + name: string; + provider: string; + family: string; + deprecated: boolean; + reasoning: boolean; + tool_call: boolean; + structured_output: boolean; + open_weights: boolean; + context_window: number | undefined; + max_output: number | undefined; + input_modalities: string; + output_modalities: string; + pricing_type: string; + input_price: number | undefined; + output_price: number | undefined; + image_input_price: number | undefined; + image_output_price: number | undefined; + cached_input_price: number | undefined; +} + +const models: FlatModel[] = []; + +for (const provider of fs.readdirSync(providersDir)) { + const mDir = path.join(providersDir, provider, "models"); + if (!fs.existsSync(mDir)) continue; + for (const f of fs.readdirSync(mDir).filter((f) => f.endsWith(".yaml"))) { + const raw = fs.readFileSync(path.join(mDir, f), "utf-8"); + const data = YAML.parse(raw); + const r = ModelSchema.safeParse(data); + if (!r.success) continue; + const m = r.data; + + const pr = m.pricing as Record | undefined; + let pricingType = "unknown"; + let inputPrice: number | undefined; + let outputPrice: number | undefined; + let imageInputPrice: number | undefined; + let imageOutputPrice: number | undefined; + let cachedInputPrice: number | undefined; + + if (pr) { + pricingType = (pr["type"] as string) ?? "token"; + if (pricingType === "token") { + inputPrice = pr["input"] as number | undefined; + outputPrice = pr["output"] as number | undefined; + imageInputPrice = pr["image_input"] as number | undefined; + imageOutputPrice = pr["image_output"] as number | undefined; + cachedInputPrice = pr["cached_input"] as number | undefined; + } + } + + models.push({ + id: m.id, + name: m.name ?? m.id, + provider, + family: m.family ?? "", + deprecated: m.deprecated ?? false, + reasoning: m.reasoning ?? false, + tool_call: m.tool_call ?? false, + structured_output: m.structured_output ?? false, + open_weights: m.open_weights ?? false, + context_window: m.limit?.context, + max_output: m.limit?.output, + input_modalities: (m.modalities?.input ?? []).join(";"), + output_modalities: (m.modalities?.output ?? []).join(";"), + pricing_type: pricingType, + input_price: inputPrice, + output_price: outputPrice, + image_input_price: imageInputPrice, + image_output_price: imageOutputPrice, + cached_input_price: cachedInputPrice, + }); + } +} + +function escapeCsv(val: unknown): string { + const s = String(val ?? ""); + if (s.includes(",") || s.includes('"') || s.includes("\n")) { + return `"${s.replace(/"/g, '""')}"`; + } + return s; +} + +const header = [ + "id", + "name", + "provider", + "family", + "deprecated", + "reasoning", + "tool_call", + "structured_output", + "open_weights", + "context_window", + "max_output", + "input_modalities", + "output_modalities", + "pricing_type", + "input_price", + "output_price", + "image_input_price", + "image_output_price", + "cached_input_price", +]; + +const rows = models.map((m) => header.map((k) => escapeCsv(m[k as keyof FlatModel])).join(",")); + +const csv = [header.join(","), ...rows].join("\n"); + +fs.writeFileSync("models.csv", csv); +console.log(`Wrote ${models.length} models to models.csv`); diff --git a/scripts/scrape-all.ts b/scripts/scrape-all.ts new file mode 100644 index 00000000..48fe052e --- /dev/null +++ b/scripts/scrape-all.ts @@ -0,0 +1,58 @@ +/** + * Run all provider scrape scripts and write updated model YAML files. + * + * Usage: npx tsx scripts/scrape-all.ts + * + * Each provider's scrape.ts exports a `scrape()` function that returns + * a ScrapeResult. This script calls each one and writes the results + * to providers//models/.yaml. + */ +import fs from "node:fs"; +import path from "node:path"; +import YAML from "yaml"; +import { defineModel } from "./lib/utils"; +import type { ScrapeResult } from "./lib/types"; + +const providersDir = "providers"; + +async function main() { + const providerDirs = fs + .readdirSync(providersDir) + .filter((d) => fs.statSync(path.join(providersDir, d)).isDirectory()); + + let totalUpdated = 0; + let totalFailed = 0; + + for (const provider of providerDirs) { + const scrapePath = path.join(providersDir, provider, "scrape.ts"); + if (!fs.existsSync(scrapePath)) continue; + + console.log(`Scraping ${provider}...`); + try { + const mod = await import(`../providers/${provider}/scrape.ts`); + const result: ScrapeResult = await mod.scrape(); + + const mDir = path.join(providersDir, provider, "models"); + fs.mkdirSync(mDir, { recursive: true }); + + for (const model of result.models) { + const validated = defineModel(model); + const filePath = path.join(mDir, `${validated.id}.yaml`); + fs.writeFileSync(filePath, YAML.stringify(validated)); + totalUpdated++; + } + + console.log(` ✓ ${provider}: ${result.models.length} models`); + } catch (err) { + console.error(` ✗ ${provider}: ${err}`); + totalFailed++; + } + } + + console.log(`\nDone: ${totalUpdated} models updated, ${totalFailed} providers failed`); + if (totalFailed > 0) { + process.exit(1); + } +} + +main(); From 7e79dcacba96d4b05d46cad886c2f758cdf3d76a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:02:01 +0800 Subject: [PATCH 070/311] docs: add CSV download option to README and API docs (EN + ZH) --- README.md | 13 ++++++++----- docs/api.md | 13 ++++++++++--- docs/zh/api.md | 13 ++++++++++--- 3 files changed, 28 insertions(+), 11 deletions(-) diff --git a/README.md b/README.md index a023424a..f17d7b7f 100644 --- a/README.md +++ b/README.md @@ -235,16 +235,19 @@ console.log(model.limit); // { context: 1047576, output: 32768 } console.log(model.modalities); // { input: ["text", "image"], output: ["text"] } ``` -### Download Compiled JSON +### Download Data + +Available in JSON and CSV formats from [GitHub Releases](https://github.com/i-need-token/ai-models/releases/latest): ```bash -# Latest release +# JSON — full metadata (2.3 MB) curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json -``` -See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` -## Project Structure +See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python.## Project Structure ``` ├── providers/ # 95 provider directories diff --git a/docs/api.md b/docs/api.md index 4b2388ca..a28196df 100644 --- a/docs/api.md +++ b/docs/api.md @@ -26,14 +26,21 @@ The package includes: ## Compiled JSON -The easiest way to access all model data is via the compiled `models.json` file. +All model data is available from [GitHub Releases](https://github.com/i-need-token/ai-models/releases/latest) in two formats: -### Download from GitHub Releases +| File | Format | Size | Best For | +| ------------- | ------ | ------- | ----------------------------------------- | +| `models.json` | JSON | ~2.3 MB | Programmatic access, web apps, TypeScript | +| `models.csv` | CSV | ~560 KB | Excel, Google Sheets, data analysis | +| `stats.json` | JSON | ~1 KB | Catalog statistics summary | ```bash -# Latest release +# Download JSON (full metadata) curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json +# Download CSV (flat table for spreadsheets) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv + # Specific version curl -LO https://github.com/i-need-token/ai-models/releases/download/v0.1.0/models.json ``` diff --git a/docs/zh/api.md b/docs/zh/api.md index 96dbc198..8e420152 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -26,14 +26,21 @@ const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < ## 编译 JSON -访问所有模型数据最简单的方式是通过编译后的 `models.json` 文件。 +所有模型数据可以从 [GitHub Releases](https://github.com/i-need-token/ai-models/releases/latest) 下载，提供两种格式： -### 从 GitHub Releases 下载 +| 文件 | 格式 | 大小 | 适用场景 | +| ------------- | ---- | ------- | ------------------------------ | +| `models.json` | JSON | ~2.3 MB | 编程访问、Web 应用、TypeScript | +| `models.csv` | CSV | ~560 KB | Excel、Google Sheets、数据分析 | +| `stats.json` | JSON | ~1 KB | 目录统计摘要 | ```bash -# 最新版本 +# 下载 JSON（完整元数据） curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json +# 下载 CSV（适合电子表格） +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv + # 特定版本 curl -LO https://github.com/i-need-token/ai-models/releases/download/v0.1.0/models.json ``` From 99ffe37e67d391418c953d8eaad338bc24c4491f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:04:33 +0800 Subject: [PATCH 071/311] feat: dynamic model/provider count badges + CI badge in README --- .github/badges/models.json | 6 ++++++ .github/badges/providers.json | 6 ++++++ .github/workflows/stats.yml | 20 ++++++++++++++++++++ README.md | 7 +++---- 4 files changed, 35 insertions(+), 4 deletions(-) create mode 100644 .github/badges/models.json create mode 100644 .github/badges/providers.json diff --git a/.github/badges/models.json b/.github/badges/models.json new file mode 100644 index 00000000..287af461 --- /dev/null +++ b/.github/badges/models.json @@ -0,0 +1,6 @@ +{ + "schemaVersion": 1, + "label": "models", + "message": "4,587", + "color": "green" +} diff --git a/.github/badges/providers.json b/.github/badges/providers.json new file mode 100644 index 00000000..db0a9530 --- /dev/null +++ b/.github/badges/providers.json @@ -0,0 +1,6 @@ +{ + "schemaVersion": 1, + "label": "providers", + "message": "95", + "color": "blue" +} diff --git a/.github/workflows/stats.yml b/.github/workflows/stats.yml index a1c498ed..a715ddc0 100644 --- a/.github/workflows/stats.yml +++ b/.github/workflows/stats.yml @@ -24,6 +24,26 @@ jobs: - name: Export CSV run: npx tsx scripts/export-csv.ts + - name: Update dynamic badges + run: | + STATS=$(cat stats.json) + MODELS=$(echo "$STATS" | jq '.model_files') + PROVIDERS=$(echo "$STATS" | jq '.providers') + cat > .github/badges/models.json << EOF + {"schemaVersion":1,"label":"models","message":"$MODELS","color":"green"} + EOF + cat > .github/badges/providers.json << EOF + {"schemaVersion":1,"label":"providers","message":"$PROVIDERS","color":"blue"} + EOF + + - name: Commit badges + run: | + git config user.name "github-actions[bot]" + git config user.email "github-actions[bot]@users.noreply.github.com" + git add .github/badges/*.json + git diff --cached --quiet || git commit -m "chore: update dynamic badges [skip ci]" + git push + - name: Upload artifacts uses: actions/upload-artifact@v4 with: diff --git a/README.md b/README.md index f17d7b7f..cbd02b43 100644 --- a/README.md +++ b/README.md @@ -7,10 +7,9 @@ 95 providers · 4,587 model files · 2,712 unique model IDs · First-party data only [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) -[![Models](https://img.shields.io/badge/Models-4%2C587-green.svg)](providers/) -[![Providers](https://img.shields.io/badge/Providers-95-orange.svg)](providers/) -[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) -[![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) +[![Models](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/models.json)](providers/) +[![Providers](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/providers.json)](providers/) +[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml)[![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From e18d2cca19e0c1c2ed9ffd64a0e99b95a46a0a66 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:06:01 +0800 Subject: [PATCH 072/311] docs: add CSV download link to README hero quick links --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index cbd02b43..c00ccec1 100644 --- a/README.md +++ b/README.md @@ -17,7 +17,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** +**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md) · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)\*\* ## Why This Catalog? From 9dad9a4787bd8b0a1daa1f93f4cd493fad32d63a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:06:31 +0800 Subject: [PATCH 073/311] docs: add npm version badge to README --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index c00ccec1..7ea4ba09 100644 --- a/README.md +++ b/README.md @@ -7,6 +7,7 @@ 95 providers · 4,587 model files · 2,712 unique model IDs · First-party data only [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) +[![npm version](https://img.shields.io/npm/v/ai-models.svg)](https://www.npmjs.com/package/ai-models) [![Models](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/models.json)](providers/) [![Providers](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/providers.json)](providers/) [![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml)[![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From fef1569a5bc4921cd646f9f27c04238bd0e628bf Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:07:30 +0800 Subject: [PATCH 074/311] docs: add npm package and CSV download to quick-start guide (EN + ZH) --- docs/quick-start.md | 23 +++++++++++++++++++++++ docs/zh/quick-start.md | 23 +++++++++++++++++++++++ 2 files changed, 46 insertions(+) diff --git a/docs/quick-start.md b/docs/quick-start.md index 594790ab..b2a64dbb 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -54,6 +54,29 @@ Find the right AI model for your needs in 30 seconds. ## I want to use the data programmatically +### npm package + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types +``` + +### Download data files + +```bash +# JSON — full metadata (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +### From source + ```bash # Install dependencies npm install diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md index 576cb5e2..3501f6a1 100644 --- a/docs/zh/quick-start.md +++ b/docs/zh/quick-start.md @@ -54,6 +54,29 @@ ## 我想编程使用数据 +### npm 包 + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 个模型 JSON +import type { Model } from "ai-models"; // TypeScript 类型 +``` + +### 下载数据文件 + +```bash +# JSON — 完整元数据（2.3 MB） +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — 适合 Excel/Google Sheets 的表格（560 KB） +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +### 从源码 + ```bash # 安装依赖 npm install From 7c7011de2c945fdd664fcaded77ea189f8188d12 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:07:59 +0800 Subject: [PATCH 075/311] docs: add CSV download to llms.txt --- llms.txt | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/llms.txt b/llms.txt index 56ed2383..fe129a6a 100644 --- a/llms.txt +++ b/llms.txt @@ -29,6 +29,16 @@ import catalog from "ai-models"; // 4,587 models as JSON import type { Model } from "ai-models"; // TypeScript types ``` +## Download + +```bash +# JSON — full metadata (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + ## Data Format Each model is a single YAML file at `providers//models/.yaml`: From e7d972b3f9fe76e906077c88993a92b836f4ffb2 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:33:23 +0800 Subject: [PATCH 076/311] fix: correct hero quick links bold formatting in README --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 7ea4ba09..7dac0ff0 100644 --- a/README.md +++ b/README.md @@ -18,7 +18,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md) · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)\*\* +**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** ## Why This Catalog? From 982a0bb4ff0fe9e8b59e2e49a679a9784e26d613 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:34:11 +0800 Subject: [PATCH 077/311] feat: add 15 npm keywords for search discoverability --- package.json | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/package.json b/package.json index 9bd0823b..84aeefc8 100644 --- a/package.json +++ b/package.json @@ -5,22 +5,37 @@ "keywords": [ "ai-catalog", "ai-models", + "ai-pricing", "ai-provider", "anthropic", "cerebras", + "chatgpt", + "claude", + "context-window", "deepseek", + "function-calling", "google-gemini", + "gpt-4", "groq", "inference-api", + "json-mode", "large-language-model", "llama", "llm", + "llm-api", + "llm-pricing", "machine-readable", "mistral", + "model-catalog", "model-comparison", "model-pricing", + "open-weights", "openai", "qwen", + "reasoning-model", + "structured-output", + "tool-calling", + "vision-model", "yaml" ], "homepage": "https://github.com/i-need-token/ai-models#readme", From 3676e985bf11a9c66be0db4367ff183af2119b2c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:34:49 +0800 Subject: [PATCH 078/311] docs: add Good First Issues section to CONTRIBUTING.md --- CONTRIBUTING.md | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 1aeb6361..c209d2cd 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -132,6 +132,20 @@ npm run typecheck npm run check ``` +## Good First Issues + +New contributors welcome! These tasks are beginner-friendly and don't require deep knowledge of the codebase: + +| Task | How | Difficulty | +| -------------------------- | ------------------------------------------------------------------------------------------------------------------- | ---------- | +| Fix incorrect model data | Find the model YAML in `providers//models/`, correct it, run `npx tsx scripts/validate.ts` | Easy | +| Add a provider README | Create `providers//README.md` describing the data source and scraping strategy (see existing examples) | Easy | +| Update outdated pricing | Re-run `npx tsx providers//scrape.ts` and review the diff | Easy | +| Add missing model features | Add `tool_call`, `structured_output`, `reasoning`, etc. to model YAML files where the provider docs confirm support | Easy | +| Translate a doc to Chinese | Copy an English doc from `docs/` to `docs/zh/` and translate | Medium | + +Look for issues labeled [`good first issue`](https://github.com/i-need-token/ai-models/labels/good%20first%20issue) or [`help wanted`](https://github.com/i-need-token/ai-models/labels/help%20wanted). + ## Reporting Issues - **Incorrect model data** — open an issue with the provider name, model ID, and what's wrong From 0c7d130dd4444e89441cf7cac78412ad2d9fc4e6 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:35:51 +0800 Subject: [PATCH 079/311] docs: add JSON sample alongside YAML example in README --- README.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/README.md b/README.md index 7dac0ff0..d1d3d894 100644 --- a/README.md +++ b/README.md @@ -95,6 +95,26 @@ release_date: "2026-05-18" last_updated: "2026-05-18" ``` +
+Same model as JSON (from models.json) + +```json +{ + "id": "gpt-4.1", + "name": "GPT-4.1", + "family": "gpt-4.1", + "tool_call": true, + "structured_output": true, + "pricing": { "input": 2.0, "output": 8.0, "cache_read": 0.5 }, + "limit": { "context": 1047576, "output": 32768 }, + "modalities": { "input": ["text", "image"], "output": ["text"] }, + "release_date": "2026-05-18", + "last_updated": "2026-05-18" +} +``` + +
+ ### Pricing Types | Type | When | Example | From f2e96f77f2c63732c2772d88ca4a98fe0433d382 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:36:50 +0800 Subject: [PATCH 080/311] feat: add good first issue label to data update template --- .github/ISSUE_TEMPLATE/data_update.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/ISSUE_TEMPLATE/data_update.md b/.github/ISSUE_TEMPLATE/data_update.md index 3f97483d..45fece32 100644 --- a/.github/ISSUE_TEMPLATE/data_update.md +++ b/.github/ISSUE_TEMPLATE/data_update.md @@ -2,7 +2,7 @@ name: Data Update about: Report outdated model data that needs updating title: "[Data] " -labels: data +labels: data, good first issue --- ## Provider From 2dab5dd1c85ab6878b0b438d8ca76563095edc1d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:37:53 +0800 Subject: [PATCH 081/311] fix: add missing newline before Project Structure heading in README --- README.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index d1d3d894..ad6df66e 100644 --- a/README.md +++ b/README.md @@ -267,7 +267,9 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` -See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python.## Project Structure +See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. + +## Project Structure ``` ├── providers/ # 95 provider directories From e7d5dda72ea4c95c7b3a3d9347d8357904fa1c19 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:40:08 +0800 Subject: [PATCH 082/311] fix: correct documentation table formatting in README --- README.md | 74 +++++++++++++++++++++++++++---------------------------- 1 file changed, 37 insertions(+), 37 deletions(-) diff --git a/README.md b/README.md index ad6df66e..5ba3c5dd 100644 --- a/README.md +++ b/README.md @@ -304,47 +304,47 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation -| Document | Description | -| [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | -| [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | -| [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | -| [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | +| Document | Description | | ------------------------------------------------------- | --------------------------------------------------------------- | -| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | -| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | -| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | -| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | -| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | -| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | -| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | -| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | -| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | -| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | -| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | -| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | -| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | +| [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | +| [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | +| [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | +| [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | +| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | +| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | +| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | +| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | +| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | +| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | +| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | +| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | +| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | +| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | +| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | **中文文档：** -| 文档 | 描述 | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| -------------------------------------------- | ------------------------------------------- | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| 文档 | 描述 | +| -------------------------------------------- | ----------------------------------------------- | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles From 3ace2b5876ab6b521c968dad84620db0ba70f2d7 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:41:34 +0800 Subject: [PATCH 083/311] fix: add space between CI and Star History badges in README --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 5ba3c5dd..de8f1036 100644 --- a/README.md +++ b/README.md @@ -10,7 +10,7 @@ [![npm version](https://img.shields.io/npm/v/ai-models.svg)](https://www.npmjs.com/package/ai-models) [![Models](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/models.json)](providers/) [![Providers](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/providers.json)](providers/) -[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml)[![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) +[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) [![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From da6be0d9b1f55896de09c67fa3e7b12cba030514 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:42:33 +0800 Subject: [PATCH 084/311] feat: add stale bot config for issue tracker maintenance --- .github/stale.yml | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) create mode 100644 .github/stale.yml diff --git a/.github/stale.yml b/.github/stale.yml new file mode 100644 index 00000000..b020f753 --- /dev/null +++ b/.github/stale.yml @@ -0,0 +1,32 @@ +# Configuration for probot/stale - https://github.com/probot/stale + +# Number of days of inactivity before an issue becomes stale +daysUntilStale: 60 + +# Number of days of inactivity before a stale issue is closed +daysUntilClose: 7 + +# Issues with these labels will never be considered stale +exemptLabels: + - pinned + - security + - good first issue + - help wanted + +# Label to use when marking an issue as stale +staleLabel: wontfix + +# Comment to post when marking an issue as stale +markComment: > + This issue has been automatically marked as stale because it has not had + recent activity. It will be closed if no further activity occurs in 7 days. + Thank you for your contributions! + +# Comment to post when closing a stale issue +closeComment: > + This issue has been automatically closed due to inactivity. + If you believe this was closed in error, please reopen the issue + or leave a comment explaining why it should remain open. + +# Limit to only issues (not PRs) +only: issues From 7321d77d9086e8b3e2475c73f57833f1900d47e1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 12:59:40 +0800 Subject: [PATCH 085/311] feat: add llms-full.txt for AI discoverability --- llms-full.txt | 367 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 367 insertions(+) create mode 100644 llms-full.txt diff --git a/llms-full.txt b/llms-full.txt new file mode 100644 index 00000000..a13ae845 --- /dev/null +++ b/llms-full.txt @@ -0,0 +1,367 @@ +# AI Models Catalog + +> Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data only. npm package available. + +## What is this? + +A machine-readable YAML catalog of every major AI model provider and their models. Every data point comes from the provider's own API or documentation, never third-party aggregators. + +## Key Stats + +- 95 providers +- 4,587 model files +- 2,712 unique model IDs +- 441 model families +- 1,306 reasoning models +- 2,350 tool-calling models +- 527 open-weight models +- 81 free models +- 1,487 vision models +- 829 structured output models +- 28 image generation models +- 118 audio input models +- 167 video input models + +## Install + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types +``` + +## Download + +```bash +# JSON — full metadata (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +## Data Format + +Each model is a single YAML file at `providers//models/.yaml`: + +```yaml +id: gpt-4.1 +name: GPT-4.1 +family: gpt-4.1 +tool_call: true +structured_output: true +pricing: + input: 2.0 + output: 8.0 + cache_read: 0.5 +limit: + context: 1047576 + output: 32768 +modalities: + input: [text, image] + output: [text] +release_date: "2026-05-18" +last_updated: "2026-05-18" +``` + +## Pricing Types + +| Type | When | Example | +| -------------- | ------------------------- | ------------------------------- | +| `TokenPricing` | Per-million-token pricing | `input: 2.5, output: 10` | +| `VideoPricing` | Per-second pricing | `unit: per_second, price: 0.03` | +| `UnitPricing` | Per-image or per-request | `unit: per_image, price: 0.04` | +| `FreePricing` | No cost | `unit: free` | + +## Covered Providers + +OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, Amazon Bedrock, Azure OpenAI, Google Vertex AI, OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, SiliconFlow, Novita AI, SambaNova, Cloudflare Workers AI, Chutes, Kluster AI, NanoGPT, and 75+ more. + +Full list: 01.AI, 302.AI, AI21 Labs, AIHubMix, AI/ML API, Aion Labs, Alibaba Cloud, Amazon Bedrock, Amazon Nova, Anthropic, Arcee AI, Auriko, Azure OpenAI, Baichuan AI, Baidu, Baseten, Berget, ByteDance, Cerebras, Chutes, Clarifai, CloudFerro Sherlock, Cloudflare Workers AI, Cohere, Cortecs, DInference, Databricks, DeepInfra, DeepSeek, DigitalOcean, evroc, FastRouter, Fireworks AI, FriendliAI, GMI Cloud, Google, Google Vertex AI, Groq, HPC-AI Cloud, Hyperbolic, IBM Granite, iFlytek SparkDesk, Inception Labs, InclusionAI, Inference.net, Kluster AI, LLM Gateway, Martian, MegaNova, Meta Llama, Microsoft Phi, MiniMax, Mistral AI, Mixlayer, MoArk AI, Moonshot AI, Morph, NanoGPT, Nebius, NeuralWatt, Nous Research, Novita AI, NVIDIA, OpenAI, OpenRouter, OrcaRouter, OVHcloud, PPIO, Perplexity, Privatemode AI, Qiniu AI, Regolo, Reka AI, Requesty, SambaNova, Sarvam AI, Scaleway, SiliconFlow, SiliconFlow CN, StepFun, SubModel, Tencent Cloud TokenHub, Tencent Hunyuan, TextSynth, Together AI, Upstage, Venice AI, Voyage AI, Vultr, Wafer, Writer, xAI Grok, Xiaomi, Zhipu AI, 接口 AI + +## Data Schema + +### Model Schema (Required Fields) + +| Field | Type | Description | Example | +| -------------- | ------- | ---------------------------------------- | ------------------------------------------ | +| `id` | string | Stable model ID (no date suffix) | `gpt-4o`, `claude-sonnet-4-5` | +| `name` | string | Display name | `GPT-4o`, `Claude Sonnet 4.5` | +| `family` | string | Model family (broad lineage) | `gpt-4o`, `claude-sonnet` | +| `pricing` | Pricing | Model pricing (see below) | — | +| `modalities` | object | Input/output modalities | `{ input: [text, image], output: [text] }` | +| `last_updated` | string | Last data update (YYYY-MM-DD or YYYY-MM) | `2024-08-06` | + +### Model Schema (Optional Fields) + +| Field | Type | Default | Description | Example | +| ------------------- | ------- | ------- | -------------------------------- | ------------------------------------ | +| `reasoning` | boolean | `false` | Supports reasoning/thinking mode | `true` | +| `temperature` | boolean | `true` | Supports temperature parameter | `false` | +| `tool_call` | boolean | `false` | Supports tool/function calling | `true` | +| `attachment` | boolean | `false` | Supports file attachments | `true` | +| `structured_output` | boolean | `false` | Supports structured/JSON output | `true` | +| `open_weights` | boolean | `false` | Open-weight model | `true` | +| `deprecated` | boolean | `false` | Deprecated but still accessible | `true` | +| `limit` | object | — | Token limits | `{ context: 128000, output: 16384 }` | +| `knowledge` | string | — | Training data cutoff | `2023-10` | +| `release_date` | string | — | Model release date | `2024-05-13` | +| `snapshots` | array | — | Dated model versions | See below | + +### Modality Types + +| Modality | Description | +| -------- | --------------------- | +| `text` | Text input or output | +| `image` | Image input or output | +| `video` | Video input | +| `audio` | Audio input or output | +| `pdf` | PDF document input | + +### Pricing Schema + +Pricing is a union of four types: + +**TokenPricing** (most common — per-million-token pricing): + +```yaml +pricing: + input: 2.5 # $/M input tokens + output: 10 # $/M output tokens + cache_write: 1.25 # optional + cache_read: 0.625 # optional +``` + +**VideoPricing** (per-second, optionally tiered by resolution): + +```yaml +pricing: + unit: per_second + price: 0.03 +``` + +**UnitPricing** (per-image or per-request): + +```yaml +pricing: + unit: per_image + price: 0.04 +``` + +**FreePricing** (no cost): + +```yaml +pricing: + unit: free +``` + +### Snapshot Schema + +Snapshots represent dated versions of a model. They inherit all parent fields and only override what differs: + +```yaml +id: gpt-4o +name: GPT-4o +snapshots: + - id: gpt-4o-2024-08-06 + last_updated: "2024-08-06" + - id: gpt-4o-2024-05-13 + deprecated: true + last_updated: "2024-05-13" +``` + +### Provider Schema + +Each provider has a `provider.yaml` file: + +| Field | Type | Required | Description | Example | +| ---------------- | ------ | -------- | ------------------------------------ | ---------------------------------- | +| `id` | string | ✅ | Provider ID (matches directory name) | `openai` | +| `name` | string | ✅ | Display name | `OpenAI` | +| `url` | string | ✅ | Official website URL | `https://openai.com` | +| `api_docs` | string | ❌ | API documentation URL | `https://platform.openai.com/docs` | +| `apis` | object | ✅ | API endpoints keyed by format | See below | +| `apis.openai` | string | ❌ | OpenAI-compatible API endpoint | `https://api.openai.com/v1` | +| `apis.anthropic` | string | ❌ | Anthropic API endpoint | — | +| `apis.google` | string | ❌ | Google AI API endpoint | — | +| `currency` | string | ❌ | Default currency (USD/CNY/EUR) | `USD` | + +## Quick Start + +### Find the cheapest model + +→ See [Pricing Comparison](docs/pricing-comparison.md) + +**Cheapest models with tool calling:** + +| Model | Provider | Input (per 1M tokens) | Output (per 1M tokens) | +| ---------------- | ------------- | --------------------: | ---------------------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-235B-A22B | Alibaba Cloud | $0.14 | $0.42 | +| Llama 4 Maverick | Together AI | $0.20 | $0.80 | + +### Find the most capable model + +→ See [Model Comparison](docs/model-comparison.md) + +**Top-tier flagships:** + +| Model | Context | Tool Call | Vision | Input $/1M | Output $/1M | +| -------------- | ------- | --------- | ------ | ---------: | ----------: | +| GPT-4.1 | 1M | ✅ | ✅ | $2.00 | $8.00 | +| Claude Opus 4 | 200K | ✅ | ✅ | $15.00 | $75.00 | +| Gemini 2.5 Pro | 1M | ✅ | ✅ | $1.25 | $10.00 | + +### Find a free model + +→ See [Free AI Models](docs/free-models.md) + +- Google Gemini 2.0 Flash (1M context, tool calling, vision, reasoning) +- 70+ models on Chutes (Llama 4, Qwen3, DeepSeek-R1, etc.) + +### Find the largest context window + +→ See [Context Window Comparison](docs/context-windows.md) + +| Model | Context Window | +| --------------- | -------------: | +| Llama 4 Scout | 10M tokens | +| Gemini 2.5 Pro | 1M tokens | +| GPT-4.1 | ~1M tokens | + +## Programmatic Access + +### npm package + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types + +// Find models with tool calling under $1/1M input +const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < 1); +``` + +### Download data files + +```bash +# JSON — full metadata (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +### Python usage + +```python +import json + +with open("models.json") as f: + catalog = json.load(f) + +# Find all reasoning models +reasoning = [m for m in catalog["models"] if m.get("reasoning")] + +# Find models with largest context windows +by_context = sorted( + catalog["models"], + key=lambda m: (m.get("limit", {}) or {}).get("context", 0), + reverse=True, +)[:10] +``` + +### JSON structure + +```json +{ + "generated_at": "2026-05-21T02:13:04.076Z", + "stats": { + "providers": 95, + "models": 4587, + "unique_model_ids": 2712, + "families": 441 + }, + "providers": { + "openai": { "name": "OpenAI", "model_count": 28 }, + "anthropic": { "name": "Anthropic", "model_count": 11 } + }, + "models": [ + { + "id": "gpt-4.1", + "name": "GPT-4.1", + "family": "gpt-4.1", + "provider": "openai", + "tool_call": true, + "structured_output": true, + "pricing": { "input": 2, "output": 8, "cache_read": 0.5 }, + "limit": { "context": 1047576, "output": 32768 }, + "modalities": { "input": ["text", "image"], "output": ["text"] } + } + ] +} +``` + +## Documentation + +- [Quick Start](docs/quick-start.md) — find the right model in 30 seconds +- [API & Programmatic Access](docs/api.md) — download models.json, code examples +- [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight +- [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers +- [Context Window Comparison](docs/context-windows.md) — largest context windows, best value +- [Free AI Models](docs/free-models.md) — 81 free models by capability +- [Open-Weight Models](docs/open-weights.md) — 513 open-weight models +- [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models +- [Tool Calling Models](docs/tool-calling.md) — 2,350 tool-calling models +- [Vision Models](docs/vision-models.md) — 1,487 vision models +- [Image Generation](docs/image-generation.md) — 28 image generation models +- [Structured Output](docs/structured-output.md) — 829 JSON-mode models +- [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video +- [Provider Overview](docs/providers.md) — all 95 providers by type and market +- [Data Schema Reference](docs/data-schema.md) — complete YAML schema +- [Data Acquisition](docs/data-acquisition.md) — how we acquire and update data +- [Design Principles](docs/lessons-learned.md) — lessons learned + +## Design Principles + +- **First-party data only** — all model data comes from the provider's own API or website +- **Dynamic discovery** — scrape functions discover models from the source, not from hardcoded lists +- **Include deprecated, exclude retired** — deprecated models are included with `deprecated: true`; retired (inaccessible) models are excluded +- **Never fabricate data** — if required data is missing, skip the model with a warning rather than filling in guessed values +- **YAML source format** — human-readable, supports comments, machine-parseable +- **Snapshot inheritance** — dated model versions are nested within the parent model, inheriting all fields + +## Adding a New Provider + +1. Create `providers//scrape.ts` with a `scrape()` function that returns `{ provider, models }` +2. Data must come from a first-party source (provider's API or website) +3. Include a discovery step — no hardcoded model ID lists +4. Run `npx tsx scripts/sync.ts ` to generate initial data +5. Validate with `npx tsx scripts/validate.ts` + +## CLI Tools + +```bash +# Validate all YAML data +npx tsx scripts/validate.ts + +# Compute catalog statistics +npx tsx scripts/stats.ts # table format +npx tsx scripts/stats.ts json # JSON format + +# Compile to models.json +npx tsx scripts/compile.ts + +# Sync data from providers +npx tsx scripts/sync.ts openai # single provider +npx tsx scripts/sync.ts # all providers + +# Export to CSV +npx tsx scripts/export-csv.ts +``` \ No newline at end of file From 6f3ddb63468afd0b6205845fa7d7899db4dea88e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:02:03 +0800 Subject: [PATCH 086/311] feat: add missing provider README.md files for 8 providers --- providers/aion/README.md | 12 ++++++++++++ providers/auriko/README.md | 12 ++++++++++++ providers/baichuan/README.md | 13 +++++++++++++ providers/cloudferro-sherlock/README.md | 13 +++++++++++++ providers/iflytek/README.md | 13 +++++++++++++ providers/llmgateway/README.md | 13 +++++++++++++ providers/martian/README.md | 13 +++++++++++++ providers/tencent-tokenhub/README.md | 13 +++++++++++++ 8 files changed, 102 insertions(+) create mode 100644 providers/aion/README.md create mode 100644 providers/auriko/README.md create mode 100644 providers/baichuan/README.md create mode 100644 providers/cloudferro-sherlock/README.md create mode 100644 providers/iflytek/README.md create mode 100644 providers/llmgateway/README.md create mode 100644 providers/martian/README.md create mode 100644 providers/tencent-tokenhub/README.md diff --git a/providers/aion/README.md b/providers/aion/README.md new file mode 100644 index 00000000..0f7281f7 --- /dev/null +++ b/providers/aion/README.md @@ -0,0 +1,12 @@ +# Aion Labs + +[Aion Labs](https://aionlabs.ai) provides AI model inference services. + +## Data Source + +Model data is fetched from the Aion Labs API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Pricing follows the provider's published rates diff --git a/providers/auriko/README.md b/providers/auriko/README.md new file mode 100644 index 00000000..d0d46dd5 --- /dev/null +++ b/providers/auriko/README.md @@ -0,0 +1,12 @@ +# Auriko + +[Auriko](https://auriko.com) provides AI model inference services. + +## Data Source + +Model data is fetched from the Auriko API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Pricing follows the provider's published rates diff --git a/providers/baichuan/README.md b/providers/baichuan/README.md new file mode 100644 index 00000000..0ab56006 --- /dev/null +++ b/providers/baichuan/README.md @@ -0,0 +1,13 @@ +# Baichuan AI + +[Baichuan AI](https://platform.baichuan-ai.com) is a Chinese AI model producer, developer of the Baichuan series of large language models. + +## Data Source + +Model data is fetched from the Baichuan AI API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Pricing is in CNY +- Baichuan models include both general-purpose and specialized variants diff --git a/providers/cloudferro-sherlock/README.md b/providers/cloudferro-sherlock/README.md new file mode 100644 index 00000000..efc94f84 --- /dev/null +++ b/providers/cloudferro-sherlock/README.md @@ -0,0 +1,13 @@ +# CloudFerro Sherlock + +[CloudFerro Sherlock](https://sherlock.cloudferro.com) provides AI model inference on European cloud infrastructure. + +## Data Source + +Model data is fetched from the CloudFerro Sherlock API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Pricing follows the provider's published rates +- Hosted on European infrastructure (GDPR-friendly) diff --git a/providers/iflytek/README.md b/providers/iflytek/README.md new file mode 100644 index 00000000..baa4b6e7 --- /dev/null +++ b/providers/iflytek/README.md @@ -0,0 +1,13 @@ +# iFlytek SparkDesk + +[iFlytek SparkDesk](https://xinghuo.xfyun.cn) is a Chinese AI model platform by iFlytek (科大讯飞), providing the SparkDesk (星火) series of large language models. + +## Data Source + +Model data is fetched from the iFlytek SparkDesk API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Pricing is in CNY +- SparkDesk models support both Chinese and English diff --git a/providers/llmgateway/README.md b/providers/llmgateway/README.md new file mode 100644 index 00000000..f04c7505 --- /dev/null +++ b/providers/llmgateway/README.md @@ -0,0 +1,13 @@ +# LLM Gateway + +[LLM Gateway](https://llmgateway.com) provides a unified API gateway for accessing multiple AI model providers. + +## Data Source + +Model data is fetched from the LLM Gateway API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Aggregates models from multiple upstream providers +- Pricing follows the provider's published rates diff --git a/providers/martian/README.md b/providers/martian/README.md new file mode 100644 index 00000000..3a21898a --- /dev/null +++ b/providers/martian/README.md @@ -0,0 +1,13 @@ +# Martian + +[Martian](https://withmartian.com) provides AI model inference services with a unified API. + +## Data Source + +Model data is fetched from the Martian API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Aggregates models from multiple upstream providers +- Pricing follows the provider's published rates diff --git a/providers/tencent-tokenhub/README.md b/providers/tencent-tokenhub/README.md new file mode 100644 index 00000000..b32611e0 --- /dev/null +++ b/providers/tencent-tokenhub/README.md @@ -0,0 +1,13 @@ +# Tencent Cloud TokenHub + +[Tencent Cloud TokenHub](https://cloud.tencent.com/product/ti) is a Chinese AI model platform by Tencent, providing access to various large language models. + +## Data Source + +Model data is fetched from the Tencent Cloud TokenHub API endpoint. + +## Notes + +- Models are discovered dynamically from the API +- Pricing is in CNY +- Provides access to both Tencent's own models and third-party models From 8f7aac4e43c35f01c2c57aebf9ee5f6bccb14d8c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:02:53 +0800 Subject: [PATCH 087/311] feat: add bilingual language links to all documentation pages --- docs/api.md | 2 ++ docs/context-windows.md | 2 ++ docs/free-models.md | 2 ++ docs/image-generation.md | 2 ++ docs/modality-matrix.md | 2 ++ docs/quick-start.md | 2 ++ docs/structured-output.md | 2 ++ docs/tool-calling.md | 2 ++ docs/vision-models.md | 2 ++ docs/zh/api.md | 2 ++ docs/zh/context-windows.md | 2 ++ docs/zh/free-models.md | 2 ++ docs/zh/image-generation.md | 2 ++ docs/zh/modality-matrix.md | 2 ++ docs/zh/quick-start.md | 2 ++ docs/zh/structured-output.md | 2 ++ docs/zh/tool-calling.md | 2 ++ docs/zh/vision-models.md | 2 ++ 18 files changed, 36 insertions(+) diff --git a/docs/api.md b/docs/api.md index a28196df..12760da0 100644 --- a/docs/api.md +++ b/docs/api.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/api.md) + # API & Programmatic Access Use the catalog data in your applications. diff --git a/docs/context-windows.md b/docs/context-windows.md index 5aec0a0c..525dcff7 100644 --- a/docs/context-windows.md +++ b/docs/context-windows.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/context-windows.md) + # Context Window Comparison Which models have the largest context windows? This page lists models by context window size and pricing. diff --git a/docs/free-models.md b/docs/free-models.md index da24c017..0caf6759 100644 --- a/docs/free-models.md +++ b/docs/free-models.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/free-models.md) + # Free AI Models 81 models in this catalog are free to use. This page lists them by capability so you can find the right free model for your project. diff --git a/docs/image-generation.md b/docs/image-generation.md index 9e5c3e32..bf049bb0 100644 --- a/docs/image-generation.md +++ b/docs/image-generation.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/image-generation.md) + # AI Image Generation Models 28 models in this catalog can generate images (image output modality). This page covers text-to-image, image editing, and multimodal generation models. diff --git a/docs/modality-matrix.md b/docs/modality-matrix.md index b9aa5998..dc2fff3c 100644 --- a/docs/modality-matrix.md +++ b/docs/modality-matrix.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/modality-matrix.md) + # Modality Matrix Which models support vision, audio, image generation, and video? This page lists the top models for each modality. diff --git a/docs/quick-start.md b/docs/quick-start.md index b2a64dbb..8000da43 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/quick-start.md) + # Quick Start Guide Find the right AI model for your needs in 30 seconds. diff --git a/docs/structured-output.md b/docs/structured-output.md index 812e5af3..cdaef929 100644 --- a/docs/structured-output.md +++ b/docs/structured-output.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/structured-output.md) + # AI Structured Output Models (JSON Mode) 829 models in this catalog support structured output — the ability to generate responses that conform to a specified JSON schema. This is essential for building reliable AI-powered APIs, data pipelines, and automation. diff --git a/docs/tool-calling.md b/docs/tool-calling.md index 16a376cc..08679aaa 100644 --- a/docs/tool-calling.md +++ b/docs/tool-calling.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/tool-calling.md) + # AI Tool Calling Models 2,350 models in this catalog support tool calling (function calling). This page highlights the most capable and cost-effective models for building AI agents and automation. diff --git a/docs/vision-models.md b/docs/vision-models.md index 98157051..1fdca6fa 100644 --- a/docs/vision-models.md +++ b/docs/vision-models.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/vision-models.md) + # AI Vision Models 1,487 models in this catalog accept image input (vision). This page highlights the most capable and cost-effective vision models for image understanding, document analysis, and visual reasoning. diff --git a/docs/zh/api.md b/docs/zh/api.md index 8e420152..2d216526 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -1,3 +1,5 @@ +[English](../api.md) | **中文** + # API 与编程访问 在你的应用中使用目录数据。 diff --git a/docs/zh/context-windows.md b/docs/zh/context-windows.md index 4547cd1f..5a194d82 100644 --- a/docs/zh/context-windows.md +++ b/docs/zh/context-windows.md @@ -1,3 +1,5 @@ +[English](../context-windows.md) | **中文** + # 上下文窗口对比 哪些模型拥有最大的上下文窗口？本页按上下文窗口大小和定价列出模型。 diff --git a/docs/zh/free-models.md b/docs/zh/free-models.md index 5253fdc9..8c0dc0ef 100644 --- a/docs/zh/free-models.md +++ b/docs/zh/free-models.md @@ -1,3 +1,5 @@ +[English](../free-models.md) | **中文** + # 免费 AI 模型 本目录中有 81 个模型可免费使用。本页按能力分类列出，帮助你找到适合项目的免费模型。 diff --git a/docs/zh/image-generation.md b/docs/zh/image-generation.md index 07af5e2c..4fd17c1d 100644 --- a/docs/zh/image-generation.md +++ b/docs/zh/image-generation.md @@ -1,3 +1,5 @@ +[English](../image-generation.md) | **中文** + # AI 图像生成模型 本目录中有 28 个模型可以生成图像（图像输出模态）。本页涵盖文本生成图像、图像编辑和多模态生成模型。 diff --git a/docs/zh/modality-matrix.md b/docs/zh/modality-matrix.md index 91cea9fb..91d77e07 100644 --- a/docs/zh/modality-matrix.md +++ b/docs/zh/modality-matrix.md @@ -1,3 +1,5 @@ +[English](../modality-matrix.md) | **中文** + # 模态矩阵 哪些模型支持视觉、音频、图像生成和视频？本页列出各模态的顶级模型。 diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md index 3501f6a1..830e206b 100644 --- a/docs/zh/quick-start.md +++ b/docs/zh/quick-start.md @@ -1,3 +1,5 @@ +[English](../quick-start.md) | **中文** + # 快速入门指南 30 秒内找到适合你需求的 AI 模型。 diff --git a/docs/zh/structured-output.md b/docs/zh/structured-output.md index 0f26dcc4..10bc2bf6 100644 --- a/docs/zh/structured-output.md +++ b/docs/zh/structured-output.md @@ -1,3 +1,5 @@ +[English](../structured-output.md) | **中文** + # AI 结构化输出模型（JSON 模式） 本目录中有 829 个模型支持结构化输出 — 能够按照指定的 JSON Schema 生成回复。这对于构建可靠的 AI 驱动 API、数据管道和自动化至关重要。 diff --git a/docs/zh/tool-calling.md b/docs/zh/tool-calling.md index 8abd7e78..e37cb9ad 100644 --- a/docs/zh/tool-calling.md +++ b/docs/zh/tool-calling.md @@ -1,3 +1,5 @@ +[English](../tool-calling.md) | **中文** + # AI 工具调用模型 本目录中有 2,350 个模型支持工具调用（函数调用）。本页重点介绍构建 AI 代理和自动化最有能力和最具性价比的模型。 diff --git a/docs/zh/vision-models.md b/docs/zh/vision-models.md index 52de56f1..06e6fc5c 100644 --- a/docs/zh/vision-models.md +++ b/docs/zh/vision-models.md @@ -1,3 +1,5 @@ +[English](../vision-models.md) | **中文** + # AI 视觉模型 本目录中有 1,487 个模型接受图像输入（视觉）。本页重点介绍图像理解、文档分析和视觉推理最有能力和最具性价比的模型。 From 07e72484b0d7106fc91482e319793e33e005d5f7 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:03:56 +0800 Subject: [PATCH 088/311] feat: add bilingual language links to open-weights and reasoning-models docs --- docs/open-weights.md | 2 ++ docs/reasoning-models.md | 2 ++ docs/zh/open-weights.md | 2 ++ docs/zh/reasoning-models.md | 2 ++ 4 files changed, 8 insertions(+) diff --git a/docs/open-weights.md b/docs/open-weights.md index db4f193c..58785975 100644 --- a/docs/open-weights.md +++ b/docs/open-weights.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/open-weights.md) + # Open-Weight AI Models 513 models in this catalog are open-weight — you can download the weights and run them on your own infrastructure. This page highlights the most capable and widely available open-weight models. diff --git a/docs/reasoning-models.md b/docs/reasoning-models.md index 32380292..e49521d3 100644 --- a/docs/reasoning-models.md +++ b/docs/reasoning-models.md @@ -1,3 +1,5 @@ +**English** | [中文](./zh/reasoning-models.md) + # AI Reasoning Models 1,306 models in this catalog support reasoning (chain-of-thought / extended thinking). This page highlights the most capable and cost-effective reasoning models available. diff --git a/docs/zh/open-weights.md b/docs/zh/open-weights.md index 788f3fcf..fc1d0872 100644 --- a/docs/zh/open-weights.md +++ b/docs/zh/open-weights.md @@ -1,3 +1,5 @@ +[English](../open-weights.md) | **中文** + # 开源权重 AI 模型 本目录中有 513 个开源权重模型 — 你可以下载权重并在自己的基础设施上运行。本页重点介绍最有能力和最广泛可用的开源权重模型。 diff --git a/docs/zh/reasoning-models.md b/docs/zh/reasoning-models.md index 0aba021d..276bc175 100644 --- a/docs/zh/reasoning-models.md +++ b/docs/zh/reasoning-models.md @@ -1,3 +1,5 @@ +[English](../reasoning-models.md) | **中文** + # AI 推理模型 本目录中有 1,306 个模型支持推理（链式思维 / 扩展思考）。本页重点介绍最有能力和最具性价比的推理模型。 From 992265ac1f734f308dde509ae0ba6a2da6b44535 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:04:45 +0800 Subject: [PATCH 089/311] feat: add llms-full.txt reference in llms.txt --- llms.txt | 2 ++ 1 file changed, 2 insertions(+) diff --git a/llms.txt b/llms.txt index fe129a6a..848a525f 100644 --- a/llms.txt +++ b/llms.txt @@ -2,6 +2,8 @@ > Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data only. +See [llms-full.txt](./llms-full.txt) for the complete data schema, pricing types, provider list, and usage examples. + ## What is this? A machine-readable YAML catalog of every major AI model provider and their models. Every data point comes from the provider's own API or documentation, never third-party aggregators. From f66a808561f9d72a45dab766c945c3fbb3f39221 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:04:45 +0800 Subject: [PATCH 090/311] feat: add llms-full.txt reference in llms.txt --- .github/ISSUE_TEMPLATE/feature_request.md | 27 ++ AGENTS.md | 1 + CITATION.cff | 23 ++ README.md | 2 + docs/code-examples.md | 352 ++++++++++++++++++++++ docs/zh/code-examples.md | 352 ++++++++++++++++++++++ llms-full.txt | 1 + llms.txt | 1 + 8 files changed, 759 insertions(+) create mode 100644 .github/ISSUE_TEMPLATE/feature_request.md create mode 100644 CITATION.cff create mode 100644 docs/code-examples.md create mode 100644 docs/zh/code-examples.md diff --git a/.github/ISSUE_TEMPLATE/feature_request.md b/.github/ISSUE_TEMPLATE/feature_request.md new file mode 100644 index 00000000..162dcb55 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/feature_request.md @@ -0,0 +1,27 @@ +--- +name: Feature request +about: Suggest a new feature, documentation page, or capability +title: "[Feature] " +labels: ["enhancement"] +assignees: "" +--- + +## Problem + + + +## Proposed Solution + + + +## Alternatives Considered + + + +## Additional Context + + + +## Would you be willing to submit a PR? + +- [ ] Yes, I'd like to contribute this feature diff --git a/AGENTS.md b/AGENTS.md index 7b1f9aa7..33e7a0e5 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -25,6 +25,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) - [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) - [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) +- [`docs/code-examples.md`](docs/code-examples.md) — code examples in multiple languages ([中文](docs/zh/code-examples.md)) ## Key Design Decisions diff --git a/CITATION.cff b/CITATION.cff new file mode 100644 index 00000000..26fab1b0 --- /dev/null +++ b/CITATION.cff @@ -0,0 +1,23 @@ +cff-version: 1.2.0 +message: "If you use this catalog in your research, please cite it as below." +title: "AI Models Catalog" +type: dataset +authors: + - given-names: "i-need-token" +repository-code: "https://github.com/i-need-token/ai-models" +url: "https://github.com/i-need-token/ai-models" +abstract: > + A structured YAML catalog of 4,587 AI models across 95 providers, + including pricing, context windows, modalities, and capabilities. + All data sourced from first-party APIs and official documentation. +keywords: + - ai-models + - llm + - pricing + - model-catalog + - yaml + - structured-data + - open-data +license: MIT +version: 0.1.0 +date-released: "2026-05-21" diff --git a/README.md b/README.md index de8f1036..12bfd669 100644 --- a/README.md +++ b/README.md @@ -312,6 +312,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | +| [Code Examples](docs/code-examples.md) | Practical examples in TypeScript, Python, Go, Rust, jq | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | @@ -333,6 +334,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | | [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | | [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | | [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | | [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | | [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | diff --git a/docs/code-examples.md b/docs/code-examples.md new file mode 100644 index 00000000..6ecd99b7 --- /dev/null +++ b/docs/code-examples.md @@ -0,0 +1,352 @@ +# Code Examples + +**English** | [中文](./zh/code-examples.md) + +Practical code examples for working with the AI Models Catalog data in multiple languages. + +## Download the Data + +```bash +# JSON — full metadata (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +## TypeScript / JavaScript + +### Install the npm package + +```bash +npm install ai-models +``` + +### Basic usage + +```typescript +import catalog from "ai-models"; +import type { Model } from "ai-models"; + +// Total number of models +console.log(`Total models: ${catalog.models.length}`); + +// Find models by provider +const openaiModels = catalog.models.filter((m) => m.provider === "openai"); +console.log(`OpenAI models: ${openaiModels.length}`); +``` + +### Find the cheapest tool-calling models + +```typescript +import catalog from "ai-models"; + +const affordable = catalog.models + .filter((m) => m.tool_call && m.pricing?.input && m.pricing.input < 1) + .sort((a, b) => a.pricing.input - b.pricing.input) + .slice(0, 10); + +for (const m of affordable) { + console.log(`${m.name}: $${m.pricing.input}/1M input, $${m.pricing.output}/1M output`); +} +``` + +### Find models with the largest context windows + +```typescript +import catalog from "ai-models"; + +const largest = catalog.models + .filter((m) => m.limit?.context) + .sort((a, b) => b.limit.context - a.limit.context) + .slice(0, 10); + +for (const m of largest) { + console.log(`${m.name}: ${(m.limit.context / 1_000_000).toFixed(1)}M context`); +} +``` + +### Find free models with vision + +```typescript +import catalog from "ai-models"; + +const freeVision = catalog.models.filter( + (m) => m.pricing?.unit === "free" && m.modalities?.input?.includes("image"), +); + +console.log(`Free vision models: ${freeVision.length}`); +for (const m of freeVision) { + console.log(`- ${m.name} (${m.provider})`); +} +``` + +## Python + +### Using the JSON data + +```python +import json +import urllib.request + +# Download the latest data +url = "https://github.com/i-need-token/ai-models/releases/latest/download/models.json" +urllib.request.urlretrieve(url, "models.json") + +with open("models.json") as f: + catalog = json.load(f) + +print(f"Total models: {len(catalog['models'])}") +``` + +### Find reasoning models under $5/1M output + +```python +reasoning_cheap = [ + m for m in catalog["models"] + if m.get("reasoning") + and m.get("pricing", {}).get("output") + and m["pricing"]["output"] < 5 +] + +for m in sorted(reasoning_cheap, key=lambda x: x["pricing"]["output"]): + print(f"{m['name']}: ${m['pricing']['output']}/1M output") +``` + +### Using pandas with the CSV + +```python +import pandas as pd + +df = pd.read_csv("https://github.com/i-need-token/ai-models/releases/latest/download/models.csv") + +# Filter and sort +tool_calling = df[df["tool_call"] == True].sort_values("pricing_input") +print(tool_calling[["name", "provider", "pricing_input", "pricing_output"]].head(10)) +``` + +### Find open-weight models with tool calling + +```python +open_tool = [ + m for m in catalog["models"] + if m.get("open_weights") and m.get("tool_call") +] + +print(f"Open-weight models with tool calling: {len(open_tool)}") +for m in open_tool[:10]: + print(f" - {m['name']} ({m['provider']})") +``` + +## Go + +```go +package main + +import ( + "encoding/json" + "fmt" + "net/http" +) + +type Catalog struct { + Models []Model `json:"models"` +} + +type Model struct { + ID string `json:"id"` + Name string `json:"name"` + Provider string `json:"provider"` + ToolCall bool `json:"tool_call"` + Pricing Pricing `json:"pricing"` +} + +type Pricing struct { + Input float64 `json:"input"` + Output float64 `json:"output"` +} + +func main() { + resp, err := http.Get("https://github.com/i-need-token/ai-models/releases/latest/download/models.json") + if err != nil { + panic(err) + } + defer resp.Body.Close() + + var catalog Catalog + json.NewDecoder(resp.Body).Decode(&catalog) + + fmt.Printf("Total models: %d\n", len(catalog.Models)) + + // Find tool-calling models under $1/1M input + for _, m := range catalog.Models { + if m.ToolCall && m.Pricing.Input > 0 && m.Pricing.Input < 1 { + fmt.Printf("%s: $%.2f/1M input\n", m.Name, m.Pricing.Input) + } + } +} +``` + +## Rust + +```rust +use serde::Deserialize; + +#[derive(Deserialize)] +struct Catalog { + models: Vec, +} + +#[derive(Deserialize)] +struct Model { + id: String, + name: String, + provider: String, + #[serde(default)] + tool_call: bool, + pricing: Option, +} + +#[derive(Deserialize)] +struct Pricing { + input: f64, + output: f64, +} + +fn main() -> Result<(), Box> { + let data = reqwest::blocking::get( + "https://github.com/i-need-token/ai-models/releases/latest/download/models.json" + )?.text()?; + + let catalog: Catalog = serde_json::from_str(&data)?; + println!("Total models: {}", catalog.models.len()); + + // Find reasoning models + let reasoning: Vec<_> = catalog.models.iter() + .filter(|m| m.tool_call) + .collect(); + + println!("Tool-calling models: {}", reasoning.len()); + Ok(()) +} +``` + +## Shell / jq + +```bash +# Download the data +curl -sLO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# Count total models +jq '.models | length' models.json + +# Find all OpenAI models +jq '.models[] | select(.provider == "openai") | .name' models.json + +# Find the cheapest models with tool calling +jq '[.models[] | select(.tool_call == true and .pricing.input != null)] | sort_by(.pricing.input) | .[:5] | .[] | {name, provider, input: .pricing.input}' models.json + +# Find free models +jq '[.models[] | select(.pricing.unit == "free")] | length' models.json + +# List all providers +jq '.providers | keys' models.json +``` + +## Excel / Google Sheets + +1. Download the CSV: `https://github.com/i-need-token/ai-models/releases/latest/download/models.csv` +2. Open in Excel or import into Google Sheets +3. Use filters to find models by provider, capability, or price range + +## Common Queries + +### Find the best model for coding + +```typescript +import catalog from "ai-models"; + +const codingModels = catalog.models.filter( + (m) => + m.tool_call && + m.structured_output && + m.limit?.context >= 128000 && + m.pricing?.input && + m.pricing.input <= 5, +); + +// Sort by context window (descending), then price (ascending) +codingModels.sort((a, b) => { + const ctxDiff = (b.limit?.context ?? 0) - (a.limit?.context ?? 0); + if (ctxDiff !== 0) return ctxDiff; + return (a.pricing?.input ?? 0) - (b.pricing?.input ?? 0); +}); +``` + +### Compare pricing across providers for the same model family + +```typescript +import catalog from "ai-models"; + +// Group by family +const families = new Map(); +for (const m of catalog.models) { + if (!m.family) continue; + const list = families.get(m.family) ?? []; + list.push(m); + families.set(m.family, list); +} + +// Find families available on multiple providers +for (const [family, models] of families) { + const providers = new Set(models.map((m) => m.provider)); + if (providers.size > 1) { + console.log(`\n${family}:`); + for (const m of models) { + console.log(` ${m.provider}: $${m.pricing?.input}/1M in, $${m.pricing?.output}/1M out`); + } + } +} +``` + +### Build a model selector for your app + +```typescript +import catalog from "ai-models"; +import type { Model } from "ai-models"; + +interface ModelRequirements { + toolCall?: boolean; + vision?: boolean; + reasoning?: boolean; + structuredOutput?: boolean; + minContext?: number; + maxInputPrice?: number; + maxOutputPrice?: number; + openWeights?: boolean; + provider?: string; +} + +function findModels(req: ModelRequirements): Model[] { + return catalog.models.filter((m) => { + if (req.toolCall && !m.tool_call) return false; + if (req.vision && !m.modalities?.input?.includes("image")) return false; + if (req.reasoning && !m.reasoning) return false; + if (req.structuredOutput && !m.structured_output) return false; + if (req.minContext && (m.limit?.context ?? 0) < req.minContext) return false; + if (req.maxInputPrice && (m.pricing?.input ?? Infinity) > req.maxInputPrice) return false; + if (req.maxOutputPrice && (m.pricing?.output ?? Infinity) > req.maxOutputPrice) return false; + if (req.openWeights && !m.open_weights) return false; + if (req.provider && m.provider !== req.provider) return false; + return true; + }); +} + +// Example: Find a cheap vision model with tool calling +const results = findModels({ + vision: true, + toolCall: true, + maxInputPrice: 1, + maxOutputPrice: 5, +}); +``` diff --git a/docs/zh/code-examples.md b/docs/zh/code-examples.md new file mode 100644 index 00000000..9e8256f3 --- /dev/null +++ b/docs/zh/code-examples.md @@ -0,0 +1,352 @@ +# 代码示例 + +[English](../code-examples.md) | **中文** + +多种语言使用 AI Models Catalog 数据的实用代码示例。 + +## 下载数据 + +```bash +# JSON — 完整元数据 (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — 适用于 Excel/Google Sheets 的平面表格 (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +## TypeScript / JavaScript + +### 安装 npm 包 + +```bash +npm install ai-models +``` + +### 基本用法 + +```typescript +import catalog from "ai-models"; +import type { Model } from "ai-models"; + +// 模型总数 +console.log(`Total models: ${catalog.models.length}`); + +// 按提供商查找模型 +const openaiModels = catalog.models.filter((m) => m.provider === "openai"); +console.log(`OpenAI models: ${openaiModels.length}`); +``` + +### 查找最便宜的工具调用模型 + +```typescript +import catalog from "ai-models"; + +const affordable = catalog.models + .filter((m) => m.tool_call && m.pricing?.input && m.pricing.input < 1) + .sort((a, b) => a.pricing.input - b.pricing.input) + .slice(0, 10); + +for (const m of affordable) { + console.log(`${m.name}: $${m.pricing.input}/1M input, $${m.pricing.output}/1M output`); +} +``` + +### 查找最大上下文窗口的模型 + +```typescript +import catalog from "ai-models"; + +const largest = catalog.models + .filter((m) => m.limit?.context) + .sort((a, b) => b.limit.context - a.limit.context) + .slice(0, 10); + +for (const m of largest) { + console.log(`${m.name}: ${(m.limit.context / 1_000_000).toFixed(1)}M context`); +} +``` + +### 查找支持视觉的免费模型 + +```typescript +import catalog from "ai-models"; + +const freeVision = catalog.models.filter( + (m) => m.pricing?.unit === "free" && m.modalities?.input?.includes("image"), +); + +console.log(`Free vision models: ${freeVision.length}`); +for (const m of freeVision) { + console.log(`- ${m.name} (${m.provider})`); +} +``` + +## Python + +### 使用 JSON 数据 + +```python +import json +import urllib.request + +# 下载最新数据 +url = "https://github.com/i-need-token/ai-models/releases/latest/download/models.json" +urllib.request.urlretrieve(url, "models.json") + +with open("models.json") as f: + catalog = json.load(f) + +print(f"Total models: {len(catalog['models'])}") +``` + +### 查找输出价格低于 $5/1M 的推理模型 + +```python +reasoning_cheap = [ + m for m in catalog["models"] + if m.get("reasoning") + and m.get("pricing", {}).get("output") + and m["pricing"]["output"] < 5 +] + +for m in sorted(reasoning_cheap, key=lambda x: x["pricing"]["output"]): + print(f"{m['name']}: ${m['pricing']['output']}/1M output") +``` + +### 使用 pandas 处理 CSV + +```python +import pandas as pd + +df = pd.read_csv("https://github.com/i-need-token/ai-models/releases/latest/download/models.csv") + +# 筛选和排序 +tool_calling = df[df["tool_call"] == True].sort_values("pricing_input") +print(tool_calling[["name", "provider", "pricing_input", "pricing_output"]].head(10)) +``` + +### 查找支持工具调用的开源模型 + +```python +open_tool = [ + m for m in catalog["models"] + if m.get("open_weights") and m.get("tool_call") +] + +print(f"Open-weight models with tool calling: {len(open_tool)}") +for m in open_tool[:10]: + print(f" - {m['name']} ({m['provider']})") +``` + +## Go + +```go +package main + +import ( + "encoding/json" + "fmt" + "net/http" +) + +type Catalog struct { + Models []Model `json:"models"` +} + +type Model struct { + ID string `json:"id"` + Name string `json:"name"` + Provider string `json:"provider"` + ToolCall bool `json:"tool_call"` + Pricing Pricing `json:"pricing"` +} + +type Pricing struct { + Input float64 `json:"input"` + Output float64 `json:"output"` +} + +func main() { + resp, err := http.Get("https://github.com/i-need-token/ai-models/releases/latest/download/models.json") + if err != nil { + panic(err) + } + defer resp.Body.Close() + + var catalog Catalog + json.NewDecoder(resp.Body).Decode(&catalog) + + fmt.Printf("Total models: %d\n", len(catalog.Models)) + + // 查找输入价格低于 $1/1M 的工具调用模型 + for _, m := range catalog.Models { + if m.ToolCall && m.Pricing.Input > 0 && m.Pricing.Input < 1 { + fmt.Printf("%s: $%.2f/1M input\n", m.Name, m.Pricing.Input) + } + } +} +``` + +## Rust + +```rust +use serde::Deserialize; + +#[derive(Deserialize)] +struct Catalog { + models: Vec, +} + +#[derive(Deserialize)] +struct Model { + id: String, + name: String, + provider: String, + #[serde(default)] + tool_call: bool, + pricing: Option, +} + +#[derive(Deserialize)] +struct Pricing { + input: f64, + output: f64, +} + +fn main() -> Result<(), Box> { + let data = reqwest::blocking::get( + "https://github.com/i-need-token/ai-models/releases/latest/download/models.json" + )?.text()?; + + let catalog: Catalog = serde_json::from_str(&data)?; + println!("Total models: {}", catalog.models.len()); + + // 查找推理模型 + let reasoning: Vec<_> = catalog.models.iter() + .filter(|m| m.tool_call) + .collect(); + + println!("Tool-calling models: {}", reasoning.len()); + Ok(()) +} +``` + +## Shell / jq + +```bash +# 下载数据 +curl -sLO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# 统计模型总数 +jq '.models | length' models.json + +# 查找所有 OpenAI 模型 +jq '.models[] | select(.provider == "openai") | .name' models.json + +# 查找最便宜的工具调用模型 +jq '[.models[] | select(.tool_call == true and .pricing.input != null)] | sort_by(.pricing.input) | .[:5] | .[] | {name, provider, input: .pricing.input}' models.json + +# 查找免费模型 +jq '[.models[] | select(.pricing.unit == "free")] | length' models.json + +# 列出所有提供商 +jq '.providers | keys' models.json +``` + +## Excel / Google Sheets + +1. 下载 CSV：`https://github.com/i-need-token/ai-models/releases/latest/download/models.csv` +2. 在 Excel 中打开或导入 Google Sheets +3. 使用筛选器按提供商、能力或价格范围查找模型 + +## 常见查询 + +### 查找最适合编程的模型 + +```typescript +import catalog from "ai-models"; + +const codingModels = catalog.models.filter( + (m) => + m.tool_call && + m.structured_output && + m.limit?.context >= 128000 && + m.pricing?.input && + m.pricing.input <= 5, +); + +// 按上下文窗口（降序）排序，然后按价格（升序）排序 +codingModels.sort((a, b) => { + const ctxDiff = (b.limit?.context ?? 0) - (a.limit?.context ?? 0); + if (ctxDiff !== 0) return ctxDiff; + return (a.pricing?.input ?? 0) - (b.pricing?.input ?? 0); +}); +``` + +### 比较同一模型系列在不同提供商的价格 + +```typescript +import catalog from "ai-models"; + +// 按系列分组 +const families = new Map(); +for (const m of catalog.models) { + if (!m.family) continue; + const list = families.get(m.family) ?? []; + list.push(m); + families.set(m.family, list); +} + +// 查找在多个提供商上可用的系列 +for (const [family, models] of families) { + const providers = new Set(models.map((m) => m.provider)); + if (providers.size > 1) { + console.log(`\n${family}:`); + for (const m of models) { + console.log(` ${m.provider}: $${m.pricing?.input}/1M in, $${m.pricing?.output}/1M out`); + } + } +} +``` + +### 为你的应用构建模型选择器 + +```typescript +import catalog from "ai-models"; +import type { Model } from "ai-models"; + +interface ModelRequirements { + toolCall?: boolean; + vision?: boolean; + reasoning?: boolean; + structuredOutput?: boolean; + minContext?: number; + maxInputPrice?: number; + maxOutputPrice?: number; + openWeights?: boolean; + provider?: string; +} + +function findModels(req: ModelRequirements): Model[] { + return catalog.models.filter((m) => { + if (req.toolCall && !m.tool_call) return false; + if (req.vision && !m.modalities?.input?.includes("image")) return false; + if (req.reasoning && !m.reasoning) return false; + if (req.structuredOutput && !m.structured_output) return false; + if (req.minContext && (m.limit?.context ?? 0) < req.minContext) return false; + if (req.maxInputPrice && (m.pricing?.input ?? Infinity) > req.maxInputPrice) return false; + if (req.maxOutputPrice && (m.pricing?.output ?? Infinity) > req.maxOutputPrice) return false; + if (req.openWeights && !m.open_weights) return false; + if (req.provider && m.provider !== req.provider) return false; + return true; + }); +} + +// 示例：查找支持视觉和工具调用的便宜模型 +const results = findModels({ + vision: true, + toolCall: true, + maxInputPrice: 1, + maxOutputPrice: 5, +}); +``` diff --git a/llms-full.txt b/llms-full.txt index a13ae845..79a87e58 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -312,6 +312,7 @@ by_context = sorted( - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds - [API & Programmatic Access](docs/api.md) — download models.json, code examples +- [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value diff --git a/llms.txt b/llms.txt index 848a525f..1e1ec33d 100644 --- a/llms.txt +++ b/llms.txt @@ -67,6 +67,7 @@ modalities: - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds - [API & Programmatic Access](docs/api.md) — download models.json, code examples +- [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value From 8d2a672b0ffe4492869943036d5eb2c9241e081b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:12:09 +0800 Subject: [PATCH 091/311] feat: add Roadmap section to README --- README.md | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/README.md b/README.md index 12bfd669..50251f85 100644 --- a/README.md +++ b/README.md @@ -369,6 +369,14 @@ Contributions are welcome! Whether it's adding a new provider, fixing data, or i Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. +## Roadmap + +- 🔜 **More providers** — continuously adding new AI model providers +- 🔜 **REST API** — hosted API for querying the catalog +- 🔜 **Historical pricing** — track pricing changes over time +- 🔜 **Model benchmarks** — integrate performance benchmarks from standard benchmarks +- 🔜 **Community scrapers** — enable community-contributed scrape scripts with automated validation + ## Who's Using This? Built something with this catalog? [Open a PR](https://github.com/i-need-token/ai-models/edit/main/README.md) to add your project! From a8e1d992ec3ddc3d3e84a2b803db154cc2e8c82d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:13:47 +0800 Subject: [PATCH 092/311] feat: add JSON Schema for YAML model data validation --- schema.json | 281 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 281 insertions(+) create mode 100644 schema.json diff --git a/schema.json b/schema.json new file mode 100644 index 00000000..477fe74d --- /dev/null +++ b/schema.json @@ -0,0 +1,281 @@ +{ + "$schema": "https://json-schema.org/draft/2020-12/schema", + "$id": "https://raw.githubusercontent.com/i-need-token/ai-models/main/schema.json", + "title": "AI Models Catalog", + "description": "Schema for the AI Models Catalog YAML data files", + "definitions": { + "TokenPricing": { + "type": "object", + "description": "Per-million-token pricing (most common)", + "properties": { + "input": { + "type": "number", + "minimum": 0, + "description": "Price per 1M input tokens (USD or CNY)" + }, + "output": { + "type": "number", + "minimum": 0, + "description": "Price per 1M output tokens (USD or CNY)" + }, + "cache_write": { + "type": "number", + "minimum": 0, + "description": "Price per 1M cached write tokens" + }, + "cache_read": { + "type": "number", + "minimum": 0, + "description": "Price per 1M cached read tokens" + } + }, + "required": ["input", "output"] + }, + "VideoPricing": { + "type": "object", + "description": "Per-second video pricing", + "properties": { + "unit": { + "type": "string", + "const": "per_second", + "description": "Pricing unit" + }, + "price": { + "type": "number", + "minimum": 0, + "description": "Price per second" + } + }, + "required": ["unit", "price"] + }, + "UnitPricing": { + "type": "object", + "description": "Per-unit pricing (per image, per request, etc.)", + "properties": { + "unit": { + "type": "string", + "enum": ["per_image", "per_request"], + "description": "Pricing unit" + }, + "price": { + "type": "number", + "minimum": 0, + "description": "Price per unit" + } + }, + "required": ["unit", "price"] + }, + "FreePricing": { + "type": "object", + "description": "Free pricing (no cost)", + "properties": { + "unit": { + "type": "string", + "const": "free", + "description": "Pricing unit" + } + }, + "required": ["unit"] + }, + "Pricing": { + "oneOf": [ + { "$ref": "#/definitions/TokenPricing" }, + { "$ref": "#/definitions/VideoPricing" }, + { "$ref": "#/definitions/UnitPricing" }, + { "$ref": "#/definitions/FreePricing" } + ] + }, + "Modality": { + "type": "string", + "enum": ["text", "image", "video", "audio", "pdf"], + "description": "Input or output modality" + }, + "Snapshot": { + "type": "object", + "description": "A dated version of a model (inherits parent fields)", + "properties": { + "id": { + "type": "string", + "description": "Snapshot ID (typically includes date, e.g. gpt-4o-2024-08-06)" + }, + "name": { + "type": "string", + "description": "Display name override" + }, + "reasoning": { + "type": "boolean", + "description": "Supports reasoning/thinking mode" + }, + "temperature": { + "type": "boolean", + "description": "Supports temperature parameter" + }, + "tool_call": { + "type": "boolean", + "description": "Supports tool/function calling" + }, + "attachment": { + "type": "boolean", + "description": "Supports file attachments" + }, + "structured_output": { + "type": "boolean", + "description": "Supports structured/JSON output" + }, + "open_weights": { + "type": "boolean", + "description": "Open-weight model" + }, + "deprecated": { + "type": "boolean", + "description": "Deprecated but still accessible" + }, + "limit": { + "type": "object", + "properties": { + "context": { + "type": "integer", + "minimum": 1, + "description": "Maximum context window in tokens" + }, + "output": { + "type": "integer", + "minimum": 1, + "description": "Maximum output tokens" + } + } + }, + "pricing": { + "$ref": "#/definitions/Pricing" + }, + "modalities": { + "type": "object", + "properties": { + "input": { + "type": "array", + "items": { "$ref": "#/definitions/Modality" } + }, + "output": { + "type": "array", + "items": { "$ref": "#/definitions/Modality" } + } + } + }, + "knowledge": { + "type": "string", + "description": "Training data cutoff (e.g. 2023-10)" + }, + "release_date": { + "type": "string", + "description": "Model release date (YYYY-MM-DD or YYYY-MM)" + }, + "last_updated": { + "type": "string", + "description": "Last data update (YYYY-MM-DD or YYYY-MM)" + } + }, + "required": ["id", "last_updated"] + } + }, + "type": "object", + "description": "A single AI model definition", + "properties": { + "id": { + "type": "string", + "description": "Stable model ID (no date suffix)", + "pattern": "^[a-z0-9][a-z0-9._-]*[a-z0-9]$" + }, + "name": { + "type": "string", + "description": "Display name" + }, + "family": { + "type": "string", + "description": "Model family (broad lineage)" + }, + "reasoning": { + "type": "boolean", + "default": false, + "description": "Supports reasoning/thinking mode" + }, + "temperature": { + "type": "boolean", + "default": true, + "description": "Supports temperature parameter" + }, + "tool_call": { + "type": "boolean", + "default": false, + "description": "Supports tool/function calling" + }, + "attachment": { + "type": "boolean", + "default": false, + "description": "Supports file attachments" + }, + "structured_output": { + "type": "boolean", + "default": false, + "description": "Supports structured/JSON output" + }, + "open_weights": { + "type": "boolean", + "default": false, + "description": "Open-weight model" + }, + "deprecated": { + "type": "boolean", + "default": false, + "description": "Deprecated but still accessible" + }, + "limit": { + "type": "object", + "properties": { + "context": { + "type": "integer", + "minimum": 1, + "description": "Maximum context window in tokens" + }, + "output": { + "type": "integer", + "minimum": 1, + "description": "Maximum output tokens" + } + } + }, + "pricing": { + "$ref": "#/definitions/Pricing" + }, + "modalities": { + "type": "object", + "properties": { + "input": { + "type": "array", + "items": { "$ref": "#/definitions/Modality" } + }, + "output": { + "type": "array", + "items": { "$ref": "#/definitions/Modality" } + } + } + }, + "knowledge": { + "type": "string", + "description": "Training data cutoff (e.g. 2023-10)" + }, + "release_date": { + "type": "string", + "description": "Model release date (YYYY-MM-DD or YYYY-MM)" + }, + "last_updated": { + "type": "string", + "description": "Last data update (YYYY-MM-DD or YYYY-MM)" + }, + "snapshots": { + "type": "array", + "items": { "$ref": "#/definitions/Snapshot" }, + "description": "Dated model versions" + } + }, + "required": ["id", "name", "family", "pricing", "modalities", "last_updated"] +} From 1d5e940ae31c676171f92ffab92c4ff28b132c40 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:30:10 +0800 Subject: [PATCH 093/311] =?UTF-8?q?docs:=20add=20jsDelivr=20CDN=20access?= =?UTF-8?q?=20=E2=80=94=20free,=20CORS-enabled,=20no-install=20JSON=20endp?= =?UTF-8?q?oint?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 28 ++++++++++++++++++++++++- docs/api.md | 44 ++++++++++++++++++++++++++++++++++++++++ docs/code-examples.md | 7 +++++++ docs/quick-start.md | 18 ++++++++++++++++ docs/zh/api.md | 44 ++++++++++++++++++++++++++++++++++++++++ docs/zh/code-examples.md | 7 +++++++ docs/zh/quick-start.md | 18 ++++++++++++++++ llms-full.txt | 3 +++ llms.txt | 3 +++ 9 files changed, 171 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 50251f85..f72abfc9 100644 --- a/README.md +++ b/README.md @@ -18,7 +18,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** +**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN access →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** ## Why This Catalog? @@ -267,6 +267,32 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` +### CDN Access (no install) + +The compiled JSON is available via [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) — no download or install needed: + +```html + + +``` + +```bash +# Direct curl (always up-to-date) +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +``` + +```python +# Python — no pip install needed +import urllib.request, json +catalog = json.loads(urllib.request.urlopen("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json").read()) +print(len(catalog["models"])) # 4587 +``` + See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. ## Project Structure diff --git a/docs/api.md b/docs/api.md index 12760da0..4ead1de3 100644 --- a/docs/api.md +++ b/docs/api.md @@ -26,6 +26,50 @@ The package includes: - `dist/index.d.ts` — TypeScript type definitions - `types/` — source type definitions (Model, Snapshot, Provider, Pricing) +## CDN Access (No Install) + +The compiled JSON is available via [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) — no download or install needed. The CDN automatically serves the latest npm release: + +```html + + +``` + +```bash +# Direct curl (always up-to-date) +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' + +# Pin to a specific version +curl -s https://cdn.jsdelivr.net/npm/ai-models@0.1.0/models.json | jq '.stats' +``` + +```python +# Python — no pip install needed +import urllib.request, json +catalog = json.loads(urllib.request.urlopen("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json").read()) +print(len(catalog["models"])) # 4587 +``` + +```go +// Go — no dependencies needed +resp, err := http.Get("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json") +``` + +### CDN vs GitHub Releases + +| Feature | jsDelivr CDN | GitHub Releases | +| ------------- | -------------------------------------------- | --------------------------------------------- | +| URL stability | `cdn.jsdelivr.net/npm/ai-models@latest/...` | `github.com/.../releases/latest/download/...` | +| CORS | ✅ Yes — works in browsers | ❌ No — download only | +| Caching | 7 days (versioned), 5 min (`@latest`) | No caching | +| Speed | Global CDN, 300+ edge locations | GitHub CDN | +| Best for | Web apps, browser scripts, quick prototyping | CLI tools, CI/CD, batch processing | + ## Compiled JSON All model data is available from [GitHub Releases](https://github.com/i-need-token/ai-models/releases/latest) in two formats: diff --git a/docs/code-examples.md b/docs/code-examples.md index 6ecd99b7..68522147 100644 --- a/docs/code-examples.md +++ b/docs/code-examples.md @@ -14,6 +14,13 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` +Or use the [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) (no download needed, CORS-enabled): + +```bash +# Always up-to-date, works in browsers +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +``` + ## TypeScript / JavaScript ### Install the npm package diff --git a/docs/quick-start.md b/docs/quick-start.md index 8000da43..68daa734 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -77,6 +77,24 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` +### CDN access (no install) + +The compiled JSON is available via [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) — no download or install needed: + +```bash +# Always up-to-date, CORS-enabled, works in browsers +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +``` + +```html + +``` + ### From source ```bash diff --git a/docs/zh/api.md b/docs/zh/api.md index 2d216526..8a13296c 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -26,6 +26,50 @@ const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < - `dist/index.d.ts` — TypeScript 类型定义 - `types/` — 源类型定义（Model、Snapshot、Provider、Pricing） +## CDN 访问（无需安装） + +编译后的 JSON 可通过 [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) 访问 — 无需下载或安装。CDN 自动提供最新的 npm 发布版本： + +```html + + +``` + +```bash +# 直接 curl（始终最新） +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' + +# 锁定特定版本 +curl -s https://cdn.jsdelivr.net/npm/ai-models@0.1.0/models.json | jq '.stats' +``` + +```python +# Python — 无需 pip install +import urllib.request, json +catalog = json.loads(urllib.request.urlopen("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json").read()) +print(len(catalog["models"])) # 4587 +``` + +```go +// Go — 无需依赖 +resp, err := http.Get("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json") +``` + +### CDN 与 GitHub Releases 对比 + +| 特性 | jsDelivr CDN | GitHub Releases | +| ---------- | ------------------------------------------- | --------------------------------------------- | +| URL 稳定性 | `cdn.jsdelivr.net/npm/ai-models@latest/...` | `github.com/.../releases/latest/download/...` | +| CORS | ✅ 支持 — 可在浏览器中使用 | ❌ 不支持 — 仅下载 | +| 缓存 | 7 天（版本化），5 分钟（`@latest`） | 无缓存 | +| 速度 | 全球 CDN，300+ 边缘节点 | GitHub CDN | +| 适用场景 | Web 应用、浏览器脚本、快速原型 | CLI 工具、CI/CD、批处理 | + ## 编译 JSON 所有模型数据可以从 [GitHub Releases](https://github.com/i-need-token/ai-models/releases/latest) 下载，提供两种格式： diff --git a/docs/zh/code-examples.md b/docs/zh/code-examples.md index 9e8256f3..f84b7a24 100644 --- a/docs/zh/code-examples.md +++ b/docs/zh/code-examples.md @@ -14,6 +14,13 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` +或使用 [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models)（无需下载，支持 CORS）： + +```bash +# 始终最新，可在浏览器中使用 +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +``` + ## TypeScript / JavaScript ### 安装 npm 包 diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md index 830e206b..4a9821b6 100644 --- a/docs/zh/quick-start.md +++ b/docs/zh/quick-start.md @@ -77,6 +77,24 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` +### CDN 访问（无需安装） + +编译后的 JSON 可通过 [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) 访问 — 无需下载或安装： + +```bash +# 始终最新，支持 CORS，可在浏览器中使用 +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +``` + +```html + +``` + ### 从源码 ```bash diff --git a/llms-full.txt b/llms-full.txt index 79a87e58..dcb9478f 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -41,6 +41,9 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode # CSV — flat table for Excel/Google Sheets (560 KB) curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv + +# CDN (no install, CORS-enabled, always up-to-date) +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' ``` ## Data Format diff --git a/llms.txt b/llms.txt index 1e1ec33d..68c1906b 100644 --- a/llms.txt +++ b/llms.txt @@ -39,6 +39,9 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode # CSV — flat table for Excel/Google Sheets (560 KB) curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv + +# CDN (no install, CORS-enabled, always up-to-date) +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' ``` ## Data Format From dc7db8696f10242a82fe7645fb2e488e11b1fb1c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:35:11 +0800 Subject: [PATCH 094/311] =?UTF-8?q?feat:=20add=20interactive=20model=20cat?= =?UTF-8?q?alog=20on=20GitHub=20Pages=20=E2=80=94=20search,=20sort,=20filt?= =?UTF-8?q?er=204,587=20models?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .github/workflows/pages.yml | 39 +++ README.md | 60 ++-- site/index.html | 562 ++++++++++++++++++++++++++++++++++++ 3 files changed, 632 insertions(+), 29 deletions(-) create mode 100644 .github/workflows/pages.yml create mode 100644 site/index.html diff --git a/.github/workflows/pages.yml b/.github/workflows/pages.yml new file mode 100644 index 00000000..d2e023c5 --- /dev/null +++ b/.github/workflows/pages.yml @@ -0,0 +1,39 @@ +name: Deploy Pages + +on: + push: + branches: [main] + paths: + - "site/**" + - ".github/workflows/pages.yml" + workflow_dispatch: + +permissions: + contents: read + pages: write + id-token: write + +concurrency: + group: pages + cancel-in-progress: true + +jobs: + deploy: + runs-on: ubuntu-latest + environment: + name: github-pages + url: ${{ steps.deployment.outputs.page_url }} + steps: + - uses: actions/checkout@v4 + + - name: Setup Pages + uses: actions/configure-pages@v5 + + - name: Upload artifact + uses: actions/upload-pages-artifact@v3 + with: + path: site + + - name: Deploy to GitHub Pages + id: deployment + uses: actions/deploy-pages@v4 diff --git a/README.md b/README.md index f72abfc9..fbfe4237 100644 --- a/README.md +++ b/README.md @@ -18,7 +18,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN access →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** +**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[🔍 Search models →](https://i-need-token.github.io/ai-models/)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN access →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** ## Why This Catalog? @@ -44,14 +44,15 @@ Machine-readable YAML catalog of every major AI model provider and their models ## Use Cases -| Use Case | How This Catalog Helps | -| ------------------------------ | ------------------------------------------------------------------------- | -| 💰 **Find the cheapest model** | [Pricing comparison](docs/pricing-comparison.md) across 95 providers | -| 🔎 **Pick the right model** | [Model comparison](docs/model-comparison.md) by capability, context, cost | -| 🔌 **Build an API gateway** | Structured pricing + modality data for routing decisions | -| 📊 **Track the AI landscape** | 2,712 models with release dates, deprecation status | -| 🤖 **Power an AI tool** | TypeScript types + Zod validation = type-safe access | -| 🌍 **Find local/EU providers** | [Provider overview](docs/providers.md) with market segmentation | +| Use Case | How This Catalog Helps | +| ------------------------------ | --------------------------------------------------------------------------------------- | +| 💰 **Find the cheapest model** | [Pricing comparison](docs/pricing-comparison.md) across 95 providers | +| 🔎 **Pick the right model** | [Model comparison](docs/model-comparison.md) by capability, context, cost | +| 🔍 **Search & filter models** | [Interactive catalog](https://i-need-token.github.io/ai-models/) — search, sort, filter | +| 🔌 **Build an API gateway** | Structured pricing + modality data for routing decisions | +| 📊 **Track the AI landscape** | 2,712 models with release dates, deprecation status | +| 🤖 **Power an AI tool** | TypeScript types + Zod validation = type-safe access | +| 🌍 **Find local/EU providers** | [Provider overview](docs/providers.md) with market segmentation | ## Quick Numbers @@ -330,26 +331,27 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin ## Documentation -| Document | Description | -| ------------------------------------------------------- | --------------------------------------------------------------- | -| [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | -| [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | -| [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | -| [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | -| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | -| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | -| [Code Examples](docs/code-examples.md) | Practical examples in TypeScript, Python, Go, Rust, jq | -| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | -| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | -| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | -| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | -| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | -| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | -| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | -| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | -| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | -| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | -| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | +| Document | Description | +| ------------------------------------------------------------------- | --------------------------------------------------------------- | +| [Tool Calling Models](docs/tool-calling.md) | 2,350 tool-calling models — cheapest, largest context, free | +| [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | +| [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | +| [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | +| [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | +| [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | +| [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | +| [Code Examples](docs/code-examples.md) | Practical examples in TypeScript, Python, Go, Rust, jq | +| [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | +| [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | +| [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | +| [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | +| [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | +| [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | +| [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | +| [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | +| [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | +| [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | +| [Design Principles & Pitfalls](docs/lessons-learned.md) | Lessons learned from building the catalog | **中文文档：** diff --git a/site/index.html b/site/index.html new file mode 100644 index 00000000..20c6096c --- /dev/null +++ b/site/index.html @@ -0,0 +1,562 @@ + + + + + + AI Models Catalog — Search & Compare 4,587 AI Models + + + + + + + + + + + +
+

🤖 AI Models Catalog

+
+ Search and compare AI models by pricing, context window, capabilities, and more +
+
+
— models
+
— providers
+
— tool calling
+
— vision
+
— reasoning
+
— free
+
+
+ +
+ + + + + + + + +
+
+ +
+ + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContextToolReasonVisionJSONTags
+
+ +

Loading catalog…

+
+
+
+ +
+ GitHub · + Download JSON + · + Download CSV + · npm · + CDN +
+ + + + From 4a1173e9860a03e1b42637a5a04617f696fe36c5 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:35:32 +0800 Subject: [PATCH 095/311] fix: correct CHANGELOG date from 2025-05 to 2026-05 --- CHANGELOG.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index 12eb326c..f74263f0 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,7 +2,7 @@ All notable changes to the AI Models Catalog. -## 2025-05 +## 2026-05 ### Added From 8157c12a58ae8fcf52675ac1e03450f3dc38975a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:37:16 +0800 Subject: [PATCH 096/311] feat: add Hugging Face dataset badge, sync workflow, and cdn keyword --- .github/workflows/sync-hf.yml | 59 +++++++++++++++++++++++++++++++++++ README.md | 1 + package.json | 1 + 3 files changed, 61 insertions(+) create mode 100644 .github/workflows/sync-hf.yml diff --git a/.github/workflows/sync-hf.yml b/.github/workflows/sync-hf.yml new file mode 100644 index 00000000..726bbdae --- /dev/null +++ b/.github/workflows/sync-hf.yml @@ -0,0 +1,59 @@ +name: Sync to Hugging Face + +on: + release: + types: [published] + workflow_dispatch: + +jobs: + sync: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + + - name: Compile JSON + run: npx tsx scripts/compile.ts + + - name: Export CSV + run: npx tsx scripts/export-csv.ts + + - name: Push to Hugging Face + env: + HF_TOKEN: ${{ secrets.HF_TOKEN }} + run: | + if [ -z "$HF_TOKEN" ]; then + echo "HF_TOKEN not set — skipping Hugging Face sync" + exit 0 + fi + + pip install huggingface-hub + + python3 -c " + from huggingface_hub import HfApi + import os + + api = HfApi(token=os.environ['HF_TOKEN']) + repo_id = os.environ.get('HF_REPO_ID', 'i-need-token/ai-models') + + api.upload_file( + path_or_fileobj='dist/models.json', + path_in_repo='models.json', + repo_id=repo_id, + repo_type='dataset', + commit_message='Sync models.json from GitHub release', + ) + + api.upload_file( + path_or_fileobj='models.csv', + path_in_repo='models.csv', + repo_id=repo_id, + repo_type='dataset', + commit_message='Sync models.csv from GitHub release', + ) + + print(f'Synced to https://huggingface.co/datasets/{repo_id}') + " diff --git a/README.md b/README.md index fbfe4237..23752951 100644 --- a/README.md +++ b/README.md @@ -8,6 +8,7 @@ [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) [![npm version](https://img.shields.io/npm/v/ai-models.svg)](https://www.npmjs.com/package/ai-models) +[![Hugging Face](https://img.shields.io/badge/%F0%9F%A4%97-Dataset-blue)](https://huggingface.co/datasets/i-need-token/ai-models) [![Models](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/models.json)](providers/) [![Providers](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/providers.json)](providers/) [![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) [![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) diff --git a/package.json b/package.json index 84aeefc8..51f08b8d 100644 --- a/package.json +++ b/package.json @@ -8,6 +8,7 @@ "ai-pricing", "ai-provider", "anthropic", + "cdn", "cerebras", "chatgpt", "claude", From 8b903dcdf1f9908b362592d2df448ff6f605f7d9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:37:51 +0800 Subject: [PATCH 097/311] docs: add Hugging Face dataset link to API docs --- docs/api.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/api.md b/docs/api.md index 4ead1de3..b26c356a 100644 --- a/docs/api.md +++ b/docs/api.md @@ -80,6 +80,8 @@ All model data is available from [GitHub Releases](https://github.com/i-need-tok | `models.csv` | CSV | ~560 KB | Excel, Google Sheets, data analysis | | `stats.json` | JSON | ~1 KB | Catalog statistics summary | +Also available on [Hugging Face Datasets](https://huggingface.co/datasets/i-need-token/ai-models) for the ML community. + ```bash # Download JSON (full metadata) curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json From 4156ba4292602cd95725d03c6f327eeac48fbbb0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:38:43 +0800 Subject: [PATCH 098/311] feat: add reusable GitHub Action for fetching catalog data in CI workflows --- action.yml | 57 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+) create mode 100644 action.yml diff --git a/action.yml b/action.yml new file mode 100644 index 00000000..da29ef27 --- /dev/null +++ b/action.yml @@ -0,0 +1,57 @@ +name: "AI Models Catalog" +description: "Fetch model data from the AI Models Catalog for use in your workflows" +branding: + icon: database + color: purple +inputs: + format: + description: "Output format — json or csv" + required: false + default: "json" + filter: + description: "JMESPath filter expression (e.g. 'models[?tool_call==`true`]')" + required: false + default: "" +outputs: + data: + description: "The fetched model data" + value: ${{ steps.fetch.outputs.data }} + model-count: + description: "Number of models in the catalog" + value: ${{ steps.fetch.outputs.model_count }} + provider-count: + description: "Number of providers in the catalog" + value: ${{ steps.fetch.outputs.provider_count }} + +runs: + using: "composite" + steps: + - name: Fetch catalog + id: fetch + shell: bash + run: | + if [ "${{ inputs.format }}" = "csv" ]; then + URL="https://github.com/i-need-token/ai-models/releases/latest/download/models.csv" + curl -sL "$URL" -o models.csv + echo "data<> "$GITHUB_OUTPUT" + cat models.csv >> "$GITHUB_OUTPUT" + echo "EOF" >> "$GITHUB_OUTPUT" + COUNT=$(tail -n +2 models.csv | wc -l | tr -d ' ') + echo "model_count=$COUNT" >> "$GITHUB_OUTPUT" + echo "provider_count=0" >> "$GITHUB_OUTPUT" + else + URL="https://cdn.jsdelivr.net/npm/ai-models@latest/models.json" + curl -sL "$URL" -o models.json + if [ -n "${{ inputs.filter }}" ]; then + DATA=$(jq -c '${{ inputs.filter }}' models.json) + else + DATA=$(cat models.json) + fi + echo "data<> "$GITHUB_OUTPUT" + echo "$DATA" >> "$GITHUB_OUTPUT" + echo "EOF" >> "$GITHUB_OUTPUT" + MC=$(jq '.models | length' models.json) + PC=$(jq '.stats.providers' models.json) + echo "model_count=$MC" >> "$GITHUB_OUTPUT" + echo "provider_count=$PC" >> "$GITHUB_OUTPUT" + fi From b2142da23d6fc03eab5e1a2a9fc02f9da5f2c601 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:39:25 +0800 Subject: [PATCH 099/311] docs: add GitHub Action usage section to README --- README.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/README.md b/README.md index 23752951..4ccda6c0 100644 --- a/README.md +++ b/README.md @@ -297,6 +297,12 @@ print(len(catalog["models"])) # 4587 See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. +### Use as GitHub Action + +Fetch catalog data directly in your CI workflows: + +true + ## Project Structure ``` From 86bd5c0c0c8a785f4fb514ffc4eedc5944e061fa Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:40:26 +0800 Subject: [PATCH 100/311] docs: add huggingface and github-action topics to repo-settings --- .github/repo-settings.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/.github/repo-settings.md b/.github/repo-settings.md index 573d4f88..79430127 100644 --- a/.github/repo-settings.md +++ b/.github/repo-settings.md @@ -44,3 +44,5 @@ Add these topics to the repository (Settings → General → Topics): - ai-model-catalog - structured-data - npm-package +- huggingface +- github-action From 462d14124af22bb109ac2d8a78a68ab00645677d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:41:08 +0800 Subject: [PATCH 101/311] docs: add interactive catalog link to CONTRIBUTING.md --- CONTRIBUTING.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index c209d2cd..9f10e1b7 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -4,6 +4,8 @@ Thank you for your interest in contributing! This guide covers everything you ne ## Quick Start +> 💡 **Try the [Interactive Catalog](https://i-need-token.github.io/ai-models/)** — search, sort, and filter all 4,587 models in your browser. + 1. Fork the repository 2. Create your feature branch: `git checkout -b feature/my-provider` 3. Make your changes From 0c370a3a95f1045c2007f295b83172c9f48923cd Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:42:58 +0800 Subject: [PATCH 102/311] docs(zh): add Hugging Face dataset link to Chinese API docs --- docs/zh/api.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/zh/api.md b/docs/zh/api.md index 8a13296c..c63bc61c 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -80,6 +80,8 @@ resp, err := http.Get("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | `models.csv` | CSV | ~560 KB | Excel、Google Sheets、数据分析 | | `stats.json` | JSON | ~1 KB | 目录统计摘要 | +也可在 [Hugging Face Datasets](https://huggingface.co/datasets/i-need-token/ai-models) 上获取，方便机器学习社区使用。 + ```bash # 下载 JSON（完整元数据） curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json From 5f91c04120bcd025e5242892cdf1e0f51a4c47f0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 13:45:10 +0800 Subject: [PATCH 103/311] feat: add JSON-LD structured data to catalog page for SEO --- site/index.html | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/site/index.html b/site/index.html index 20c6096c..1f783a3c 100644 --- a/site/index.html +++ b/site/index.html @@ -24,6 +24,27 @@ rel="icon" href="data:image/svg+xml,🤖" /> + -
-

🤖 AI Models Catalog

-
- Search and compare AI models by pricing, context window, capabilities, and more +
+
+

🤖 AI Models Catalog

+
+ / to search + +
-
-
— models
-
— providers
-
— tool calling
-
— vision
-
— reasoning
-
— free
+
+
+ + + +
+
+ + + + + + + + + +
+
+
+ + + + + + + + + + + + +
Model Provider Context Input $/M Output $/M Capabilities
+
+
+ ⭐ Star on GitHub · + CDN · + GitHub Action · + npm · + Hugging Face
- -
- - - - - - - - -
-
- -
- - - - - - - - - - - - - - - - - - - - -
ModelProviderInput $/1MOutput $/1MContextToolReasonVisionJSONTags
-
- -

Loading catalog…

-
-
-
- -
- GitHub · - Download JSON - · - Download CSV - · npm · - CDN · - GitHub Action -
+
From 7f6b3281b0ecc33470e81df1af8161fde4e9e337 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:38:25 +0800 Subject: [PATCH 133/311] =?UTF-8?q?Add=20Quick=20Compare=20table=20to=20RE?= =?UTF-8?q?ADME=20=E2=80=94=2019=20flagship=20models=20at=20a=20glance?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Shows model, provider, context, pricing, tools/reasoning/vision at a glance - Covers OpenAI, Anthropic, Google, DeepSeek, Meta, xAI, Mistral, Alibaba - Collapsible explanation for table columns - Links to full model-comparison doc for 4,587 models - Immediate value demonstration within first screen of README --- README.md | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) diff --git a/README.md b/README.md index 3e885248..0ce89f8a 100644 --- a/README.md +++ b/README.md @@ -44,6 +44,43 @@ Machine-readable YAML catalog of every major AI model provider and their models - [Who's Using This](#whos-using-this) - [License](#license) +## Quick Compare + +> Popular models at a glance — full data for [4,587 models](docs/model-comparison.md) + +| Model | Provider | Context | Input $/M | Output $/M | Tools | Reason | Vision | +| ---------------- | --------- | ------- | --------- | ---------- | ----- | ------ | ------ | +| gpt-4.1 | openai | 1M | $2 | $8 | ✓ | | ✓ | +| gpt-4.1-mini | openai | 1M | $0.40 | $1.60 | ✓ | | ✓ | +| gpt-4.1-nano | openai | 1M | $0.10 | $0.40 | ✓ | | ✓ | +| o3 | openai | 200K | $10 | $40 | ✓ | ✓ | ✓ | +| o4-mini | openai | 200K | $1.10 | $4.40 | ✓ | ✓ | ✓ | +| claude-opus-4 | anthropic | 200K | $15 | $75 | ✓ | ✓ | ✓ | +| claude-sonnet-4 | anthropic | 200K | $3 | $15 | ✓ | ✓ | ✓ | +| claude-haiku-4 | anthropic | 200K | $1 | $5 | ✓ | ✓ | ✓ | +| gemini-2.5-pro | google | 1M | $1.25 | $10 | ✓ | ✓ | ✓ | +| gemini-2.5-flash | google | 1M | $0.15 | $0.60 | ✓ | ✓ | ✓ | +| deepseek-r1 | deepseek | 128K | $0.55 | $2.19 | | ✓ | | +| deepseek-chat | deepseek | 128K | $0.14 | $0.28 | ✓ | | | +| llama-4-maverick | meta | 1M | $0.20 | $0.20 | ✓ | | ✓ | +| llama-4-scout | meta | 10M | $0.03 | $0.03 | ✓ | | ✓ | +| grok-3 | xai | 131K | $3 | $15 | ✓ | | ✓ | +| grok-3-mini | xai | 131K | $0.30 | $0.50 | ✓ | ✓ | ✓ | +| mistral-large | mistral | 128K | $2 | $6 | ✓ | | ✓ | +| qwen3-235b-a22b | alibaba | 128K | $0.14 | $0.42 | ✓ | ✓ | ✓ | +| qwen3-30b-a3b | alibaba | 128K | $0.03 | $0.05 | ✓ | ✓ | ✓ | + +
📖 How to read this table + +- **Context**: Maximum context window (input + output tokens) +- **Input/Output $/M**: Price per million tokens +- **Tools**: Supports function/tool calling +- **Reason**: Uses chain-of-thought reasoning +- **Vision**: Accepts image input +- Prices shown are for standard (non-cached) API calls. Many providers offer 50-90% discounts for cached inputs. + +
+ ## Use Cases | Use Case | How This Catalog Helps | From e1e74c78216ff4bc0a1f383040a312b8a100f547 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:39:00 +0800 Subject: [PATCH 134/311] Add Quick Compare to README table of contents --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 0ce89f8a..ecece18c 100644 --- a/README.md +++ b/README.md @@ -33,6 +33,7 @@ Machine-readable YAML catalog of every major AI model provider and their models ## Contents +- [Quick Compare](#quick-compare) - [Use Cases](#use-cases) - [Quick Numbers](#quick-numbers) - [Example Model](#example-model) From ec9ced263c4dc9cd8a434b29b9ebeae09c2409a1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:40:38 +0800 Subject: [PATCH 135/311] Expand CHANGELOG.md with comprehensive v0.1.0 release notes - Core data: 95 providers, 4,587 models, TypeScript types, Zod schemas - Programmatic access: npm, CDN, CSV, GitHub Action, HF dataset, interactive catalog - Documentation: 22 EN + 22 ZH = 44 pages - Community: CONTRIBUTING, issue templates, 10 CI workflows - Provider coverage: producers, inference platforms, cloud hosted, Chinese/European markets - Data highlights: reasoning, tool calling, vision, free, open weights counts --- CHANGELOG.md | 68 ++++++++++++++++++++++++++++++++++++++++------------ 1 file changed, 53 insertions(+), 15 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index f74263f0..2612d240 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,28 +2,65 @@ All notable changes to the AI Models Catalog. -## 2026-05 +## v0.1.0 (2026-05) -### Added +### Core Data - **95 providers** with structured YAML model data - **4,587 model files** covering 2,712 unique model IDs across 441 families - TypeScript type definitions (`types/model.ts`, `types/pricing.ts`, `types/provider.ts`) - Zod runtime validation schemas (`types/schemas.ts`) +- JSON Schema for YAML validation (`schema.json`) - Automated scrape scripts for each provider (`providers//scrape.ts`) -- CLI tools: `scripts/sync.ts` (data sync), `scripts/validate.ts` (validation) -- Documentation: - - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, vision, open-weight comparisons - - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers and platforms - - [Provider Overview](docs/providers.md) — all providers organized by type and market - - [Data Schema Reference](docs/data-schema.md) — complete YAML schema documentation - - [Data Acquisition Guide](docs/data-acquisition.md) — how we acquire and update model data - - [Design Principles & Pitfalls](docs/lessons-learned.md) — lessons learned - - All docs available in Chinese (`docs/zh/`) -- Community files: CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md -- GitHub issue templates (bug report, provider request, data update) -- GitHub PR template -- GitHub Actions CI workflow (validate, type check, lint) +- CLI tools: `scripts/sync.ts`, `scripts/validate.ts`, `scripts/stats.ts`, `scripts/compile.ts`, `scripts/export-csv.ts` + +### Programmatic Access + +- npm package (`ai-models`) with TypeScript types and JSON data +- jsDelivr CDN access (`cdn.jsdelivr.net/npm/ai-models@latest/models.json`) +- CSV export (`models.csv`) available from GitHub Releases +- Reusable GitHub Action (`action.yml`) for CI/CD pipelines +- Hugging Face dataset sync (`huggingface.co/datasets/i-need-token/ai-models`) +- Interactive model catalog on GitHub Pages (`i-need-token.github.io/ai-models/`) + +### Documentation (22 EN + 22 ZH = 44 pages) + +- [Quick Start](docs/quick-start.md) — find the right model in 30 seconds +- [API & Programmatic Access](docs/api.md) — npm, CDN, CSV, GitHub Action, Hugging Face +- [FAQ](docs/faq.md) — common questions about the catalog, data, and contributing +- [Glossary](docs/glossary.md) — key terms and definitions for AI model terminology +- [Code Examples](docs/code-examples.md) — TypeScript, Python, Go, Rust, jq, Excel +- [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight +- [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers +- [Context Window Comparison](docs/context-windows.md) — largest context windows by tier +- [Modality Matrix](docs/modality-matrix.md) — which models support what modalities +- [Tool Calling Models](docs/tool-calling.md) — 2,350 tool-calling models +- [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models +- [Structured Output](docs/structured-output.md) — 829 JSON-mode models +- [Vision Models](docs/vision-models.md) — 1,487 vision models +- [Video Models](docs/video-models.md) — 167 video input models +- [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models +- [Image Generation](docs/image-generation.md) — 28 image generation models +- [Free AI Models](docs/free-models.md) — 81 free models +- [Open-Weight Models](docs/open-weights.md) — 527 open-weight models +- [Provider Overview](docs/providers.md) — all 95 providers by type and market +- [Data Schema Reference](docs/data-schema.md) — complete YAML schema +- [Data Acquisition](docs/data-acquisition.md) — how we acquire and update model data +- [Design Principles](docs/lessons-learned.md) — lessons learned from building the catalog +- All docs available in Chinese (`docs/zh/`) with cross-language links + +### Community & Infrastructure + +- CONTRIBUTING.md with Good First Issues table +- CODE_OF_CONDUCT.md, SECURITY.md, CHANGELOG.md +- CITATION.cff for academic discoverability (18 keywords) +- 6 GitHub issue templates + PR template + config.yml +- 10 CI workflows (validate, stats, release, publish, sync, sync-hf, pages, labeler, auto-merge, welcome) +- PR auto-labeler, Dependabot auto-merge, Renovate config +- Welcome workflow for first-time contributors +- CODEOWNERS, .editorconfig, .npmignore, Makefile +- llms.txt + llms-full.txt for AI discoverability +- Social preview SVG, robots.txt, sitemap.xml for SEO ### Provider Coverage @@ -37,6 +74,7 @@ All notable changes to the AI Models Catalog. - 1,306 reasoning models - 2,350 tool-calling models +- 829 structured output models - 527 open-weight models - 81 free models - 1,487 vision (image input) models From b58be6d0f33a37ff8065cbf4eecd11ec9c3c0f4f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:43:47 +0800 Subject: [PATCH 136/311] Add cached pricing documentation (EN + ZH) - 75 models with prompt caching across 7 providers - Shows standard vs cached read/write pricing with savings percentage - Explains why cached pricing matters (cost, latency, throughput) - High-value content for developers optimizing API costs --- docs/cached-pricing.md | 1411 +++++++++++++++++++++++++++++++++++++ docs/zh/cached-pricing.md | 1411 +++++++++++++++++++++++++++++++++++++ 2 files changed, 2822 insertions(+) create mode 100644 docs/cached-pricing.md create mode 100644 docs/zh/cached-pricing.md diff --git a/docs/cached-pricing.md b/docs/cached-pricing.md new file mode 100644 index 00000000..8cd3a6e1 --- /dev/null +++ b/docs/cached-pricing.md @@ -0,0 +1,1411 @@ +# Cached Pricing + +[中文](zh/cached-pricing.md) + +AI models with prompt caching support, showing standard vs. cached pricing. Cached inputs can be **50-90% cheaper** than standard input tokens. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Cached Pricing Matters + +Prompt caching lets you store repeated prompt prefixes (system prompts, few-shot examples, tool definitions) and reuse them across requests. This dramatically reduces: + +- **Cost**: 50-90% savings on input tokens +- **Latency**: Faster time-to-first-token for cached content +- **Throughput**: More efficient use of rate limits + +## Stats + +| Metric | Count | +| ------------------------- | ----- | +| Models with cache pricing | 1374 | +| Providers | 39 | + +## Providers + +`aihubmix`, `aion`, `amazon-bedrock`, `auriko`, `baidu`, `baseten`, `chutes`, `clarifai`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `deepseek`, `digitalocean`, `fastrouter`, `friendli`, `google`, `google-vertex`, `groq`, `hpc-ai`, `inception`, `jiekou`, `llmgateway`, `martian`, `minimax`, `moonshotai`, `nanogpt`, `openai`, `openrouter`, `ppio`, `privatemode`, `requesty`, `siliconflow`, `stepfun`, `tencent-tokenhub`, `togetherai`, `upstage`, `venice`, `wafer` + +## Model Pricing + +| Model | Provider | Context | Input $/M | Cache Read $/M | Cache Write $/M | Savings | +| ---------------------------------------------------------- | ---------------- | ------- | ------------------- | --------------------- | --------------- | ------- | +| aistudio_gemini-2.0-flash | aihubmix | — | $0.05 | $0.125 | — | -150% | +| aistudio_gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| anthropic-opus-4-6 | aihubmix | — | $2.5 | $0.25 | $3.125 | 90% | +| claude-haiku-4-5 | aihubmix | — | $0.55 | $0.055 | $0.6875 | 90% | +| claude-sonnet-4-0 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5-think | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| codex-mini-latest | aihubmix | — | $0.75 | $0.1875 | — | 75% | +| deepseek-v3.2 | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| deepseek-v3.2-exp | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-exp-think | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-think | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| doubao-1.5-lite-32k | aihubmix | — | $0.025 | $0.005 | — | 80% | +| doubao-1.5-pro-32k | aihubmix | — | $0.067 | $0.0134 | — | 80% | +| doubao-lite-32k | aihubmix | — | $0.03 | $0.006 | — | 80% | +| doubao-pro-32k | aihubmix | — | $0.07 | $0.014 | — | 80% | +| doubao-seed-1-6 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-flash | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-flash-250615 | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-lite | aihubmix | — | $0.041 | $0.0082 | — | 80% | +| doubao-seed-1-6-thinking | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-thinking-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-vision-250815 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| doubao-seed-1-8 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| gemini-2.0-flash | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.0-flash-001 | aihubmix | — | $0.05 | $0.125 | — | -150% | +| gemini-2.0-flash-search | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.5-flash | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-lite | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-pro | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| glm-4.5-airx | aihubmix | — | $0.55 | $0.11 | — | 80% | +| glm-4.5-x | aihubmix | — | $1.1 | $0.22 | — | 80% | +| glm-4.6 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| glm-4.6v | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| glm-4.7 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| gpt-4.1 | aihubmix | — | $1 | $0.25 | — | 75% | +| gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| gpt-4.1-nano | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gpt-4o | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06-global | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-11-20 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-mini | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-2024-07-18 | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-global | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-search-preview | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-search-preview | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-5 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5-nano | aihubmix | — | $0.025 | $0.0025 | — | 90% | +| gpt-5.1 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-max | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5.2 | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-chat-latest | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-codex | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-high | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-low | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-pro | aihubmix | — | $10.5 | $1.05 | — | 90% | +| grok-4 | aihubmix | — | $1.65 | $0.4125 | — | 75% | +| grok-4-1-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-1-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4.20-beta-0309-non-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-beta-0309-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-beta-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-code-fast-1 | aihubmix | — | $0.1 | $0.025 | — | 75% | +| kimi-k2-thinking | aihubmix | — | $0.274 | $0.0685 | — | 75% | +| kimi-k2-turbo-preview | aihubmix | — | $0.6 | $0.15 | — | 75% | +| kimi-k2.5 | aihubmix | — | $0.3 | $0.0525 | — | 82% | +| mimo-v2-flash | aihubmix | — | $0.0959 | $0.01918 | — | 80% | +| mimo-v2-omni | aihubmix | — | $0.22 | $0.044 | — | 80% | +| mimo-v2-pro | aihubmix | — | $0.55 | $0.11 | — | 80% | +| nvidia-nemotron-3-super-120b-a12b | aihubmix | — | $0.055 | $0.01375 | — | 75% | +| o1 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-2024-12-17 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-global | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-mini | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-mini-2024-09-12 | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-preview | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-preview-2024-09-12 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o3 | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-deep-research | aihubmix | — | $5 | $1.25 | — | 75% | +| o3-global | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-mini | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o3-mini-global | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o4-mini | aihubmix | — | $0.55 | $0.1375 | — | 75% | +| qwen-plus | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-04-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-07-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-latest | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-turbo | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen-turbo-latest | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen3-coder-plus | aihubmix | — | $0.27 | $0.054 | — | 80% | +| qwen3-max | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-2026-01-23 | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-preview | aihubmix | — | $0.423 | $0.0846 | — | 80% | +| qwen3-vl-flash | aihubmix | — | $0.0103 | $0.00206 | — | 80% | +| qwen3-vl-plus | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| zai-glm-5-turbo | aihubmix | — | $0.6 | $0.12 | — | 80% | +| aion-2.0 | aion | — | $0.7999999999999999 | $0.19999999999999998 | — | 75% | +| aion-2.5 | aion | — | $1 | $0.35 | — | 65% | +| amazon-nova-2-lite | amazon-bedrock | — | $0.33 | $0.0825 | — | 75% | +| amazon-nova-lite | amazon-bedrock | — | $0.06 | $0.015 | — | 75% | +| amazon-nova-micro | amazon-bedrock | — | $0.035 | $0.00875 | — | 75% | +| amazon-nova-premier | amazon-bedrock | — | $2.5 | $0.625 | — | 75% | +| amazon-nova-pro | amazon-bedrock | — | $0.8 | $0.2 | — | 75% | +| claude-haiku-4-5-20251001 | auriko | — | $1 | $0.1 | $1.25 | 90% | +| claude-opus-4-1-20250805 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-20250514 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-5-20251101 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-6 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-7 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-sonnet-4-20250514 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5-20250929 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-6 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| deepseek-r1-0528 | auriko | — | $0.5 | $0.35 | — | 30% | +| deepseek-v3-0324 | auriko | — | $0.2 | $0.135 | — | 32% | +| deepseek-v3.1 | auriko | — | $0.21 | $0.13 | — | 38% | +| deepseek-v3.1-terminus | auriko | — | $0.27 | $0.13 | — | 52% | +| deepseek-v3.2 | auriko | — | $0.26 | $0.13 | — | 50% | +| deepseek-v4-flash | auriko | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | auriko | — | $0.435 | $0.003625 | — | 99% | +| gemini-2.5-flash | auriko | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-lite | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-pro | auriko | — | $1.25 | $0.125 | — | 90% | +| gemini-3-flash-preview | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-3.1-flash-lite | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-flash-lite-preview | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-pro-preview | auriko | — | $2 | $0.2 | — | 90% | +| gemini-3.1-pro-preview-customtools | auriko | — | $2 | $0.2 | — | 90% | +| gemini-flash-latest | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-flash-lite-latest | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-pro-latest | auriko | — | $2 | $0.2 | — | 90% | +| glm-4.5 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.5-air | auriko | — | $0.2 | $0.03 | — | 85% | +| glm-4.5-airx | auriko | — | $1.1 | $0.22 | — | 80% | +| glm-4.5-x | auriko | — | $2.2 | $0.45 | — | 80% | +| glm-4.5v | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6v | auriko | — | $0.3 | $0.05 | — | 83% | +| glm-4.6v-flashx | auriko | — | $0.04 | $0.004 | — | 90% | +| glm-4.7 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.7-flashx | auriko | — | $0.07 | $0.01 | — | 86% | +| glm-5 | auriko | — | $1 | $0.2 | — | 80% | +| glm-5-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| glm-5.1 | auriko | — | $1.4 | $0.26 | — | 81% | +| glm-5v-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| gpt-4.1-2025-04-14 | auriko | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini-2025-04-14 | auriko | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano-2025-04-14 | auriko | — | $0.1 | $0.025 | — | 75% | +| gpt-4o-2024-08-06 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-2024-11-20 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-mini-2024-07-18 | auriko | — | $0.15 | $0.075 | — | 50% | +| gpt-5-2025-08-07 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-mini-2025-08-07 | auriko | — | $0.25 | $0.025 | — | 90% | +| gpt-5-nano-2025-08-07 | auriko | — | $0.05 | $0.005 | — | 90% | +| gpt-5.1-2025-11-13 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.1-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.2-2025-12-11 | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.2-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.3-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.4-2026-03-05 | auriko | — | $2.5 | $0.25 | — | 90% | +| gpt-5.4-mini-2026-03-17 | auriko | — | $0.75 | $0.075 | — | 90% | +| gpt-5.4-nano-2026-03-17 | auriko | — | $0.2 | $0.02 | — | 90% | +| gpt-5.5-2026-04-23 | auriko | — | $5 | $0.5 | — | 90% | +| gpt-oss-120b | auriko | — | $0.15 | $0.01 | — | 93% | +| gpt-oss-20b | auriko | — | $0.07 | $0.04 | — | 43% | +| grok-4.20-0309-non-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.20-0309-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.3 | auriko | — | $1.25 | $0.2 | — | 84% | +| hy3-preview | auriko | — | $0.066 | $0.029 | — | 56% | +| kimi-k2-0711-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-0905-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking-turbo | auriko | — | $1.15 | $0.15 | — | 87% | +| kimi-k2-turbo-preview | auriko | — | $1.15 | $0.15 | — | 87% | +| kimi-k2.5 | auriko | — | $0.6 | $0.1 | — | 83% | +| kimi-k2.6 | auriko | — | $0.95 | $0.16 | — | 83% | +| mimo-v2.5 | auriko | — | $0.4 | $0.08 | — | 80% | +| mimo-v2.5-pro | auriko | — | $1 | $0.2 | — | 80% | +| minimax-m2 | auriko | — | $0.3 | — | $0.375 | — | +| minimax-m2-1 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | +| minimax-m2-1-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | +| minimax-m2-5 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | +| minimax-m2-5-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | +| minimax-m2-7 | auriko | — | $0.3 | — | $0.375 | — | +| minimax-m2-7-highspeed | auriko | — | $0.6 | — | $0.375 | — | +| o3-2025-04-16 | auriko | — | $2 | $0.5 | — | 75% | +| o3-mini-2025-01-31 | auriko | — | $1.1 | $0.55 | — | 50% | +| o4-mini-2025-04-16 | auriko | — | $1.1 | $0.275 | — | 75% | +| qwen-3-coder-480b-a35b-instruct | auriko | — | $0.3 | $0.1 | — | 67% | +| qwen-3-max | auriko | — | $1.2 | $0.24 | — | 80% | +| qwen-3-max-thinking | auriko | — | $1.2 | $0.24 | — | 80% | +| qwen-3-vl-235b-a22b-instruct | auriko | — | $0.2 | $0.11 | — | 45% | +| qwen-3.5-35b-a3b | auriko | — | $0.14 | $0.05 | — | 64% | +| seed-1.8 | auriko | — | $0.25 | $0.05 | — | 80% | +| seed-2.0-code | auriko | — | $0.5 | $0.1 | — | 80% | +| seed-2.0-mini | auriko | — | $0.1 | $0.02 | — | 80% | +| seed-2.0-pro | auriko | — | $0.5 | $0.1 | — | 80% | +| step-3.5-flash | auriko | — | $0.1 | $0.02 | — | 80% | +| deepseek-v3.2 | baidu | — | $0.252 | $0.0252 | — | 90% | +| deepseek-v4-flash | baidu | — | $0.126 | $0.0252 | — | 80% | +| deepseek-v4-pro | baidu | — | $1.521 | $0.126 | — | 92% | +| glm-5 | baidu | — | $0.7 | $0.14 | — | 80% | +| glm-5.1 | baidu | — | $0.98 | $0.182 | — | 81% | +| minimax-m2.5 | baidu | — | $0.27 | $0.027 | — | 90% | +| deepseek-v3-1 | baseten | — | $0.5 | $0.25 | — | 50% | +| deepseek-v4 | baseten | — | $1.74 | $0.145 | — | 92% | +| glm-4-7 | baseten | — | $0.6 | $0.12 | — | 80% | +| glm-5 | baseten | — | $0.95 | $0.2 | — | 79% | +| kimi-k2-5 | baseten | — | $0.6 | $0.12 | — | 80% | +| kimi-k2-6 | baseten | — | $1 | $0.2 | — | 80% | +| minimax-m2-5 | baseten | — | $0.3 | $0.06 | — | 80% | +| nvidia-nemotron-3-super | baseten | — | $0.3 | $0.06 | — | 80% | +| MiniMaxAI--MiniMax-M2.5-TEE | chutes | — | $0.15 | $0.075 | — | 50% | +| Qwen--Qwen3-235B-A22B-Thinking-2507 | chutes | — | $0.11 | $0.055 | — | 50% | +| Qwen--Qwen3-32B-TEE | chutes | — | $0.08 | $0.04 | — | 50% | +| Qwen--Qwen3.5-397B-A17B-TEE | chutes | — | $0.39 | $0.195 | — | 50% | +| Qwen--Qwen3.6-27B-TEE | chutes | — | $0.5 | $0.25 | — | 50% | +| deepseek-ai--DeepSeek-V3.2-TEE | chutes | — | $0.28 | $0.14 | — | 50% | +| google--gemma-4-31B-turbo-TEE | chutes | — | $0.13 | $0.065 | — | 50% | +| moonshotai--Kimi-K2.5-TEE | chutes | — | $0.44 | $0.22 | — | 50% | +| moonshotai--Kimi-K2.6-TEE | chutes | — | $0.74 | $0.37 | — | 50% | +| zai-org--GLM-5-TEE | chutes | — | $0.95 | $0.475 | — | 50% | +| zai-org--GLM-5-Turbo | chutes | — | $0.4891 | $0.24455 | — | 50% | +| zai-org--GLM-5.1-TEE | chutes | — | $1.05 | $0.525 | — | 50% | +| kimi-k2-thinking | clarifai | — | $1.5 | $0.15 | — | 90% | +| kimi-k2.5 | cloudflare | — | $0.6 | $0.1 | — | 83% | +| kimi-k2.6 | cloudflare | — | $0.95 | $0.16 | — | 83% | +| claude-4-5-sonnet | cortecs | — | $2.683 | $0.293 | — | 89% | +| claude-4-6-sonnet | cortecs | — | $2.869 | $0.287 | — | 90% | +| claude-haiku-4-5 | cortecs | — | $0.894 | $0.089 | — | 90% | +| claude-opus4-5 | cortecs | — | $4.769 | $0.477 | — | 90% | +| claude-opus4-6 | cortecs | — | $4.769 | $0.477 | — | 90% | +| claude-opus4-7 | cortecs | — | $4.769 | $0.477 | — | 90% | +| claude-sonnet-4 | cortecs | — | $2.601 | $0.26 | — | 90% | +| codestral-2508 | cortecs | — | $0.3 | $0.03 | — | 90% | +| deepseek-chat-v3.1 | cortecs | — | $0.177 | $0.044 | — | 75% | +| deepseek-r1-0528 | cortecs | — | $0.585 | $0.146 | — | 75% | +| deepseek-v3-0324 | cortecs | — | $0.266 | $0.067 | — | 75% | +| deepseek-v3.2 | cortecs | — | $0.266 | $0.067 | — | 75% | +| deepseek-v4-flash | cortecs | — | $0.133 | $0.033 | — | 75% | +| deepseek-v4-pro | cortecs | — | $1.553 | $0.388 | — | 75% | +| devstral-2512 | cortecs | — | $0.4 | $0.04 | — | 90% | +| devstral-medium-2507 | cortecs | — | $0.4 | $0.04 | — | 90% | +| devstral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-flash | cortecs | — | $0.268 | $0.026 | — | 90% | +| gemini-2.5-pro | cortecs | — | $1.342 | $0.217 | — | 84% | +| gemini-3.1-flash-lite | cortecs | — | $0.244 | $0.022 | — | 91% | +| glm-4.6 | cortecs | — | $0.355 | $0.089 | — | 75% | +| glm-4.7 | cortecs | — | $0.532 | $0.133 | — | 75% | +| glm-5 | cortecs | — | $0.887 | $0.222 | — | 75% | +| glm-5-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | +| glm-5.1 | cortecs | — | $1.242 | $0.311 | — | 75% | +| glm-5v-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | +| gpt-4.1 | cortecs | — | $1.968 | $0.49 | — | 75% | +| gpt-4.1-mini | cortecs | — | $0.39 | $0.12 | — | 69% | +| gpt-4.1-nano | cortecs | — | $0.1 | $0.05 | — | 50% | +| gpt-4o | cortecs | — | $2.387 | $1.194 | — | 50% | +| gpt-4o-mini | cortecs | — | $0.143 | $0.073 | — | 49% | +| gpt-5 | cortecs | — | $1.234 | $0.14 | — | 89% | +| gpt-5-mini | cortecs | — | $0.25 | $0.05 | — | 80% | +| gpt-5-nano | cortecs | — | $0.054 | $0.017 | — | 69% | +| gpt-5.1 | cortecs | — | $1.234 | $0.14 | — | 89% | +| gpt-5.4 | cortecs | — | $2.601 | $0.217 | — | 92% | +| gpt-oss-120b | cortecs | — | $0.035 | $0.009 | — | 74% | +| gpt-oss-20b | cortecs | — | $0.027 | $0.007 | — | 74% | +| kimi-k2.5 | cortecs | — | $0.444 | $0.111 | — | 75% | +| kimi-k2.6 | cortecs | — | $0.694 | $0.173 | — | 75% | +| llama-3.3-70b-instruct | cortecs | — | $0.089 | $0.023 | — | 74% | +| llama-4-maverick | cortecs | — | $0.124 | $0.03 | — | 76% | +| magistral-medium-2509 | cortecs | — | $2 | $0.2 | — | 90% | +| magistral-small-2509 | cortecs | — | $0.5 | $0.05 | — | 90% | +| minimax-m2 | cortecs | — | $0.222 | $0.055 | — | 75% | +| minimax-m2.5 | cortecs | — | $0.266 | $0.027 | — | 90% | +| minimax-m2.7 | cortecs | — | $0.444 | $0.111 | — | 75% | +| ministral-14b-2512 | cortecs | — | $0.2 | $0.02 | — | 90% | +| ministral-3b-2512 | cortecs | — | $0.1 | $0.01 | — | 90% | +| ministral-8b-2512 | cortecs | — | $0.15 | $0.015 | — | 90% | +| mistral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | +| mistral-large-2512 | cortecs | — | $0.5 | $0.05 | — | 90% | +| mistral-medium-2508 | cortecs | — | $0.4 | $0.04 | — | 90% | +| mistral-medium-3.5 | cortecs | — | $1.25 | $0.125 | — | 90% | +| mistral-nemo-instruct-2407 | cortecs | — | $0.13 | $0.013 | — | 90% | +| mistral-small-2506 | cortecs | — | $0.1 | $0.01 | — | 90% | +| mistral-small-2603 | cortecs | — | $0.128 | $0.013 | — | 90% | +| nemotron-3-super-120b-a12b | cortecs | — | $0.266 | $0.067 | — | 75% | +| pixtral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | +| qwen-2.5-72b-instruct | cortecs | — | $0.062 | $0.016 | — | 74% | +| qwen3-235b-a22b-instruct-2507 | cortecs | — | $0.062 | $0.016 | — | 74% | +| qwen3-coder-30b-a3b-instruct | cortecs | — | $0.053 | $0.013 | — | 75% | +| qwen3-vl-235b-a22b | cortecs | — | $0.186 | $0.047 | — | 75% | +| qwen3.5-122b-a10b | cortecs | — | $0.444 | $0.111 | — | 75% | +| voxtral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | +| databricks-claude-haiku-4-5 | databricks | — | $1 | $0.1 | $1.25 | 90% | +| databricks-claude-opus-4-1 | databricks | — | $15 | $1.5 | $18.75 | 90% | +| databricks-claude-opus-4-5 | databricks | — | $5 | $0.5 | $6.25 | 90% | +| databricks-claude-sonnet-4 | databricks | — | $3 | $0.3 | $3.75 | 90% | +| databricks-claude-sonnet-4-5 | databricks | — | $3 | $0.3 | $3.75 | 90% | +| databricks-gemini-3-1-flash-lite | databricks | — | $0.25 | $0.02 | $0.25 | 92% | +| databricks-gemini-3-1-pro | databricks | — | $2.5 | $0.25 | $2.5 | 90% | +| databricks-gemini-3-flash | databricks | — | $0.63 | $0.06 | $0.63 | 90% | +| databricks-gpt-5 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | +| databricks-gpt-5-1 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | +| databricks-gpt-5-1-codex-max | databricks | — | $1.25 | $0.13 | $1.25 | 90% | +| databricks-gpt-5-1-codex-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | +| databricks-gpt-5-2 | databricks | — | $1.75 | $0.18 | $1.75 | 90% | +| databricks-gpt-5-2-codex | databricks | — | $1.75 | $0.18 | $1.75 | 90% | +| databricks-gpt-5-4 | databricks | — | $2.5 | $0.25 | $2.5 | 90% | +| databricks-gpt-5-4-mini | databricks | — | $0.75 | $0.07 | $0.75 | 91% | +| databricks-gpt-5-4-nano | databricks | — | $0.2 | $0.02 | $0.2 | 90% | +| databricks-gpt-5-5 | databricks | — | $5 | $0.5 | $5 | 90% | +| databricks-gpt-5-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | +| databricks-gpt-5-nano | databricks | — | $0.05 | Free | $0.05 | — | +| deepseek-r1-0528 | deepinfra | — | $0.5 | $0.35 | — | 30% | +| deepseek-v3-0324 | deepinfra | — | $0.2 | $0.135 | — | 32% | +| deepseek-v3.1 | deepinfra | — | $0.21 | $0.13 | — | 38% | +| deepseek-v3.1-terminus | deepinfra | — | $0.27 | $0.13 | — | 52% | +| deepseek-v3.2 | deepinfra | — | $0.26 | $0.13 | — | 50% | +| deepseek-v4-flash | deepinfra | — | $0.14 | $0.028 | — | 80% | +| deepseek-v4-pro | deepinfra | — | $1.74 | $0.145 | — | 92% | +| glm-4.6 | deepinfra | — | $0.43 | $0.08 | — | 81% | +| glm-4.7 | deepinfra | — | $0.4 | $0.08 | — | 80% | +| glm-4.7-flash | deepinfra | — | $0.06 | $0.01 | — | 83% | +| glm-5 | deepinfra | — | $0.6 | $0.12 | — | 80% | +| glm-5.1 | deepinfra | — | $1.05 | $0.205 | — | 80% | +| kimi-k2.5 | deepinfra | — | $0.45 | $0.07 | — | 84% | +| kimi-k2.6 | deepinfra | — | $0.75 | $0.15 | — | 80% | +| mimo-v2.5 | deepinfra | — | $0.4 | $0.08 | — | 80% | +| mimo-v2.5-pro | deepinfra | — | $1 | $0.2 | — | 80% | +| minimax-m2.5 | deepinfra | — | $0.15 | $0.03 | — | 80% | +| qwen3-235b-a22b-thinking-2507 | deepinfra | — | $0.23 | $0.2 | — | 13% | +| qwen3-coder-480b-a35b-instruct-turbo | deepinfra | — | $0.3 | $0.1 | — | 67% | +| qwen3-max | deepinfra | — | $1.2 | $0.24 | — | 80% | +| qwen3-max-thinking | deepinfra | — | $1.2 | $0.24 | — | 80% | +| qwen3-vl-235b-a22b-instruct | deepinfra | — | $0.2 | $0.11 | — | 45% | +| qwen3.5-35b-a3b | deepinfra | — | $0.14 | $0.05 | — | 64% | +| qwen3.5-397b-a17b | deepinfra | — | $0.49 | $0.3 | — | 39% | +| seed-1.8 | deepinfra | — | $0.25 | $0.05 | — | 80% | +| seed-2.0-code | deepinfra | — | $0.5 | $0.1 | — | 80% | +| seed-2.0-mini | deepinfra | — | $0.1 | $0.02 | — | 80% | +| seed-2.0-pro | deepinfra | — | $0.5 | $0.1 | — | 80% | +| step-3.5-flash | deepinfra | — | $0.1 | $0.02 | — | 80% | +| deepseek-chat | deepseek | — | $0.14 | $0.0028 | — | 98% | +| deepseek-reasoner | deepseek | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-flash | deepseek | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | deepseek | — | $0.435 | $0.003625 | — | 99% | +| arcee-trinity-large-thinking | digitalocean | — | $0.25 | $0.06 | — | 76% | +| anthropic--claude-3-5-haiku-20241022 | fastrouter | — | $0.8 | $0.08 | $1 | 90% | +| anthropic--claude-4.5-sonnet | fastrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-haiku-4.5 | fastrouter | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4-20250514 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.1 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.5 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.6 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.7 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-sonnet-4-20250514 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4.6 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | +| deepseek--deepseek-v3.2 | fastrouter | — | $0.27 | $0.216 | — | 20% | +| deepseek--deepseek-v4-pro | fastrouter | — | $2.1 | $0.2 | — | 90% | +| google--gemini-2.0-flash-001 | fastrouter | — | $0.1 | $0.025 | $0.1833 | 75% | +| google--gemini-2.5-flash | fastrouter | — | $0.3 | $0.075 | $0.3833 | 75% | +| google--gemini-2.5-pro | fastrouter | — | $1.25 | $0.31 | $1.625 | 75% | +| google--gemini-3-flash-preview | fastrouter | — | $0.5 | $0.05 | — | 90% | +| google--gemini-3.1-flash-lite-preview | fastrouter | — | $0.25 | $0.025 | $0.083333 | 90% | +| google--gemini-3.1-pro-preview | fastrouter | — | $2 | $0.31 | $1.625 | 84% | +| minimax--minimax-m2.1 | fastrouter | — | $0.27 | $0.03 | — | 89% | +| minimax--minimax-m2.5 | fastrouter | — | $0.3 | $0.03 | — | 90% | +| minimax--minimax-m2.5-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | +| minimax--minimax-m2.7 | fastrouter | — | $0.3 | $0.06 | — | 80% | +| minimax--minimax-m2.7-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | +| moonshotai--kimi-k2-thinking | fastrouter | — | $0.6 | $0.15 | — | 75% | +| moonshotai--kimi-k2.6 | fastrouter | — | $0.55 | $0.15 | — | 73% | +| openai--gpt-4.1 | fastrouter | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-mini | fastrouter | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-nano | fastrouter | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4o | fastrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-08-06 | fastrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-11-20 | fastrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-mini | fastrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-mini-2024-07-18 | fastrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-realtime-preview-2024-12-17 | fastrouter | — | $5 | $2.5 | — | 50% | +| openai--gpt-4o-realtime-preview-2025-06-03 | fastrouter | — | $5 | $2.5 | — | 50% | +| openai--gpt-5 | fastrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-mini | fastrouter | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-nano | fastrouter | — | $0.05 | $0.005 | — | 90% | +| openai--gpt-5.1 | fastrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.2 | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.4 | fastrouter | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | fastrouter | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | fastrouter | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | fastrouter | — | $5 | $0.5 | — | 90% | +| openai--gpt-realtime | fastrouter | — | $4 | $0.5 | — | 88% | +| openai--gpt-realtime-1.5 | fastrouter | — | $4 | $0.4 | — | 90% | +| openai--gpt-realtime-mini | fastrouter | — | $0.6 | $0.06 | — | 90% | +| openai--o1 | fastrouter | — | $15 | $7.5 | — | 50% | +| openai--o3 | fastrouter | — | $2 | $0.5 | — | 75% | +| openai--o3-mini | fastrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o3-mini-high | fastrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o4-mini | fastrouter | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-high | fastrouter | — | $1.1 | $0.275 | — | 75% | +| x-ai--grok-3-beta | fastrouter | — | $5 | $1.25 | — | 75% | +| x-ai--grok-3-mini-beta | fastrouter | — | $0.6 | $0.15 | — | 75% | +| x-ai--grok-4 | fastrouter | — | $3 | $0.75 | — | 75% | +| x-ai--grok-4.20-beta | fastrouter | — | $2 | $0.2 | — | 90% | +| x-ai--grok-4.3 | fastrouter | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-code-fast-1 | fastrouter | — | $0.2 | $0.02 | — | 90% | +| z-ai--glm-4.5-air | fastrouter | — | $0.2 | $0.03 | — | 85% | +| LGAI-EXAONE--K-EXAONE-236B-A23B | friendli | — | $0.2 | $0.1 | — | 50% | +| MiniMaxAI--MiniMax-M2.5 | friendli | — | $0.3 | $0.06 | — | 80% | +| deepseek-ai--DeepSeek-V3.2 | friendli | — | $0.5 | $0.25 | — | 50% | +| zai-org--GLM-5 | friendli | — | $1 | $0.5 | — | 50% | +| zai-org--GLM-5.1 | friendli | — | $1.4 | $0.26 | — | 81% | +| gemini-1.5-flash | google | — | $0.075 | $0.01875 | $0.2625 | 75% | +| gemini-1.5-flash-8b | google | — | $0.075 | $0.01875 | $0.2625 | 75% | +| gemini-1.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | +| gemini-2.0-flash | google | — | $0.1 | $0.025 | $0.35 | 75% | +| gemini-2.0-flash-lite | google | — | $0.075 | $0.01875 | $0.2625 | 75% | +| gemini-2.5-flash | google | — | $0.15 | $0.0375 | $0.525 | 75% | +| gemini-2.5-flash-lite | google | — | $0.1 | $0.025 | $0.35 | 75% | +| gemini-2.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | +| claude-haiku-4-5 | google-vertex | — | $1 | $0.1 | — | 90% | +| claude-opus-4 | google-vertex | — | $15 | $1.5 | — | 90% | +| claude-opus-4-1 | google-vertex | — | $15 | $1.5 | — | 90% | +| claude-opus-4-5 | google-vertex | — | $5 | $0.5 | — | 90% | +| claude-opus-4-6 | google-vertex | — | $5 | $0.5 | — | 90% | +| claude-opus-4-7 | google-vertex | — | $5 | $0.5 | — | 90% | +| claude-sonnet-4 | google-vertex | — | $3 | $0.3 | — | 90% | +| claude-sonnet-4-5 | google-vertex | — | $3 | $0.3 | — | 90% | +| claude-sonnet-4-6 | google-vertex | — | $3 | $0.3 | — | 90% | +| deepseek-v3-1 | google-vertex | — | $0.6 | $0.06 | — | 90% | +| deepseek-v3-2 | google-vertex | — | $0.56 | $0.056 | — | 90% | +| gemini-2-5-flash | google-vertex | — | $0.3 | $0.03 | — | 90% | +| gemini-2-5-flash-lite | google-vertex | — | $0.1 | $0.01 | — | 90% | +| gemini-2-5-pro | google-vertex | — | $1.25 | $0.13 | — | 90% | +| gemini-3-1-flash-lite | google-vertex | — | $0.25 | $0.025 | — | 90% | +| gemini-3-flash | google-vertex | — | $0.5 | $0.05 | — | 90% | +| gemini-3-pro | google-vertex | — | $2 | $0.2 | — | 90% | +| gemma-4-26b | google-vertex | — | $0.15 | $0.015 | — | 90% | +| glm-5 | google-vertex | — | $1 | $0.1 | — | 90% | +| gpt-oss-20b | google-vertex | — | $0.07 | $0.007 | — | 90% | +| grok-4-1-fast | google-vertex | — | $0.2 | $0.05 | — | 75% | +| grok-4-20 | google-vertex | — | $2 | $0.2 | — | 90% | +| kimi-k2-thinking | google-vertex | — | $0.6 | $0.06 | — | 90% | +| minimax-m2 | google-vertex | — | $0.3 | $0.03 | — | 90% | +| qwen3-coder-480b-a35b | google-vertex | — | $0.22 | $0.022 | — | 90% | +| gpt-oss-120b | groq | — | $0.15 | $0.075 | — | 50% | +| gpt-oss-20b | groq | — | $0.075 | $0.0375 | — | 50% | +| kimi-k2-instruct-0905 | groq | — | $1 | $0.5 | — | 50% | +| deepseek--deepseek-v4-flash | hpc-ai | — | $0.14 | $0.028 | — | 80% | +| deepseek--deepseek-v4-pro | hpc-ai | — | $1.74 | $0.145 | — | 92% | +| minimax--minimax-m2.5 | hpc-ai | — | $0.3 | $0.03 | — | 90% | +| moonshotai--kimi-k2.5 | hpc-ai | — | $0.6 | $0.1 | — | 83% | +| moonshotai--kimi-k2.6 | hpc-ai | — | $0.95 | $0.16 | — | 83% | +| xiaomi--mimo-v2.5 | hpc-ai | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | hpc-ai | — | $1 | $0.2 | — | 80% | +| zai--glm-5.1 | hpc-ai | — | $1.4 | $0.26 | — | 81% | +| mercury | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-coder | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-edit | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-edit-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| claude-opus-4-6 | jiekou | — | $2.75 | $0.475 | — | 83% | +| claude-opus-4-6-dd | jiekou | — | $2.85 | $0.275 | — | 90% | +| claude-opus-4-7 | jiekou | — | $4.75 | $0.475 | — | 90% | +| claude-sonnet-4-6 | jiekou | — | $1.65 | $0.285 | — | 83% | +| claude-sonnet-4-6-dd | jiekou | — | $4.75 | $0.165 | — | 97% | +| doubao-seed-1-8-251228 | jiekou | — | $0.11 | $0.0221 | — | 80% | +| gemini-2.5-flash | jiekou | — | $0.095 | $0.0285 | — | 70% | +| gemini-2.5-flash-lite | jiekou | — | $1.1875 | $0.0095 | — | 99% | +| gemini-2.5-pro | jiekou | — | $0.285 | $0.1187 | — | 58% | +| gemini-3-flash-preview | jiekou | — | $0.095 | $0.0475 | — | 50% | +| gemini-3.1-flash-lite-preview | jiekou | — | $1.9 | $0.0237 | — | 99% | +| gemini-3.1-pro-preview | jiekou | — | $0.475 | $0.19 | — | 60% | +| gpt-4.1 | jiekou | — | $9.5 | $0.5 | — | 95% | +| gpt-4.1-mini | jiekou | — | $0.1 | $0.1 | — | 0% | +| gpt-4o | jiekou | — | $1.9 | $1.1875 | — | 38% | +| gpt-5-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | +| gpt-5-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | +| gpt-5-nano | jiekou | — | $0.2375 | $0.0047 | — | 98% | +| gpt-5.1-chat-latest | jiekou | — | $1.1875 | $0.1187 | — | 90% | +| gpt-5.1-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | +| gpt-5.1-codex-max | jiekou | — | $0.2375 | $0.1187 | — | 50% | +| gpt-5.1-codex-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | +| gpt-5.2-chat-latest | jiekou | — | $1.1875 | $0.1662 | — | 86% | +| gpt-5.2-codex | jiekou | — | $1.6625 | $0.175 | — | 89% | +| gpt-5.3-chat-latest | jiekou | — | $1.6625 | $0.1662 | — | 90% | +| gpt-5.3-codex | jiekou | — | $1.75 | $0.1662 | — | 91% | +| gpt-5.4 | jiekou | — | $1.6625 | $0.2375 | — | 86% | +| gpt-5.4-nano | jiekou | — | $0.7125 | $0.019 | — | 97% | +| gpt-5.5 | jiekou | — | $0.19 | $0.5 | — | -163% | +| gpt-5.5-light | jiekou | — | $5 | $0.025 | — | 100% | +| grok-code-fast-1 | jiekou | — | $0.19 | $0.019 | — | 90% | +| minimax-minimax-m2.1 | jiekou | — | $0.55 | $0.03 | — | 95% | +| moonshotai-kimi-k2.5 | jiekou | — | $0.6 | $0.1 | — | 83% | +| o3 | jiekou | — | $1.045 | $0.475 | — | 55% | +| o3-mini | jiekou | — | $1.045 | $0.5225 | — | 50% | +| zai-org-glm-4.7-flash | jiekou | — | $0.6 | $0.01 | — | 98% | +| claude-3-7-sonnet-20250219 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-3-opus | llmgateway | — | $15 | $1.5 | $18.75 | 90% | +| claude-haiku-4-5 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | +| claude-haiku-4-5-20251001 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | +| claude-opus-4-1-20250805 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-20250514 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-5-20251101 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-6 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-7 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | +| claude-sonnet-4-20250514 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5-20250929 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-6 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| deepseek-v3.1 | llmgateway | — | $0.56 | $0.112 | — | 80% | +| deepseek-v3.2 | llmgateway | — | $0.28 | $0.028 | — | 90% | +| deepseek-v4-flash | llmgateway | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | llmgateway | — | $0.435 | $0.003625 | — | 99% | +| gemini-2.5-flash | llmgateway | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-image | llmgateway | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-lite | llmgateway | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-pro | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gemini-3-flash-preview | llmgateway | — | $0.5 | $0.05 | — | 90% | +| gemini-3-pro-image-preview | llmgateway | — | $2 | $0.2 | — | 90% | +| gemini-3.1-flash-lite | llmgateway | — | $0.25 | $0.025 | $0.08333 | 90% | +| gemini-3.1-pro-preview | llmgateway | — | $2 | $0.2 | — | 90% | +| gemini-3.5-flash | llmgateway | — | $1.5 | $0.15 | $0.08333 | 90% | +| gemini-pro-latest | llmgateway | — | $2 | $0.2 | — | 90% | +| glm-4.5 | llmgateway | — | $0.6 | $0.11 | — | 82% | +| glm-4.5-air | llmgateway | — | $0.2 | $0.03 | — | 85% | +| glm-4.5-airx | llmgateway | — | $1.1 | $0.22 | — | 80% | +| glm-4.5-x | llmgateway | — | $2.2 | $0.45 | — | 80% | +| glm-4.5v | llmgateway | — | $0.6 | $0.11 | — | 82% | +| glm-4.6v | llmgateway | — | $0.3 | $0.05 | — | 83% | +| glm-4.6v-flashx | llmgateway | — | $0.04 | $0.004 | — | 90% | +| glm-4.7 | llmgateway | — | $0.6 | $0.11 | — | 82% | +| glm-4.7-flashx | llmgateway | — | $0.07 | $0.01 | — | 86% | +| glm-5.1 | llmgateway | — | $1.4 | $0.26 | — | 81% | +| gpt-4.1 | llmgateway | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini | llmgateway | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano | llmgateway | — | $0.1 | $0.025 | — | 75% | +| gpt-4o | llmgateway | — | $2.5 | $1.25 | — | 50% | +| gpt-5 | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gpt-5-chat-latest | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gpt-5-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | +| gpt-5-nano | llmgateway | — | $0.05 | $0.005 | — | 90% | +| gpt-5.1 | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gpt-5.1-codex-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | +| gpt-5.2 | llmgateway | — | $1.75 | $0.175 | — | 90% | +| gpt-5.2-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | +| gpt-5.3-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | +| gpt-5.4 | llmgateway | — | $2.5 | $0.25 | — | 90% | +| gpt-5.4-mini | llmgateway | — | $0.75 | $0.075 | — | 90% | +| gpt-5.4-nano | llmgateway | — | $0.2 | $0.02 | — | 90% | +| gpt-5.5 | llmgateway | — | $5 | $0.5 | — | 90% | +| grok-4 | llmgateway | — | $3 | $0.75 | — | 75% | +| grok-4-20-beta-0309-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-20-beta-0309-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-20-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-20-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-3 | llmgateway | — | $1.25 | $0.3125 | — | 75% | +| kimi-k2-thinking | llmgateway | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking-turbo | llmgateway | — | $1.15 | $0.15 | — | 87% | +| kimi-k2.5 | llmgateway | — | $0.6 | $0.1 | — | 83% | +| kimi-k2.6 | llmgateway | — | $0.95 | $0.16 | — | 83% | +| mimo-v2-flash | llmgateway | — | $0.1 | $0.01 | — | 90% | +| mimo-v2-omni | llmgateway | — | $0.4 | $0.08 | — | 80% | +| mimo-v2-pro | llmgateway | — | $1 | $0.2 | — | 80% | +| mimo-v2.5 | llmgateway | — | $0.4 | $0.08 | — | 80% | +| mimo-v2.5-pro | llmgateway | — | $1 | $0.2 | — | 80% | +| minimax-m2 | llmgateway | — | $0.2 | $0.03 | — | 85% | +| minimax-m2.5-highspeed | llmgateway | — | $0.6 | $0.03 | — | 95% | +| minimax-m2.7 | llmgateway | — | $0.3 | $0.06 | — | 80% | +| minimax-m2.7-highspeed | llmgateway | — | $0.6 | $0.06 | — | 90% | +| o1 | llmgateway | — | $15 | $7.5 | — | 50% | +| o3 | llmgateway | — | $2 | $0.5 | — | 75% | +| o3-mini | llmgateway | — | $1.1 | $0.55 | — | 50% | +| o4-mini | llmgateway | — | $1.1 | $0.275 | — | 75% | +| qwen-flash | llmgateway | — | $0.05 | $0.01 | $0.0625 | 80% | +| qwen-plus | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | +| qwen-plus-latest | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | +| qwen3-coder-flash | llmgateway | — | $0.3 | $0.06 | $0.375 | 80% | +| qwen3-coder-next | llmgateway | — | $0.108 | $0.06 | — | 44% | +| qwen3-coder-plus | llmgateway | — | $6 | $1.2 | $7.5 | 80% | +| qwen3-max | llmgateway | — | $3 | $0.6 | $3.75 | 80% | +| qwen3-max-2026-01-23 | llmgateway | — | $1.2 | $0.24 | $1.5 | 80% | +| qwen3-vl-flash | llmgateway | — | $0.05 | $0.01 | — | 80% | +| qwen3-vl-plus | llmgateway | — | $0.2 | $0.04 | $0.25 | 80% | +| qwen3.6-max-preview | llmgateway | — | $1.3 | $0.13 | — | 90% | +| qwen3.6-plus | llmgateway | — | $0.5 | $0.05 | — | 90% | +| seed-1-6-250615 | llmgateway | — | $0.25 | $0.05 | — | 80% | +| seed-1-6-250915 | llmgateway | — | $0.25 | $0.05 | — | 80% | +| seed-1-6-flash-250715 | llmgateway | — | $0.07 | $0.015 | — | 79% | +| seed-1-8-251228 | llmgateway | — | $0.25 | $0.05 | — | 80% | +| aion-labs--aion-2.0 | martian | — | $0.8 | $0.2 | — | 75% | +| alibaba--tongyi-deepresearch-30b-a3b | martian | — | $0.09 | $0.09 | — | 0% | +| amazon--nova-premier-v1 | martian | — | $2.5 | $0.625 | — | 75% | +| anthropic--claude-haiku-4-5 | martian | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-haiku-4-5-20251001 | martian | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4-0 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-1 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-1-20250805 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-20250514 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-5 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-5-20251101 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-6 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-7 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-sonnet-4-0 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-20250514 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-5 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-5-20250929 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-6 | martian | — | $3 | $0.3 | $3.75 | 90% | +| arcee-ai--trinity-large-thinking | martian | — | $0.22 | $0.06 | — | 73% | +| bytedance--ui-tars-1.5-7b | martian | — | $0.1 | $0.1 | — | 0% | +| deepinfra--deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | +| deepinfra--deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | +| deepinfra--deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | +| deepinfra--deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | +| deepinfra--deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | +| deepinfra--google--gemini-2.5-flash | martian | — | $0.3 | $0.03 | $0.083333 | 90% | +| deepinfra--google--gemini-2.5-pro | martian | — | $1.25 | $0.125 | $0.375 | 90% | +| deepinfra--z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | +| deepinfra--z-ai--glm-4.7-flash | martian | — | $0.06 | $0.01 | — | 83% | +| deepinfra--z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | +| deepinfra--z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | +| deepseek--deepseek-chat-v3-0324 | martian | — | $0.2 | $0.135 | — | 32% | +| deepseek--deepseek-chat-v3.1 | martian | — | $0.21 | $0.13 | — | 38% | +| deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | +| deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | +| deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | +| deepseek--deepseek-v3.2-speciale | martian | — | $0.287 | $0.058 | — | 80% | +| deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | +| deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | +| ibm-granite--granite-4.1-8b | martian | — | $0.05 | $0.05 | — | 0% | +| inception--mercury-2 | martian | — | $0.25 | $0.025 | — | 90% | +| inclusionai--ling-2.6-1t | martian | — | $0.3 | $0.06 | — | 80% | +| inclusionai--ling-2.6-flash | martian | — | $0.08 | $0.016 | — | 80% | +| kwaipilot--kat-coder-pro-v2 | martian | — | $0.3 | $0.06 | — | 80% | +| microsoft--phi-4-mini-instruct | martian | — | $0.08 | $0.08 | — | 0% | +| minimax--minimax-m2 | martian | — | $0.255 | $0.03 | — | 88% | +| minimax--minimax-m2-her | martian | — | $0.3 | $0.03 | — | 90% | +| minimax--minimax-m2.1 | martian | — | $0.29 | $0.03 | — | 90% | +| mistralai--codestral-2508 | martian | — | $0.3 | $0.03 | — | 90% | +| mistralai--devstral-medium | martian | — | $0.4 | $0.04 | — | 90% | +| mistralai--devstral-small | martian | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-14b-2512 | martian | — | $0.2 | $0.02 | — | 90% | +| mistralai--ministral-3b-2512 | martian | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-8b-2512 | martian | — | $0.15 | $0.015 | — | 90% | +| mistralai--mistral-large | martian | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2407 | martian | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2411 | martian | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2512 | martian | — | $0.5 | $0.05 | — | 90% | +| mistralai--mistral-medium-3 | martian | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-medium-3.1 | martian | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-saba | martian | — | $0.2 | $0.02 | — | 90% | +| mistralai--mistral-small-2603 | martian | — | $0.15 | $0.015 | — | 90% | +| mistralai--mixtral-8x22b-instruct | martian | — | $2 | $0.2 | — | 90% | +| mistralai--pixtral-large-2411 | martian | — | $2 | $0.2 | — | 90% | +| mistralai--voxtral-small-24b-2507 | martian | — | $0.1 | $0.01 | — | 90% | +| moonshotai--kimi-k2-thinking | martian | — | $0.6 | $0.15 | — | 75% | +| moonshotai--kimi-k2.5 | martian | — | $0.4 | $0.05 | — | 88% | +| moonshotai--kimi-k2.6 | martian | — | $0.74 | $0.25 | — | 66% | +| openai--gpt-4.1 | martian | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-2025-04-14 | martian | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-mini | martian | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-mini-2025-04-14 | martian | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-nano | martian | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4.1-nano-2025-04-14 | martian | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4o-2024-08-06 | martian | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-11-20 | martian | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-mini | martian | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-mini-2024-07-18 | martian | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-5 | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-2025-08-07 | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-codex | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-mini | martian | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-mini-2025-08-07 | martian | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-nano | martian | — | $0.05 | $0.01 | — | 80% | +| openai--gpt-5-nano-2025-08-07 | martian | — | $0.05 | $0.01 | — | 80% | +| openai--gpt-5.1 | martian | — | $1.25 | $0.13 | — | 90% | +| openai--gpt-5.1-2025-11-13 | martian | — | $1.25 | $0.13 | — | 90% | +| openai--gpt-5.1-codex | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-max | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-mini | martian | — | $0.25 | $0.03 | — | 88% | +| openai--gpt-5.2 | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-2025-12-11 | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-codex | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-codex | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.4 | martian | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-2026-03-05 | martian | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | martian | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-mini-2026-03-17 | martian | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | martian | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.4-nano-2026-03-17 | martian | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | martian | — | $5 | $0.5 | — | 90% | +| openai--gpt-5.5-2026-04-23 | martian | — | $5 | $0.5 | — | 90% | +| openai--o1 | martian | — | $15 | $7.5 | — | 50% | +| openai--o1-2024-12-17 | martian | — | $15 | $7.5 | — | 50% | +| openai--o3 | martian | — | $2 | $0.5 | — | 75% | +| openai--o3-2025-04-16 | martian | — | $2 | $0.5 | — | 75% | +| openai--o3-mini | martian | — | $1.1 | $0.55 | — | 50% | +| openai--o3-mini-2025-01-31 | martian | — | $1.1 | $0.55 | — | 50% | +| openai--o4-mini | martian | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-2025-04-16 | martian | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-deep-research | martian | — | $2 | $0.5 | — | 75% | +| openai--o4-mini-deep-research-2025-06-26 | martian | — | $2 | $0.5 | — | 75% | +| qwen--qwen-plus | martian | — | $0.26 | $0.052 | $0.325 | 80% | +| qwen--qwen-plus-2025-07-28 | martian | — | $0.26 | — | $0.325 | — | +| qwen--qwen-plus-2025-07-28--thinking | martian | — | $0.26 | — | $0.325 | — | +| qwen--qwen3-30b-a3b-thinking-2507 | martian | — | $0.08 | $0.08 | — | 0% | +| qwen--qwen3-8b | martian | — | $0.05 | $0.05 | — | 0% | +| qwen--qwen3-coder-flash | martian | — | $0.195 | $0.039 | $0.24375 | 80% | +| qwen--qwen3-coder-next | martian | — | $0.11 | $0.07 | — | 36% | +| qwen--qwen3-coder-plus | martian | — | $0.65 | $0.13 | $0.8125 | 80% | +| qwen--qwen3-max | martian | — | $0.78 | $0.156 | $0.975 | 80% | +| qwen--qwen3-vl-235b-a22b-instruct | martian | — | $0.2 | $0.11 | — | 45% | +| qwen--qwen3.5-35b-a3b | martian | — | $0.14 | $0.05 | — | 64% | +| qwen--qwen3.5-flash-02-23 | martian | — | $0.065 | — | $0.08125 | — | +| qwen--qwen3.5-plus-02-15 | martian | — | $0.26 | — | $0.325 | — | +| qwen--qwen3.6-35b-a3b | martian | — | $0.15 | $0.05 | — | 67% | +| qwen--qwen3.6-flash | martian | — | $0.25 | — | $0.3125 | — | +| qwen--qwen3.6-max-preview | martian | — | $1.04 | — | $1.3 | — | +| qwen--qwen3.6-plus | martian | — | $0.325 | — | $0.40625 | — | +| tencent--hy3-preview | martian | — | $0.066 | $0.029 | — | 56% | +| thedrummer--cydonia-24b-v4.1 | martian | — | $0.3 | $0.15 | — | 50% | +| thedrummer--skyfall-36b-v2 | martian | — | $0.55 | $0.25 | — | 55% | +| x-ai--grok-3 | martian | — | $3 | $0.75 | — | 75% | +| x-ai--grok-3-mini | martian | — | $0.3 | $0.075 | — | 75% | +| x-ai--grok-4 | martian | — | $3 | $0.75 | — | 75% | +| x-ai--grok-4.20-multi-agent | martian | — | $2 | $0.2 | — | 90% | +| x-ai--grok-4.3 | martian | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-code-fast-1 | martian | — | $0.2 | $0.02 | — | 90% | +| xiaomi--mimo-v2-flash | martian | — | $0.1 | $0.01 | — | 90% | +| xiaomi--mimo-v2-omni | martian | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2-pro | martian | — | $1 | $0.2 | — | 80% | +| xiaomi--mimo-v2.5 | martian | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | martian | — | $1 | $0.2 | — | 80% | +| z-ai--glm-4.5 | martian | — | $0.6 | $0.11 | — | 82% | +| z-ai--glm-4.5-air | martian | — | $0.13 | $0.025 | — | 81% | +| z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | +| z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | +| z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | +| MiniMax-M2 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | +| MiniMax-M2.1 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | +| MiniMax-M2.1-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | +| MiniMax-M2.5 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | +| MiniMax-M2.5-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | +| MiniMax-M2.7 | minimax | — | $2.1 | $0.42 | $2.625 | 80% | +| MiniMax-M2.7-highspeed | minimax | — | $4.2 | $0.42 | $2.625 | 90% | +| kimi-k2-0711-preview | moonshotai | — | $4 | $1 | — | 75% | +| kimi-k2-0905-preview | moonshotai | — | $4 | $1 | — | 75% | +| kimi-k2-thinking | moonshotai | — | $4 | $1 | — | 75% | +| kimi-k2-thinking-turbo | moonshotai | — | $8 | $1 | — | 88% | +| kimi-k2-turbo-preview | moonshotai | — | $8 | $1 | — | 88% | +| kimi-k2.5 | moonshotai | — | $4 | $0.7 | — | 82% | +| kimi-k2.6 | moonshotai | — | $6.5 | $1.1 | — | 83% | +| kimi-k2.6-long | moonshotai | — | $6.5 | $1.1 | — | 83% | +| kimi-vl-a3b | moonshotai | — | $4 | $0.7 | — | 82% | +| kimi-vl-a3b-thinking | moonshotai | — | $4 | $0.7 | — | 82% | +| aion-labs--aion-2.5 | nanogpt | — | $1 | $0.35 | — | 65% | +| alibaba--qwen3.6-flash | nanogpt | — | $0.19 | $0.02 | $0.24 | 89% | +| deepseek--deepseek-latest | nanogpt | — | $1.1 | $0.11 | — | 90% | +| deepseek--deepseek-v4-flash | nanogpt | — | $0.14 | $0.028 | — | 80% | +| deepseek--deepseek-v4-flash--thinking | nanogpt | — | $0.14 | $0.028 | — | 80% | +| deepseek--deepseek-v4-pro | nanogpt | — | $1.1 | $0.11 | — | 90% | +| deepseek--deepseek-v4-pro--thinking | nanogpt | — | $1.1 | $0.11 | — | 90% | +| deepseek--deepseek-v4-pro-cheaper | nanogpt | — | $0.435 | $0.0435 | — | 90% | +| deepseek--deepseek-v4-pro-cheaper--thinking | nanogpt | — | $0.435 | $0.0435 | — | 90% | +| ernie-5.1 | nanogpt | — | $0.75 | $0.75 | — | 0% | +| ernie-5.1--thinking | nanogpt | — | $0.75 | $0.75 | — | 0% | +| google--gemini-3-flash-preview | nanogpt | — | $0.5 | — | $0.08333 | — | +| google--gemini-3.1-flash-lite | nanogpt | — | $0.25 | — | $0.08333 | — | +| google--gemini-3.1-pro-preview | nanogpt | — | $2 | — | $0.375 | — | +| google--gemini-3.1-pro-preview-customtools | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-pro-preview-high | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-pro-preview-low | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.5-flash | nanogpt | — | $1.5 | — | $0.08333 | — | +| google--gemini-flash-lite-latest | nanogpt | — | $0.25 | $0.025 | $0.08333 | 90% | +| google--gemini-pro-latest | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| ibm-granite--granite-4.1-8b | nanogpt | — | $0.05 | $0.05 | — | 0% | +| inclusionai--ling-2.6-1t | nanogpt | — | $0.3 | $0.06 | — | 80% | +| mercury-2 | nanogpt | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5.4 | nanogpt | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | nanogpt | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | nanogpt | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | nanogpt | — | $5 | $0.5 | — | 90% | +| openai--gpt-chat-latest | nanogpt | — | $5 | $0.5 | — | 90% | +| openai--gpt-latest | nanogpt | — | $5 | $0.5 | — | 90% | +| qwen-3.6-plus | nanogpt | — | $0.325 | $0.0325 | $0.40625 | 90% | +| sarvam-105b | nanogpt | — | $0.045 | $0.028 | — | 38% | +| sarvam-30b | nanogpt | — | $0.028 | $0.017 | — | 39% | +| tencent--hy3-preview | nanogpt | — | $0.066 | $0.029 | — | 56% | +| upstage--solar-pro-3 | nanogpt | — | $0.15 | $0.015 | — | 90% | +| x-ai--grok-4.3 | nanogpt | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-latest | nanogpt | — | $1.25 | $0.2 | — | 84% | +| xiaomi--mimo-v2-omni | nanogpt | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2-pro | nanogpt | — | $1 | $0.2 | — | 80% | +| xiaomi--mimo-v2.5 | nanogpt | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | nanogpt | — | $0.6 | $0.12 | — | 80% | +| z-ai--glm-5-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | +| z-ai--glm-5v-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | +| z-ai--glm-5v-turbo--thinking | nanogpt | — | $1.2 | $0.24 | — | 80% | +| zai-org--glm-4.7-original | nanogpt | — | $0.6 | $0.11 | — | 82% | +| zai-org--glm-4.7-original--thinking | nanogpt | — | $0.6 | $0.11 | — | 82% | +| gpt-4.1 | openai | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini | openai | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano | openai | — | $0.1 | $0.025 | — | 75% | +| gpt-4o | openai | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-mini | openai | — | $0.15 | $0.075 | — | 50% | +| o1 | openai | — | $15 | $7.5 | — | 50% | +| o1-mini | openai | — | $1.5 | $0.75 | — | 50% | +| o3 | openai | — | $10 | $2.5 | — | 75% | +| o3-mini | openai | — | $1.1 | $0.55 | — | 50% | +| o4-mini | openai | — | $1.1 | $0.275 | — | 75% | +| aion-labs--aion-2.0 | openrouter | — | $0.8 | $0.2 | — | 75% | +| alibaba--tongyi-deepresearch-30b-a3b | openrouter | — | $0.09 | $0.09 | — | 0% | +| amazon--nova-premier-v1 | openrouter | — | $2.5 | $0.625 | — | 75% | +| anthropic--claude-3-haiku | openrouter | — | $0.25 | $0.03 | $0.3 | 88% | +| anthropic--claude-3.5-haiku | openrouter | — | $0.8 | $0.08 | $1 | 90% | +| anthropic--claude-haiku-4.5 | openrouter | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4 | openrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.1 | openrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.5 | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.6 | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.6-fast | openrouter | — | $30 | $3 | $37.5 | 90% | +| anthropic--claude-opus-4.7 | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.7-fast | openrouter | — | $30 | $3 | $37.5 | 90% | +| anthropic--claude-sonnet-4 | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4.5 | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4.6 | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| arcee-ai--trinity-large-thinking | openrouter | — | $0.22 | $0.06 | — | 73% | +| bytedance--ui-tars-1.5-7b | openrouter | — | $0.1 | $0.1 | — | 0% | +| deepseek--deepseek-chat-v3-0324 | openrouter | — | $0.2 | $0.135 | — | 32% | +| deepseek--deepseek-chat-v3.1 | openrouter | — | $0.21 | $0.13 | — | 38% | +| deepseek--deepseek-r1-0528 | openrouter | — | $0.5 | $0.35 | — | 30% | +| deepseek--deepseek-v3.1-terminus | openrouter | — | $0.27 | $0.13 | — | 52% | +| deepseek--deepseek-v3.2 | openrouter | — | $0.252 | $0.0252 | — | 90% | +| deepseek--deepseek-v3.2-speciale | openrouter | — | $0.287 | $0.058 | — | 80% | +| deepseek--deepseek-v4-flash | openrouter | — | $0.112 | $0.022 | — | 80% | +| deepseek--deepseek-v4-pro | openrouter | — | $0.435 | $0.003625 | — | 99% | +| google--gemini-2.0-flash-001 | openrouter | — | $0.1 | $0.025 | $0.083333 | 75% | +| google--gemini-2.5-flash | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | +| google--gemini-2.5-flash-image | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | +| google--gemini-2.5-flash-lite | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | +| google--gemini-2.5-flash-lite-preview-09-2025 | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | +| google--gemini-2.5-pro | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | +| google--gemini-2.5-pro-preview | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | +| google--gemini-2.5-pro-preview-05-06 | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | +| google--gemini-3-flash-preview | openrouter | — | $0.5 | $0.05 | $0.083333 | 90% | +| google--gemini-3-pro-image-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-flash-lite | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | +| google--gemini-3.1-flash-lite-preview | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | +| google--gemini-3.1-pro-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-pro-preview-customtools | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.5-flash | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | +| ibm-granite--granite-4.1-8b | openrouter | — | $0.05 | $0.05 | — | 0% | +| inception--mercury-2 | openrouter | — | $0.25 | $0.025 | — | 90% | +| inclusionai--ling-2.6-1t | openrouter | — | $0.3 | $0.06 | — | 80% | +| inclusionai--ling-2.6-flash | openrouter | — | $0.01 | $0.002 | — | 80% | +| inclusionai--ring-2.6-1t | openrouter | — | $0.075 | $0.015 | — | 80% | +| kwaipilot--kat-coder-pro-v2 | openrouter | — | $0.3 | $0.06 | — | 80% | +| microsoft--phi-4-mini-instruct | openrouter | — | $0.08 | $0.08 | — | 0% | +| minimax--minimax-m2 | openrouter | — | $0.255 | $0.03 | — | 88% | +| minimax--minimax-m2-her | openrouter | — | $0.3 | $0.03 | — | 90% | +| minimax--minimax-m2.1 | openrouter | — | $0.29 | $0.03 | — | 90% | +| mistralai--codestral-2508 | openrouter | — | $0.3 | $0.03 | — | 90% | +| mistralai--devstral-2512 | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--devstral-medium | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--devstral-small | openrouter | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-14b-2512 | openrouter | — | $0.2 | $0.02 | — | 90% | +| mistralai--ministral-3b-2512 | openrouter | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-8b-2512 | openrouter | — | $0.15 | $0.015 | — | 90% | +| mistralai--mistral-large | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2407 | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2512 | openrouter | — | $0.5 | $0.05 | — | 90% | +| mistralai--mistral-medium-3 | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-medium-3.1 | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-saba | openrouter | — | $0.2 | $0.02 | — | 90% | +| mistralai--mistral-small-2603 | openrouter | — | $0.15 | $0.015 | — | 90% | +| mistralai--mixtral-8x22b-instruct | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--pixtral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--voxtral-small-24b-2507 | openrouter | — | $0.1 | $0.01 | — | 90% | +| moonshotai--kimi-k2.5 | openrouter | — | $0.4 | $0.09 | — | 78% | +| moonshotai--kimi-k2.6 | openrouter | — | $0.73 | $0.25 | — | 66% | +| openai--gpt-4.1 | openrouter | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-mini | openrouter | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-nano | openrouter | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4o-2024-08-06 | openrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-11-20 | openrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-mini | openrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-mini-2024-07-18 | openrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-5 | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-chat | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-codex | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-image | openrouter | — | $10 | $1.25 | — | 88% | +| openai--gpt-5-image-mini | openrouter | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5-mini | openrouter | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-nano | openrouter | — | $0.05 | $0.01 | — | 80% | +| openai--gpt-5.1 | openrouter | — | $1.25 | $0.13 | — | 90% | +| openai--gpt-5.1-chat | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-max | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-mini | openrouter | — | $0.25 | $0.03 | — | 88% | +| openai--gpt-5.2 | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-chat | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-codex | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-chat | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-codex | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.4 | openrouter | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-image-2 | openrouter | — | $8 | $2 | — | 75% | +| openai--gpt-5.4-mini | openrouter | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | openrouter | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | openrouter | — | $5 | $0.5 | — | 90% | +| openai--gpt-chat-latest | openrouter | — | $5 | $0.5 | — | 90% | +| openai--gpt-oss-safeguard-20b | openrouter | — | $0.075 | $0.037 | — | 51% | +| openai--o1 | openrouter | — | $15 | $7.5 | — | 50% | +| openai--o3 | openrouter | — | $2 | $0.5 | — | 75% | +| openai--o3-deep-research | openrouter | — | $10 | $2.5 | — | 75% | +| openai--o3-mini | openrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o3-mini-high | openrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o4-mini | openrouter | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-deep-research | openrouter | — | $2 | $0.5 | — | 75% | +| openai--o4-mini-high | openrouter | — | $1.1 | $0.275 | — | 75% | +| qwen--qwen-plus | openrouter | — | $0.26 | $0.052 | $0.325 | 80% | +| qwen--qwen-plus-2025-07-28 | openrouter | — | $0.26 | — | $0.325 | — | +| qwen--qwen-plus-2025-07-28--thinking | openrouter | — | $0.26 | — | $0.325 | — | +| qwen--qwen3-30b-a3b-thinking-2507 | openrouter | — | $0.08 | $0.08 | — | 0% | +| qwen--qwen3-8b | openrouter | — | $0.05 | $0.05 | — | 0% | +| qwen--qwen3-coder-flash | openrouter | — | $0.195 | $0.039 | $0.24375 | 80% | +| qwen--qwen3-coder-next | openrouter | — | $0.11 | $0.07 | — | 36% | +| qwen--qwen3-coder-plus | openrouter | — | $0.65 | $0.13 | $0.8125 | 80% | +| qwen--qwen3-max | openrouter | — | $0.78 | $0.156 | $0.975 | 80% | +| qwen--qwen3-vl-235b-a22b-instruct | openrouter | — | $0.2 | $0.11 | — | 45% | +| qwen--qwen3.5-397b-a17b | openrouter | — | $0.39 | $0.195 | — | 50% | +| qwen--qwen3.5-flash-02-23 | openrouter | — | $0.065 | — | $0.08125 | — | +| qwen--qwen3.5-plus-02-15 | openrouter | — | $0.26 | — | $0.325 | — | +| qwen--qwen3.6-35b-a3b | openrouter | — | $0.15 | $0.05 | — | 67% | +| qwen--qwen3.6-flash | openrouter | — | $0.1875 | — | $0.234375 | — | +| qwen--qwen3.6-max-preview | openrouter | — | $1.04 | — | $1.3 | — | +| qwen--qwen3.6-plus | openrouter | — | $0.325 | — | $0.40625 | — | +| tencent--hy3-preview | openrouter | — | $0.066 | $0.029 | — | 56% | +| thedrummer--cydonia-24b-v4.1 | openrouter | — | $0.3 | $0.15 | — | 50% | +| thedrummer--skyfall-36b-v2 | openrouter | — | $0.55 | $0.25 | — | 55% | +| upstage--solar-pro-3 | openrouter | — | $0.15 | $0.015 | — | 90% | +| x-ai--grok-4.20 | openrouter | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-4.20-multi-agent | openrouter | — | $2 | $0.2 | — | 90% | +| x-ai--grok-4.3 | openrouter | — | $1.25 | $0.2 | — | 84% | +| xiaomi--mimo-v2-flash | openrouter | — | $0.1 | $0.01 | — | 90% | +| xiaomi--mimo-v2-omni | openrouter | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2-pro | openrouter | — | $1 | $0.2 | — | 80% | +| xiaomi--mimo-v2.5 | openrouter | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | openrouter | — | $1 | $0.2 | — | 80% | +| z-ai--glm-4.5 | openrouter | — | $0.6 | $0.11 | — | 82% | +| z-ai--glm-4.5-air | openrouter | — | $0.13 | $0.025 | — | 81% | +| z-ai--glm-4.5v | openrouter | — | $0.6 | $0.11 | — | 82% | +| z-ai--glm-4.6 | openrouter | — | $0.43 | $0.08 | — | 81% | +| z-ai--glm-4.6v | openrouter | — | $0.3 | $0.05 | — | 83% | +| z-ai--glm-4.7 | openrouter | — | $0.4 | $0.08 | — | 80% | +| z-ai--glm-4.7-flash | openrouter | — | $0.06 | $0.01 | — | 83% | +| z-ai--glm-5 | openrouter | — | $0.6 | $0.12 | — | 80% | +| z-ai--glm-5-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | +| z-ai--glm-5v-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | +| ~anthropic--claude-haiku-latest | openrouter | — | $1 | $0.1 | $1.25 | 90% | +| ~anthropic--claude-opus-latest | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| ~anthropic--claude-sonnet-latest | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| ~google--gemini-flash-latest | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | +| ~google--gemini-pro-latest | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| ~moonshotai--kimi-latest | openrouter | — | $0.73 | $0.25 | — | 66% | +| ~openai--gpt-latest | openrouter | — | $5 | $0.5 | — | 90% | +| ~openai--gpt-mini-latest | openrouter | — | $0.75 | $0.075 | — | 90% | +| deepseek--deepseek-v3-0324 | ppio | — | $2 | $0.6 | — | 70% | +| deepseek--deepseek-v3.1 | ppio | — | $4 | $2 | — | 50% | +| deepseek--deepseek-v3.1-terminus | ppio | — | $4 | $2 | — | 50% | +| deepseek--deepseek-v3.2 | ppio | — | $2 | $0.2 | — | 90% | +| deepseek--deepseek-v4-flash | ppio | — | $1 | $0.2 | — | 80% | +| deepseek--deepseek-v4-pro | ppio | — | $12 | $1 | — | 92% | +| minimax--minimax-m2 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | +| minimax--minimax-m2.1 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | +| minimax--minimax-m2.5 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | +| minimax--minimax-m2.5-highspeed | ppio | — | $4.2 | $0.21 | $2.625 | 95% | +| minimax--minimax-m2.7 | ppio | — | $2.1 | $0.42 | $2.625 | 80% | +| minimax--minimax-m2.7-highspeed | ppio | — | $4.2 | $0.42 | $2.625 | 90% | +| moonshotai--kimi-k2.5 | ppio | — | $4 | $0.7 | — | 82% | +| moonshotai--kimi-k2.6 | ppio | — | $6.5 | $1.1 | — | 83% | +| xiaomimimo--mimo-v2-flash | ppio | — | $0.7 | $0.07 | — | 90% | +| zai-org--glm-4.5 | ppio | — | $4 | $0.8 | — | 80% | +| zai-org--glm-4.5-air | ppio | — | $1.2 | $0.24 | — | 80% | +| zai-org--glm-4.5v | ppio | — | $4 | $0.8 | — | 80% | +| zai-org--glm-4.6 | ppio | — | $4 | $0.8 | $0.8 | 80% | +| zai-org--glm-4.7-flash | ppio | — | $0.5 | $0.1 | — | 80% | +| gemma-3-27b | privatemode | — | $0.77 | $0.08 | — | 90% | +| gemma-4-31b | privatemode | — | $0.77 | $0.08 | — | 90% | +| gpt-oss-120b | privatemode | — | $0.43 | $0.04 | — | 91% | +| kimi-k2-5 | privatemode | — | $1.55 | $0.15 | — | 90% | +| qwen3-coder-30b-a3b | privatemode | — | $0.43 | $0.04 | — | 91% | +| alibaba--qwen-max | requesty | — | $1.6 | $1.6 | $1.6 | 0% | +| alibaba--qwen-plus | requesty | — | $0.4 | $0.4 | $0.4 | 0% | +| alibaba--qwen-turbo | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| alibaba--qwen3-30b-a3b-instruct-2507 | requesty | — | $0.2 | $0.2 | $0.8 | 0% | +| alibaba--qwen3-coder-flash | requesty | — | $0.3 | $0.08 | $0.3 | 73% | +| alibaba--qwen3.5 | requesty | — | $0.6 | $0.6 | $3.6 | 0% | +| anthropic--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| azure--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| azure--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| azure--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| azure--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| azure--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | +| azure--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | +| azure--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| azure--gpt-5.2 | requesty | — | $1.75 | $0.175 | $14 | 90% | +| azure--openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| azure--openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| azure--openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| bedrock--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | +| bedrock--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | +| bedrock--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| bedrock--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| bedrock--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| bedrock--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| bedrock--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| bedrock--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| deepinfra--deepseek-ai--deepseek-r1 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | +| deepinfra--deepseek-ai--deepseek-r1-distill-llama-70b | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--deepseek-ai--deepseek-v3 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | +| deepinfra--deepseek-ai--deepseek-v3.1 | requesty | — | $0.3 | $0.3 | $1 | 0% | +| deepinfra--kimi k2.5 | requesty | — | $0.45 | $0.07 | — | 84% | +| deepinfra--meta-llama--llama-3.2-90b-vision-instruct | requesty | — | $0.35 | $0.35 | $0.35 | 0% | +| deepinfra--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.12 | $0.12 | $0.12 | 0% | +| deepinfra--meta-llama--meta-llama-3.1-405b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | +| deepinfra--meta-llama--meta-llama-3.1-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.02 | $0.02 | $0.02 | 0% | +| deepinfra--microsoft--phi-4 | requesty | — | $0.07 | $0.07 | $0.07 | 0% | +| deepinfra--qwen--qwen2.5-72b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | — | $0.07 | $0.07 | $0.07 | 0% | +| deepinfra--qwen--qwen3-235b-a22b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| deepinfra--qwen--qwen3-32b | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| deepinfra--qwen--qwen3-coder-480b-a35b-instruct | requesty | — | $0.4 | $0.4 | $1.6 | 0% | +| deepinfra--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | +| deepinfra--zai-org--glm-4.5-air | requesty | — | $0.2 | $0.2 | $1.1 | 0% | +| deepseek--deepseek-chat | requesty | — | $0.14 | $0.028 | $0.14 | 80% | +| deepseek--deepseek-reasoner | requesty | — | $0.14 | $0.028 | $0.14 | 80% | +| deepseek--deepseek-v4-flash | requesty | — | $0.14 | $0.028 | $0.14 | 80% | +| deepseek--deepseek-v4-pro | requesty | — | $1.74 | $0.145 | $1.74 | 92% | +| fireworks--deepseek-v3.2 | requesty | — | $0.56 | $0.28 | — | 50% | +| fireworks--deepseek-v4-pro | requesty | — | $1.74 | $0.15 | — | 91% | +| fireworks--glm-5 | requesty | — | $1 | $0.2 | — | 80% | +| fireworks--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | +| fireworks--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $3 | 83% | +| fireworks--kimi-k2.6 | requesty | — | $0.95 | $0.16 | — | 83% | +| fireworks--minimax-m2.5 | requesty | — | $0.3 | $0.03 | — | 90% | +| fireworks--minimax-m2.7 | requesty | — | $0.3 | $0.06 | — | 80% | +| fireworks--qwen3.6-plus | requesty | — | $0.5 | $0.1 | — | 80% | +| google--gemini-2.0-flash-001 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| google--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | +| google--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | +| google--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | +| google--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | +| google--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| google--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | +| google--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| groq--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.75 | 0% | +| groq--openai--gpt-oss-20b | requesty | — | $0.1 | $0.1 | $0.5 | 0% | +| inceptron--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | +| inceptron--kimi-k2.6 | requesty | — | $0.8 | $0.2 | — | 75% | +| inceptron--minimax-m2.5 | requesty | — | $0.28 | $0.03 | — | 89% | +| minimaxi--minimax-m2 | requesty | — | $0.3 | $0.3 | $1.2 | 0% | +| minimaxi--minimax-m2.5 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | +| minimaxi--minimax-m2.5-highspeed | requesty | — | $0.6 | $0.06 | $2.4 | 90% | +| minimaxi--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | +| minimaxi--minimax-m2.7-highspeed | requesty | — | $0.6 | $0.06 | $1.2 | 90% | +| mistral--codestral-latest | requesty | — | $0.3 | $0.3 | $0.9 | 0% | +| mistral--devstral-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | +| mistral--devstral-medium-2507 | requesty | — | $0.4 | $0.4 | $2 | 0% | +| mistral--devstral-small-2507 | requesty | — | $0.1 | $0.1 | $0.3 | 0% | +| mistral--devstral-small-latest | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| mistral--mistral-large-latest | requesty | — | $0.5 | $0.5 | $0.5 | 0% | +| mistral--mistral-medium-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | +| mistral--mistral-small-2503 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| mistral--mistral-small-2603 | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| mistral--mistral-small-latest | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| mistral--open-mistral-7b | requesty | — | $0.25 | $0.25 | $0.25 | 0% | +| mistral--pixtral-large-latest | requesty | — | $2 | $2 | $5 | 0% | +| moonshot--kimi-k2-0711-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-0905-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-thinking | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-thinking-turbo | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-turbo-preview | requesty | — | $1.2 | $0.3 | $5 | 75% | +| moonshot--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $0.6 | 83% | +| moonshot--kimi-k2.6 | requesty | — | $0.95 | $0.16 | $0.96 | 83% | +| nebius--deepseek-ai--deepseek-v3.2 | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| nebius--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.13 | $0.13 | $0.13 | 0% | +| nebius--moonshotai--kimi-k2.5 | requesty | — | $0.5 | $0.5 | $2.5 | 0% | +| nebius--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| novita--deepseek--deepseek-prover-v2-671b | requesty | — | $0.7 | $0.7 | $0.7 | 0% | +| novita--deepseek--deepseek-r1 | requesty | — | $4 | $4 | $4 | 0% | +| novita--deepseek--deepseek-r1-distill-llama-70b | requesty | — | $0.8 | $0.8 | $0.8 | 0% | +| novita--deepseek--deepseek-r1-distill-qwen-14b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| novita--deepseek--deepseek-r1-distill-qwen-32b | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| novita--deepseek--deepseek-r1-turbo | requesty | — | $0.7 | $0.7 | $2.5 | 0% | +| novita--deepseek--deepseek-v3-0324 | requesty | — | $0.4 | $0.4 | $0.4 | 0% | +| novita--deepseek--deepseek-v3-turbo | requesty | — | $0.4 | $0.4 | $0.4 | 0% | +| novita--deepseek--deepseek-v3.2 | requesty | — | $0.269 | $0.1345 | $0.269 | 50% | +| novita--deepseek--deepseek_v3 | requesty | — | $0.89 | $0.89 | $0.89 | 0% | +| novita--glm-5 | requesty | — | $1 | $0.2 | — | 80% | +| novita--gryphe--mythomax-l2-13b | requesty | — | $0.09 | $0.09 | $0.09 | 0% | +| novita--meta-llama--llama-3-70b-instruct | requesty | — | $0.51 | $0.51 | $0.51 | 0% | +| novita--meta-llama--llama-3-8b-instruct | requesty | — | $0.04 | $0.04 | $0.04 | 0% | +| novita--meta-llama--llama-3.1-8b-instruct | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| novita--meta-llama--llama-3.2-1b-instruct | requesty | — | $0.02 | $0.02 | $0.02 | 0% | +| novita--meta-llama--llama-3.2-3b-instruct | requesty | — | $0.03 | $0.03 | $0.03 | 0% | +| novita--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.39 | $0.39 | $0.39 | 0% | +| novita--meta-llama--llama-4-maverick-17b-128e-instruct-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| novita--microsoft--wizardlm-2-8x22b | requesty | — | $0.62 | $0.62 | $0.62 | 0% | +| novita--minimax--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | +| novita--mistralai--mistral-nemo | requesty | — | $0.17 | $0.17 | $0.17 | 0% | +| novita--moonshotai--kimi-k2-instruct | requesty | — | $0.57 | $0.57 | $0.57 | 0% | +| novita--nousresearch--hermes-2-pro-llama-3-8b | requesty | — | $0.14 | $0.14 | $0.14 | 0% | +| novita--qwen--qwen-2.5-72b-instruct | requesty | — | $0.38 | $0.38 | $0.38 | 0% | +| novita--qwen--qwen2.5-vl-72b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | +| novita--qwen--qwen3-235b-a22b-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| novita--qwen--qwen3.5-397b-a17b | requesty | — | $0.6 | $0.6 | $3.6 | 0% | +| novita--sao10k--l3-70b-euryale-v2.1 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | +| novita--sao10k--l3-8b-lunaris | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| novita--sao10k--l3-8b-stheno-v3.2 | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| novita--sao10k--l31-70b-euryale-v2.2 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | +| novita--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | +| novita--zai-org--glm-4.6 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | +| openai--chatgpt-4o | requesty | — | $5 | $5 | $5 | 0% | +| openai--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| openai--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| openai--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| openai--gpt-4o | requesty | — | $2.5 | $1.25 | $2.5 | 50% | +| openai--gpt-4o-2024-05-13 | requesty | — | $2.5 | $2.5 | $2.5 | 0% | +| openai--gpt-4o-2024-08-06 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | +| openai--gpt-4o-2024-11-20 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | +| openai--gpt-4o-mini | requesty | — | $0.15 | $0.075 | $0.15 | 50% | +| openai--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | +| openai--gpt-5-mini:flex | requesty | — | $0.125 | $0.0125 | $0.125 | 90% | +| openai--gpt-5-mini:priority | requesty | — | $0.45 | $0.045 | $3.6 | 90% | +| openai--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | +| openai--gpt-5-nano:flex | requesty | — | $0.025 | $0.0025 | $0.025 | 90% | +| openai--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5.1-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | +| openai--gpt-5.2-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai--gpt-5.4 | requesty | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | +| openai--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | +| openai--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | +| openai--gpt-5:flex | requesty | — | $0.625 | $0.0625 | $0.625 | 90% | +| openai--gpt-5:priority | requesty | — | $2.5 | $0.25 | $20 | 90% | +| openai--o1 | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o1-pro | requesty | — | $150 | $150 | $150 | 0% | +| openai--o1:high | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o1:low | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o1:medium | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o3 | requesty | — | $2 | $0.5 | $2 | 75% | +| openai--o3-deep-research | requesty | — | $10 | $2.5 | $10 | 75% | +| openai--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3-mini:high | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3-mini:low | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3-mini:medium | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3:flex | requesty | — | $1 | $0.25 | $1 | 75% | +| openai--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai--o4-mini-deep-research | requesty | — | $2 | $0.5 | $2 | 75% | +| openai--o4-mini:flex | requesty | — | $0.55 | $0.138 | $0.55 | 75% | +| openai--o4-mini:high | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai--o4-mini:low | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai--o4-mini:medium | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| openai-responses--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai-responses--gpt-5-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | +| openai-responses--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | +| openai-responses--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | +| openai-responses--gpt-5-pro | requesty | — | $15 | $15 | $120 | 0% | +| openai-responses--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai-responses--gpt-5.1-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | +| openai-responses--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | +| openai-responses--gpt-5.2-codex | requesty | — | $1.75 | $0.175 | $1.75 | 90% | +| openai-responses--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai-responses--gpt-5.3-codex | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai-responses--gpt-5.4 | requesty | — | $2.5 | $0.25 | $2.5 | 90% | +| openai-responses--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | +| openai-responses--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | +| openai-responses--gpt-5.4-pro | requesty | — | $30 | $30 | $180 | 0% | +| openai-responses--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | +| openai-responses--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai-responses--o3-pro | requesty | — | $20 | $20 | $20 | 0% | +| openai-responses--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| parasail--parasail-gemma3-27b-it | requesty | — | $0.3 | $0.3 | $0.5 | 0% | +| parasail--parasail-kimi-k2-instruct | requesty | — | $0.99 | $0.99 | $2.99 | 0% | +| parasail--parasail-qwen25-vl-72b-instruct | requesty | — | $0.7 | $0.7 | $0.7 | 0% | +| parasail--parasail-qwen3-235b-a22b-instruct-2507 | requesty | — | $0.15 | $0.15 | $0.85 | 0% | +| perplexity--sonar | requesty | — | $1 | $1 | $1 | 0% | +| perplexity--sonar-pro | requesty | — | $3 | $3 | $3 | 0% | +| perplexity--sonar-reasoning-pro | requesty | — | $2 | $2 | $2 | 0% | +| together--deepseek-ai--deepseek-r1 | requesty | — | $3 | $3 | $7 | 0% | +| together--deepseek-ai--deepseek-v3 | requesty | — | $1.25 | $1.25 | $1.25 | 0% | +| together--meta-llama--llama-3.2-3b-instruct-turbo | requesty | — | $0.06 | $0.06 | $0.06 | 0% | +| together--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | +| together--meta-llama--llamaguard-2-8b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| together--meta-llama--meta-llama-3-8b-instruct-lite | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| together--meta-llama--meta-llama-3.1-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | +| together--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.18 | $0.18 | $0.18 | 0% | +| together--qwen--qwen2.5-72b-instruct-turbo | requesty | — | $1.2 | $1.2 | $1.2 | 0% | +| together--qwen--qwen2.5-7b-instruct-turbo | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| vertex--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | +| vertex--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| vertex--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| vertex--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| vertex--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| vertex--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| vertex--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--deepseek-v3.2 | requesty | — | $0.56 | $0.056 | $1.68 | 90% | +| vertex--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | +| vertex--gemini-2.5-flash-image | requesty | — | $0.3 | $0.3 | $2.5 | 0% | +| vertex--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | +| vertex--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | +| vertex--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | +| vertex--gemini-3-flash-preview:flex | requesty | — | $0.25 | $0.025 | $0.5 | 90% | +| vertex--gemini-3-pro-image-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| vertex--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| vertex--gemini-3.1-flash-lite | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | +| vertex--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | +| vertex--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| vertex--gemini-3.1-pro-preview:flex | requesty | — | $1 | $0.1 | $2.75 | 90% | +| vertex--gemini-3.5-flash | requesty | — | $1.5 | $0.15 | $1.583 | 90% | +| vertex--kimi-k2 | requesty | — | $0.6 | $0.06 | $2.5 | 90% | +| xai--grok-2-1212 | requesty | — | $2 | $2 | $2 | 0% | +| xai--grok-3 | requesty | — | $5 | $5 | $5 | 0% | +| xai--grok-3-mini | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| xai--grok-3-mini:high | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| xai--grok-3-mini:low | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| xai--grok-4 | requesty | — | $3 | $0.75 | $3 | 75% | +| xai--grok-4-1-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4-1-fast-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4-fast | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4.2-beta | requesty | — | $2 | $0.2 | $2 | 90% | +| xai--grok-4.3 | requesty | — | $1.25 | $0.2 | $1.25 | 84% | +| xai--grok-code-fast-1 | requesty | — | $0.2 | $0.02 | $0.2 | 90% | +| zai--glm-4.5 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | +| zai--glm-4.6 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | +| zai--glm-4.7 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | +| zai--glm-5 | requesty | — | $1 | $0.2 | — | 80% | +| zai--glm-5.1 | requesty | — | $1.4 | $0.26 | $4.4 | 81% | +| deepseek-v3.2 | siliconflow | — | $0.27 | $0.135 | — | 50% | +| deepseek-v4-flash | siliconflow | — | $0.14 | $0.028 | — | 80% | +| deepseek-v4-pro | siliconflow | — | $1.74 | $0.145 | — | 92% | +| glm-4.7 | siliconflow | — | $0.42 | $0.11 | — | 74% | +| glm-5 | siliconflow | — | $0.95 | $0.2 | — | 79% | +| glm-5.1 | siliconflow | — | $1.4 | $0.26 | — | 81% | +| hy3-preview | siliconflow | — | $0.066 | $0.029 | — | 56% | +| kimi-k2.5 | siliconflow | — | $0.45 | $0.07 | — | 84% | +| kimi-k2.6 | siliconflow | — | $0.9 | $0.2 | — | 78% | +| minimax-m2.5 | siliconflow | — | $0.3 | $0.03 | — | 90% | +| step-1-32k | stepfun | — | $15 | $3 | — | 80% | +| step-1-8k | stepfun | — | $5 | $1 | — | 80% | +| step-1o-audio | stepfun | — | $25 | $5 | — | 80% | +| step-1o-turbo-vision | stepfun | — | $2.5 | $0.5 | — | 80% | +| step-1o-vision-32k | stepfun | — | $15 | $3 | — | 80% | +| step-1v-32k | stepfun | — | $15 | $3 | — | 80% | +| step-1v-8k | stepfun | — | $5 | $1 | — | 80% | +| step-2-16k | stepfun | — | $38 | $7.6 | — | 80% | +| step-2-16k-exp | stepfun | — | $38 | $7.6 | — | 80% | +| step-2-mini | stepfun | — | $1 | $0.2 | — | 80% | +| step-3 | stepfun | — | $1.5 | $0.3 | — | 80% | +| step-3.5-flash | stepfun | — | $0.7 | $0.14 | — | 80% | +| step-3.5-flash-2603 | stepfun | — | $0.7 | $0.14 | — | 80% | +| step-audio-2 | stepfun | — | $10 | $2 | — | 80% | +| step-r1-v-mini | stepfun | — | $2.5 | $0.5 | — | 80% | +| stepaudio-2.5-chat | stepfun | — | $10 | $2 | — | 80% | +| stepaudio-2.5-realtime | stepfun | — | $10 | $2 | — | 80% | +| deepseek-v4-flash | tencent-tokenhub | — | $1 | $0.2 | — | 80% | +| deepseek-v4-pro | tencent-tokenhub | — | $12 | $1 | — | 92% | +| glm-5 | tencent-tokenhub | — | $4 | $1 | — | 75% | +| glm-5-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | +| glm-5.1 | tencent-tokenhub | — | $6 | $1.3 | — | 78% | +| glm-5v-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | +| hy3-preview | tencent-tokenhub | — | $1.2 | $0.4 | — | 67% | +| kimi-k2.5 | tencent-tokenhub | — | $4 | $0.7 | — | 82% | +| kimi-k2.6 | tencent-tokenhub | — | $6.5 | $1.1 | — | 83% | +| minimax-m2.5 | tencent-tokenhub | — | $2.1 | $0.21 | — | 90% | +| minimax-m2.7 | tencent-tokenhub | — | $2.1 | $0.42 | — | 80% | +| MiniMaxAI--MiniMax-M2.5 | togetherai | — | $0.3 | $0.06 | — | 80% | +| MiniMaxAI--MiniMax-M2.7 | togetherai | — | $0.3 | $0.06 | — | 80% | +| deepseek-ai--DeepSeek-V4-Pro | togetherai | — | $2.1 | $0.2 | — | 90% | +| moonshotai--Kimi-K2.6 | togetherai | — | $1.2 | $0.2 | — | 83% | +| solar-pro2 | upstage | — | $0.15 | $0.015 | — | 90% | +| solar-pro3 | upstage | — | $0.15 | $0.015 | — | 90% | +| aion-labs-aion-2-0 | venice | — | $1 | $0.25 | — | 75% | +| arcee-trinity-large-thinking | venice | — | $0.3125 | $0.075 | — | 76% | +| claude-opus-4-5 | venice | — | $6 | $0.6 | — | 90% | +| claude-opus-4-6 | venice | — | $6 | $0.6 | — | 90% | +| claude-opus-4-6-fast | venice | — | $36 | $3.6 | — | 90% | +| claude-opus-4-7 | venice | — | $6 | $0.6 | — | 90% | +| claude-opus-4-7-fast | venice | — | $36 | $3.6 | — | 90% | +| claude-sonnet-4-5 | venice | — | $3.75 | $0.375 | — | 90% | +| claude-sonnet-4-6 | venice | — | $3.6 | $0.36 | — | 90% | +| deepseek-v3.2 | venice | — | $0.33 | $0.16 | — | 52% | +| deepseek-v4-flash | venice | — | $0.17 | $0.028 | — | 84% | +| deepseek-v4-pro | venice | — | $1.73 | $0.33 | — | 81% | +| gemini-3-1-pro-preview | venice | — | $2.5 | $0.5 | — | 80% | +| gemini-3-flash-preview | venice | — | $0.7 | $0.07 | — | 90% | +| grok-4-20 | venice | — | $1.42 | $0.23 | — | 84% | +| grok-4-20-multi-agent | venice | — | $1.42 | $0.23 | — | 84% | +| grok-4-3 | venice | — | $1.42 | $0.23 | — | 84% | +| kimi-k2-5 | venice | — | $0.56 | $0.22 | — | 61% | +| kimi-k2-6 | venice | — | $0.85 | $0.22 | — | 74% | +| mercury-2 | venice | — | $0.3125 | $0.03125 | — | 90% | +| minimax-m25 | venice | — | $0.34 | $0.04 | — | 88% | +| minimax-m27 | venice | — | $0.375 | $0.075 | — | 80% | +| openai-gpt-4o-mini-2024-07-18 | venice | — | $0.1875 | $0.09375 | — | 50% | +| openai-gpt-52 | venice | — | $2.19 | $0.219 | — | 90% | +| openai-gpt-52-codex | venice | — | $2.19 | $0.219 | — | 90% | +| openai-gpt-53-codex | venice | — | $2.19 | $0.219 | — | 90% | +| openai-gpt-54 | venice | — | $3.13 | $0.313 | — | 90% | +| openai-gpt-54-mini | venice | — | $0.9375 | $0.09375 | — | 90% | +| openai-gpt-55 | venice | — | $6.25 | $0.625 | — | 90% | +| qwen-3-6-plus | venice | — | $0.625 | $0.0625 | — | 90% | +| qwen3-5-35b-a3b | venice | — | $0.3125 | $0.15625 | — | 50% | +| qwen3-coder-480b-a35b-instruct-turbo | venice | — | $0.35 | $0.04 | — | 89% | +| z-ai-glm-5-turbo | venice | — | $1.2 | $0.24 | — | 80% | +| z-ai-glm-5v-turbo | venice | — | $1.5 | $0.3 | — | 80% | +| zai-org-glm-4.6 | venice | — | $0.85 | $0.3 | — | 65% | +| zai-org-glm-4.7 | venice | — | $0.55 | $0.11 | — | 80% | +| zai-org-glm-5 | venice | — | $1 | $0.2 | — | 80% | +| zai-org-glm-5-1 | venice | — | $1.75 | $0.325 | — | 81% | +| GLM-5.1 | wafer | — | $1.5 | $0.15 | — | 90% | +| Qwen3.5-397B-A17B | wafer | — | $0.6 | $0.06 | — | 90% | + +> 💡 Use the [interactive catalog](https://i-need-token.github.io/ai-models/) to search and filter models with more criteria. + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/cached-pricing.md b/docs/zh/cached-pricing.md new file mode 100644 index 00000000..d4556d7e --- /dev/null +++ b/docs/zh/cached-pricing.md @@ -0,0 +1,1411 @@ +# 缓存定价 + +[English](../cached-pricing.md) + +支持提示缓存的 AI 模型，展示标准定价与缓存定价对比。缓存输入可比标准输入 token **便宜 50-90%**。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么缓存定价很重要 + +提示缓存允许你存储重复的提示前缀（系统提示、少样本示例、工具定义），并在多个请求中复用。这显著降低了： + +- **成本**：输入 token 节省 50-90% +- **延迟**：缓存内容的首 token 时间更快 +- **吞吐量**：更高效地利用速率限制 + +## 统计 + +| 指标 | 数量 | +| ------------------ | ---- | +| 支持缓存定价的模型 | 1374 | +| 提供商 | 39 | + +## 提供商 + +`aihubmix`, `aion`, `amazon-bedrock`, `auriko`, `baidu`, `baseten`, `chutes`, `clarifai`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `deepseek`, `digitalocean`, `fastrouter`, `friendli`, `google`, `google-vertex`, `groq`, `hpc-ai`, `inception`, `jiekou`, `llmgateway`, `martian`, `minimax`, `moonshotai`, `nanogpt`, `openai`, `openrouter`, `ppio`, `privatemode`, `requesty`, `siliconflow`, `stepfun`, `tencent-tokenhub`, `togetherai`, `upstage`, `venice`, `wafer` + +## 模型定价 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 缓存读取 $/M | 缓存写入 $/M | 节省 | +| ---------------------------------------------------------- | ---------------- | ------ | ------------------- | --------------------- | ------------ | ----- | +| aistudio_gemini-2.0-flash | aihubmix | — | $0.05 | $0.125 | — | -150% | +| aistudio_gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| anthropic-opus-4-6 | aihubmix | — | $2.5 | $0.25 | $3.125 | 90% | +| claude-haiku-4-5 | aihubmix | — | $0.55 | $0.055 | $0.6875 | 90% | +| claude-sonnet-4-0 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5-think | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| codex-mini-latest | aihubmix | — | $0.75 | $0.1875 | — | 75% | +| deepseek-v3.2 | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| deepseek-v3.2-exp | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-exp-think | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-think | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| doubao-1.5-lite-32k | aihubmix | — | $0.025 | $0.005 | — | 80% | +| doubao-1.5-pro-32k | aihubmix | — | $0.067 | $0.0134 | — | 80% | +| doubao-lite-32k | aihubmix | — | $0.03 | $0.006 | — | 80% | +| doubao-pro-32k | aihubmix | — | $0.07 | $0.014 | — | 80% | +| doubao-seed-1-6 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-flash | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-flash-250615 | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-lite | aihubmix | — | $0.041 | $0.0082 | — | 80% | +| doubao-seed-1-6-thinking | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-thinking-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-vision-250815 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| doubao-seed-1-8 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| gemini-2.0-flash | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.0-flash-001 | aihubmix | — | $0.05 | $0.125 | — | -150% | +| gemini-2.0-flash-search | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.5-flash | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-lite | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-pro | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| glm-4.5-airx | aihubmix | — | $0.55 | $0.11 | — | 80% | +| glm-4.5-x | aihubmix | — | $1.1 | $0.22 | — | 80% | +| glm-4.6 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| glm-4.6v | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| glm-4.7 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| gpt-4.1 | aihubmix | — | $1 | $0.25 | — | 75% | +| gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| gpt-4.1-nano | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gpt-4o | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06-global | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-11-20 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-mini | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-2024-07-18 | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-global | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-search-preview | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-search-preview | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-5 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5-nano | aihubmix | — | $0.025 | $0.0025 | — | 90% | +| gpt-5.1 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-max | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5.2 | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-chat-latest | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-codex | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-high | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-low | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-pro | aihubmix | — | $10.5 | $1.05 | — | 90% | +| grok-4 | aihubmix | — | $1.65 | $0.4125 | — | 75% | +| grok-4-1-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-1-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4.20-beta-0309-non-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-beta-0309-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-beta-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-code-fast-1 | aihubmix | — | $0.1 | $0.025 | — | 75% | +| kimi-k2-thinking | aihubmix | — | $0.274 | $0.0685 | — | 75% | +| kimi-k2-turbo-preview | aihubmix | — | $0.6 | $0.15 | — | 75% | +| kimi-k2.5 | aihubmix | — | $0.3 | $0.0525 | — | 82% | +| mimo-v2-flash | aihubmix | — | $0.0959 | $0.01918 | — | 80% | +| mimo-v2-omni | aihubmix | — | $0.22 | $0.044 | — | 80% | +| mimo-v2-pro | aihubmix | — | $0.55 | $0.11 | — | 80% | +| nvidia-nemotron-3-super-120b-a12b | aihubmix | — | $0.055 | $0.01375 | — | 75% | +| o1 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-2024-12-17 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-global | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-mini | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-mini-2024-09-12 | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-preview | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-preview-2024-09-12 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o3 | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-deep-research | aihubmix | — | $5 | $1.25 | — | 75% | +| o3-global | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-mini | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o3-mini-global | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o4-mini | aihubmix | — | $0.55 | $0.1375 | — | 75% | +| qwen-plus | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-04-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-07-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-latest | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-turbo | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen-turbo-latest | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen3-coder-plus | aihubmix | — | $0.27 | $0.054 | — | 80% | +| qwen3-max | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-2026-01-23 | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-preview | aihubmix | — | $0.423 | $0.0846 | — | 80% | +| qwen3-vl-flash | aihubmix | — | $0.0103 | $0.00206 | — | 80% | +| qwen3-vl-plus | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| zai-glm-5-turbo | aihubmix | — | $0.6 | $0.12 | — | 80% | +| aion-2.0 | aion | — | $0.7999999999999999 | $0.19999999999999998 | — | 75% | +| aion-2.5 | aion | — | $1 | $0.35 | — | 65% | +| amazon-nova-2-lite | amazon-bedrock | — | $0.33 | $0.0825 | — | 75% | +| amazon-nova-lite | amazon-bedrock | — | $0.06 | $0.015 | — | 75% | +| amazon-nova-micro | amazon-bedrock | — | $0.035 | $0.00875 | — | 75% | +| amazon-nova-premier | amazon-bedrock | — | $2.5 | $0.625 | — | 75% | +| amazon-nova-pro | amazon-bedrock | — | $0.8 | $0.2 | — | 75% | +| claude-haiku-4-5-20251001 | auriko | — | $1 | $0.1 | $1.25 | 90% | +| claude-opus-4-1-20250805 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-20250514 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-5-20251101 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-6 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-7 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-sonnet-4-20250514 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5-20250929 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-6 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| deepseek-r1-0528 | auriko | — | $0.5 | $0.35 | — | 30% | +| deepseek-v3-0324 | auriko | — | $0.2 | $0.135 | — | 32% | +| deepseek-v3.1 | auriko | — | $0.21 | $0.13 | — | 38% | +| deepseek-v3.1-terminus | auriko | — | $0.27 | $0.13 | — | 52% | +| deepseek-v3.2 | auriko | — | $0.26 | $0.13 | — | 50% | +| deepseek-v4-flash | auriko | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | auriko | — | $0.435 | $0.003625 | — | 99% | +| gemini-2.5-flash | auriko | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-lite | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-pro | auriko | — | $1.25 | $0.125 | — | 90% | +| gemini-3-flash-preview | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-3.1-flash-lite | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-flash-lite-preview | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-pro-preview | auriko | — | $2 | $0.2 | — | 90% | +| gemini-3.1-pro-preview-customtools | auriko | — | $2 | $0.2 | — | 90% | +| gemini-flash-latest | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-flash-lite-latest | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-pro-latest | auriko | — | $2 | $0.2 | — | 90% | +| glm-4.5 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.5-air | auriko | — | $0.2 | $0.03 | — | 85% | +| glm-4.5-airx | auriko | — | $1.1 | $0.22 | — | 80% | +| glm-4.5-x | auriko | — | $2.2 | $0.45 | — | 80% | +| glm-4.5v | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6v | auriko | — | $0.3 | $0.05 | — | 83% | +| glm-4.6v-flashx | auriko | — | $0.04 | $0.004 | — | 90% | +| glm-4.7 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.7-flashx | auriko | — | $0.07 | $0.01 | — | 86% | +| glm-5 | auriko | — | $1 | $0.2 | — | 80% | +| glm-5-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| glm-5.1 | auriko | — | $1.4 | $0.26 | — | 81% | +| glm-5v-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| gpt-4.1-2025-04-14 | auriko | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini-2025-04-14 | auriko | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano-2025-04-14 | auriko | — | $0.1 | $0.025 | — | 75% | +| gpt-4o-2024-08-06 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-2024-11-20 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-mini-2024-07-18 | auriko | — | $0.15 | $0.075 | — | 50% | +| gpt-5-2025-08-07 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-mini-2025-08-07 | auriko | — | $0.25 | $0.025 | — | 90% | +| gpt-5-nano-2025-08-07 | auriko | — | $0.05 | $0.005 | — | 90% | +| gpt-5.1-2025-11-13 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.1-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.2-2025-12-11 | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.2-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.3-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.4-2026-03-05 | auriko | — | $2.5 | $0.25 | — | 90% | +| gpt-5.4-mini-2026-03-17 | auriko | — | $0.75 | $0.075 | — | 90% | +| gpt-5.4-nano-2026-03-17 | auriko | — | $0.2 | $0.02 | — | 90% | +| gpt-5.5-2026-04-23 | auriko | — | $5 | $0.5 | — | 90% | +| gpt-oss-120b | auriko | — | $0.15 | $0.01 | — | 93% | +| gpt-oss-20b | auriko | — | $0.07 | $0.04 | — | 43% | +| grok-4.20-0309-non-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.20-0309-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.3 | auriko | — | $1.25 | $0.2 | — | 84% | +| hy3-preview | auriko | — | $0.066 | $0.029 | — | 56% | +| kimi-k2-0711-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-0905-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking-turbo | auriko | — | $1.15 | $0.15 | — | 87% | +| kimi-k2-turbo-preview | auriko | — | $1.15 | $0.15 | — | 87% | +| kimi-k2.5 | auriko | — | $0.6 | $0.1 | — | 83% | +| kimi-k2.6 | auriko | — | $0.95 | $0.16 | — | 83% | +| mimo-v2.5 | auriko | — | $0.4 | $0.08 | — | 80% | +| mimo-v2.5-pro | auriko | — | $1 | $0.2 | — | 80% | +| minimax-m2 | auriko | — | $0.3 | — | $0.375 | — | +| minimax-m2-1 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | +| minimax-m2-1-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | +| minimax-m2-5 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | +| minimax-m2-5-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | +| minimax-m2-7 | auriko | — | $0.3 | — | $0.375 | — | +| minimax-m2-7-highspeed | auriko | — | $0.6 | — | $0.375 | — | +| o3-2025-04-16 | auriko | — | $2 | $0.5 | — | 75% | +| o3-mini-2025-01-31 | auriko | — | $1.1 | $0.55 | — | 50% | +| o4-mini-2025-04-16 | auriko | — | $1.1 | $0.275 | — | 75% | +| qwen-3-coder-480b-a35b-instruct | auriko | — | $0.3 | $0.1 | — | 67% | +| qwen-3-max | auriko | — | $1.2 | $0.24 | — | 80% | +| qwen-3-max-thinking | auriko | — | $1.2 | $0.24 | — | 80% | +| qwen-3-vl-235b-a22b-instruct | auriko | — | $0.2 | $0.11 | — | 45% | +| qwen-3.5-35b-a3b | auriko | — | $0.14 | $0.05 | — | 64% | +| seed-1.8 | auriko | — | $0.25 | $0.05 | — | 80% | +| seed-2.0-code | auriko | — | $0.5 | $0.1 | — | 80% | +| seed-2.0-mini | auriko | — | $0.1 | $0.02 | — | 80% | +| seed-2.0-pro | auriko | — | $0.5 | $0.1 | — | 80% | +| step-3.5-flash | auriko | — | $0.1 | $0.02 | — | 80% | +| deepseek-v3.2 | baidu | — | $0.252 | $0.0252 | — | 90% | +| deepseek-v4-flash | baidu | — | $0.126 | $0.0252 | — | 80% | +| deepseek-v4-pro | baidu | — | $1.521 | $0.126 | — | 92% | +| glm-5 | baidu | — | $0.7 | $0.14 | — | 80% | +| glm-5.1 | baidu | — | $0.98 | $0.182 | — | 81% | +| minimax-m2.5 | baidu | — | $0.27 | $0.027 | — | 90% | +| deepseek-v3-1 | baseten | — | $0.5 | $0.25 | — | 50% | +| deepseek-v4 | baseten | — | $1.74 | $0.145 | — | 92% | +| glm-4-7 | baseten | — | $0.6 | $0.12 | — | 80% | +| glm-5 | baseten | — | $0.95 | $0.2 | — | 79% | +| kimi-k2-5 | baseten | — | $0.6 | $0.12 | — | 80% | +| kimi-k2-6 | baseten | — | $1 | $0.2 | — | 80% | +| minimax-m2-5 | baseten | — | $0.3 | $0.06 | — | 80% | +| nvidia-nemotron-3-super | baseten | — | $0.3 | $0.06 | — | 80% | +| MiniMaxAI--MiniMax-M2.5-TEE | chutes | — | $0.15 | $0.075 | — | 50% | +| Qwen--Qwen3-235B-A22B-Thinking-2507 | chutes | — | $0.11 | $0.055 | — | 50% | +| Qwen--Qwen3-32B-TEE | chutes | — | $0.08 | $0.04 | — | 50% | +| Qwen--Qwen3.5-397B-A17B-TEE | chutes | — | $0.39 | $0.195 | — | 50% | +| Qwen--Qwen3.6-27B-TEE | chutes | — | $0.5 | $0.25 | — | 50% | +| deepseek-ai--DeepSeek-V3.2-TEE | chutes | — | $0.28 | $0.14 | — | 50% | +| google--gemma-4-31B-turbo-TEE | chutes | — | $0.13 | $0.065 | — | 50% | +| moonshotai--Kimi-K2.5-TEE | chutes | — | $0.44 | $0.22 | — | 50% | +| moonshotai--Kimi-K2.6-TEE | chutes | — | $0.74 | $0.37 | — | 50% | +| zai-org--GLM-5-TEE | chutes | — | $0.95 | $0.475 | — | 50% | +| zai-org--GLM-5-Turbo | chutes | — | $0.4891 | $0.24455 | — | 50% | +| zai-org--GLM-5.1-TEE | chutes | — | $1.05 | $0.525 | — | 50% | +| kimi-k2-thinking | clarifai | — | $1.5 | $0.15 | — | 90% | +| kimi-k2.5 | cloudflare | — | $0.6 | $0.1 | — | 83% | +| kimi-k2.6 | cloudflare | — | $0.95 | $0.16 | — | 83% | +| claude-4-5-sonnet | cortecs | — | $2.683 | $0.293 | — | 89% | +| claude-4-6-sonnet | cortecs | — | $2.869 | $0.287 | — | 90% | +| claude-haiku-4-5 | cortecs | — | $0.894 | $0.089 | — | 90% | +| claude-opus4-5 | cortecs | — | $4.769 | $0.477 | — | 90% | +| claude-opus4-6 | cortecs | — | $4.769 | $0.477 | — | 90% | +| claude-opus4-7 | cortecs | — | $4.769 | $0.477 | — | 90% | +| claude-sonnet-4 | cortecs | — | $2.601 | $0.26 | — | 90% | +| codestral-2508 | cortecs | — | $0.3 | $0.03 | — | 90% | +| deepseek-chat-v3.1 | cortecs | — | $0.177 | $0.044 | — | 75% | +| deepseek-r1-0528 | cortecs | — | $0.585 | $0.146 | — | 75% | +| deepseek-v3-0324 | cortecs | — | $0.266 | $0.067 | — | 75% | +| deepseek-v3.2 | cortecs | — | $0.266 | $0.067 | — | 75% | +| deepseek-v4-flash | cortecs | — | $0.133 | $0.033 | — | 75% | +| deepseek-v4-pro | cortecs | — | $1.553 | $0.388 | — | 75% | +| devstral-2512 | cortecs | — | $0.4 | $0.04 | — | 90% | +| devstral-medium-2507 | cortecs | — | $0.4 | $0.04 | — | 90% | +| devstral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-flash | cortecs | — | $0.268 | $0.026 | — | 90% | +| gemini-2.5-pro | cortecs | — | $1.342 | $0.217 | — | 84% | +| gemini-3.1-flash-lite | cortecs | — | $0.244 | $0.022 | — | 91% | +| glm-4.6 | cortecs | — | $0.355 | $0.089 | — | 75% | +| glm-4.7 | cortecs | — | $0.532 | $0.133 | — | 75% | +| glm-5 | cortecs | — | $0.887 | $0.222 | — | 75% | +| glm-5-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | +| glm-5.1 | cortecs | — | $1.242 | $0.311 | — | 75% | +| glm-5v-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | +| gpt-4.1 | cortecs | — | $1.968 | $0.49 | — | 75% | +| gpt-4.1-mini | cortecs | — | $0.39 | $0.12 | — | 69% | +| gpt-4.1-nano | cortecs | — | $0.1 | $0.05 | — | 50% | +| gpt-4o | cortecs | — | $2.387 | $1.194 | — | 50% | +| gpt-4o-mini | cortecs | — | $0.143 | $0.073 | — | 49% | +| gpt-5 | cortecs | — | $1.234 | $0.14 | — | 89% | +| gpt-5-mini | cortecs | — | $0.25 | $0.05 | — | 80% | +| gpt-5-nano | cortecs | — | $0.054 | $0.017 | — | 69% | +| gpt-5.1 | cortecs | — | $1.234 | $0.14 | — | 89% | +| gpt-5.4 | cortecs | — | $2.601 | $0.217 | — | 92% | +| gpt-oss-120b | cortecs | — | $0.035 | $0.009 | — | 74% | +| gpt-oss-20b | cortecs | — | $0.027 | $0.007 | — | 74% | +| kimi-k2.5 | cortecs | — | $0.444 | $0.111 | — | 75% | +| kimi-k2.6 | cortecs | — | $0.694 | $0.173 | — | 75% | +| llama-3.3-70b-instruct | cortecs | — | $0.089 | $0.023 | — | 74% | +| llama-4-maverick | cortecs | — | $0.124 | $0.03 | — | 76% | +| magistral-medium-2509 | cortecs | — | $2 | $0.2 | — | 90% | +| magistral-small-2509 | cortecs | — | $0.5 | $0.05 | — | 90% | +| minimax-m2 | cortecs | — | $0.222 | $0.055 | — | 75% | +| minimax-m2.5 | cortecs | — | $0.266 | $0.027 | — | 90% | +| minimax-m2.7 | cortecs | — | $0.444 | $0.111 | — | 75% | +| ministral-14b-2512 | cortecs | — | $0.2 | $0.02 | — | 90% | +| ministral-3b-2512 | cortecs | — | $0.1 | $0.01 | — | 90% | +| ministral-8b-2512 | cortecs | — | $0.15 | $0.015 | — | 90% | +| mistral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | +| mistral-large-2512 | cortecs | — | $0.5 | $0.05 | — | 90% | +| mistral-medium-2508 | cortecs | — | $0.4 | $0.04 | — | 90% | +| mistral-medium-3.5 | cortecs | — | $1.25 | $0.125 | — | 90% | +| mistral-nemo-instruct-2407 | cortecs | — | $0.13 | $0.013 | — | 90% | +| mistral-small-2506 | cortecs | — | $0.1 | $0.01 | — | 90% | +| mistral-small-2603 | cortecs | — | $0.128 | $0.013 | — | 90% | +| nemotron-3-super-120b-a12b | cortecs | — | $0.266 | $0.067 | — | 75% | +| pixtral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | +| qwen-2.5-72b-instruct | cortecs | — | $0.062 | $0.016 | — | 74% | +| qwen3-235b-a22b-instruct-2507 | cortecs | — | $0.062 | $0.016 | — | 74% | +| qwen3-coder-30b-a3b-instruct | cortecs | — | $0.053 | $0.013 | — | 75% | +| qwen3-vl-235b-a22b | cortecs | — | $0.186 | $0.047 | — | 75% | +| qwen3.5-122b-a10b | cortecs | — | $0.444 | $0.111 | — | 75% | +| voxtral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | +| databricks-claude-haiku-4-5 | databricks | — | $1 | $0.1 | $1.25 | 90% | +| databricks-claude-opus-4-1 | databricks | — | $15 | $1.5 | $18.75 | 90% | +| databricks-claude-opus-4-5 | databricks | — | $5 | $0.5 | $6.25 | 90% | +| databricks-claude-sonnet-4 | databricks | — | $3 | $0.3 | $3.75 | 90% | +| databricks-claude-sonnet-4-5 | databricks | — | $3 | $0.3 | $3.75 | 90% | +| databricks-gemini-3-1-flash-lite | databricks | — | $0.25 | $0.02 | $0.25 | 92% | +| databricks-gemini-3-1-pro | databricks | — | $2.5 | $0.25 | $2.5 | 90% | +| databricks-gemini-3-flash | databricks | — | $0.63 | $0.06 | $0.63 | 90% | +| databricks-gpt-5 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | +| databricks-gpt-5-1 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | +| databricks-gpt-5-1-codex-max | databricks | — | $1.25 | $0.13 | $1.25 | 90% | +| databricks-gpt-5-1-codex-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | +| databricks-gpt-5-2 | databricks | — | $1.75 | $0.18 | $1.75 | 90% | +| databricks-gpt-5-2-codex | databricks | — | $1.75 | $0.18 | $1.75 | 90% | +| databricks-gpt-5-4 | databricks | — | $2.5 | $0.25 | $2.5 | 90% | +| databricks-gpt-5-4-mini | databricks | — | $0.75 | $0.07 | $0.75 | 91% | +| databricks-gpt-5-4-nano | databricks | — | $0.2 | $0.02 | $0.2 | 90% | +| databricks-gpt-5-5 | databricks | — | $5 | $0.5 | $5 | 90% | +| databricks-gpt-5-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | +| databricks-gpt-5-nano | databricks | — | $0.05 | Free | $0.05 | — | +| deepseek-r1-0528 | deepinfra | — | $0.5 | $0.35 | — | 30% | +| deepseek-v3-0324 | deepinfra | — | $0.2 | $0.135 | — | 32% | +| deepseek-v3.1 | deepinfra | — | $0.21 | $0.13 | — | 38% | +| deepseek-v3.1-terminus | deepinfra | — | $0.27 | $0.13 | — | 52% | +| deepseek-v3.2 | deepinfra | — | $0.26 | $0.13 | — | 50% | +| deepseek-v4-flash | deepinfra | — | $0.14 | $0.028 | — | 80% | +| deepseek-v4-pro | deepinfra | — | $1.74 | $0.145 | — | 92% | +| glm-4.6 | deepinfra | — | $0.43 | $0.08 | — | 81% | +| glm-4.7 | deepinfra | — | $0.4 | $0.08 | — | 80% | +| glm-4.7-flash | deepinfra | — | $0.06 | $0.01 | — | 83% | +| glm-5 | deepinfra | — | $0.6 | $0.12 | — | 80% | +| glm-5.1 | deepinfra | — | $1.05 | $0.205 | — | 80% | +| kimi-k2.5 | deepinfra | — | $0.45 | $0.07 | — | 84% | +| kimi-k2.6 | deepinfra | — | $0.75 | $0.15 | — | 80% | +| mimo-v2.5 | deepinfra | — | $0.4 | $0.08 | — | 80% | +| mimo-v2.5-pro | deepinfra | — | $1 | $0.2 | — | 80% | +| minimax-m2.5 | deepinfra | — | $0.15 | $0.03 | — | 80% | +| qwen3-235b-a22b-thinking-2507 | deepinfra | — | $0.23 | $0.2 | — | 13% | +| qwen3-coder-480b-a35b-instruct-turbo | deepinfra | — | $0.3 | $0.1 | — | 67% | +| qwen3-max | deepinfra | — | $1.2 | $0.24 | — | 80% | +| qwen3-max-thinking | deepinfra | — | $1.2 | $0.24 | — | 80% | +| qwen3-vl-235b-a22b-instruct | deepinfra | — | $0.2 | $0.11 | — | 45% | +| qwen3.5-35b-a3b | deepinfra | — | $0.14 | $0.05 | — | 64% | +| qwen3.5-397b-a17b | deepinfra | — | $0.49 | $0.3 | — | 39% | +| seed-1.8 | deepinfra | — | $0.25 | $0.05 | — | 80% | +| seed-2.0-code | deepinfra | — | $0.5 | $0.1 | — | 80% | +| seed-2.0-mini | deepinfra | — | $0.1 | $0.02 | — | 80% | +| seed-2.0-pro | deepinfra | — | $0.5 | $0.1 | — | 80% | +| step-3.5-flash | deepinfra | — | $0.1 | $0.02 | — | 80% | +| deepseek-chat | deepseek | — | $0.14 | $0.0028 | — | 98% | +| deepseek-reasoner | deepseek | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-flash | deepseek | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | deepseek | — | $0.435 | $0.003625 | — | 99% | +| arcee-trinity-large-thinking | digitalocean | — | $0.25 | $0.06 | — | 76% | +| anthropic--claude-3-5-haiku-20241022 | fastrouter | — | $0.8 | $0.08 | $1 | 90% | +| anthropic--claude-4.5-sonnet | fastrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-haiku-4.5 | fastrouter | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4-20250514 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.1 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.5 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.6 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.7 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-sonnet-4-20250514 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4.6 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | +| deepseek--deepseek-v3.2 | fastrouter | — | $0.27 | $0.216 | — | 20% | +| deepseek--deepseek-v4-pro | fastrouter | — | $2.1 | $0.2 | — | 90% | +| google--gemini-2.0-flash-001 | fastrouter | — | $0.1 | $0.025 | $0.1833 | 75% | +| google--gemini-2.5-flash | fastrouter | — | $0.3 | $0.075 | $0.3833 | 75% | +| google--gemini-2.5-pro | fastrouter | — | $1.25 | $0.31 | $1.625 | 75% | +| google--gemini-3-flash-preview | fastrouter | — | $0.5 | $0.05 | — | 90% | +| google--gemini-3.1-flash-lite-preview | fastrouter | — | $0.25 | $0.025 | $0.083333 | 90% | +| google--gemini-3.1-pro-preview | fastrouter | — | $2 | $0.31 | $1.625 | 84% | +| minimax--minimax-m2.1 | fastrouter | — | $0.27 | $0.03 | — | 89% | +| minimax--minimax-m2.5 | fastrouter | — | $0.3 | $0.03 | — | 90% | +| minimax--minimax-m2.5-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | +| minimax--minimax-m2.7 | fastrouter | — | $0.3 | $0.06 | — | 80% | +| minimax--minimax-m2.7-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | +| moonshotai--kimi-k2-thinking | fastrouter | — | $0.6 | $0.15 | — | 75% | +| moonshotai--kimi-k2.6 | fastrouter | — | $0.55 | $0.15 | — | 73% | +| openai--gpt-4.1 | fastrouter | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-mini | fastrouter | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-nano | fastrouter | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4o | fastrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-08-06 | fastrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-11-20 | fastrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-mini | fastrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-mini-2024-07-18 | fastrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-realtime-preview-2024-12-17 | fastrouter | — | $5 | $2.5 | — | 50% | +| openai--gpt-4o-realtime-preview-2025-06-03 | fastrouter | — | $5 | $2.5 | — | 50% | +| openai--gpt-5 | fastrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-mini | fastrouter | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-nano | fastrouter | — | $0.05 | $0.005 | — | 90% | +| openai--gpt-5.1 | fastrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.2 | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.4 | fastrouter | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | fastrouter | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | fastrouter | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | fastrouter | — | $5 | $0.5 | — | 90% | +| openai--gpt-realtime | fastrouter | — | $4 | $0.5 | — | 88% | +| openai--gpt-realtime-1.5 | fastrouter | — | $4 | $0.4 | — | 90% | +| openai--gpt-realtime-mini | fastrouter | — | $0.6 | $0.06 | — | 90% | +| openai--o1 | fastrouter | — | $15 | $7.5 | — | 50% | +| openai--o3 | fastrouter | — | $2 | $0.5 | — | 75% | +| openai--o3-mini | fastrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o3-mini-high | fastrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o4-mini | fastrouter | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-high | fastrouter | — | $1.1 | $0.275 | — | 75% | +| x-ai--grok-3-beta | fastrouter | — | $5 | $1.25 | — | 75% | +| x-ai--grok-3-mini-beta | fastrouter | — | $0.6 | $0.15 | — | 75% | +| x-ai--grok-4 | fastrouter | — | $3 | $0.75 | — | 75% | +| x-ai--grok-4.20-beta | fastrouter | — | $2 | $0.2 | — | 90% | +| x-ai--grok-4.3 | fastrouter | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-code-fast-1 | fastrouter | — | $0.2 | $0.02 | — | 90% | +| z-ai--glm-4.5-air | fastrouter | — | $0.2 | $0.03 | — | 85% | +| LGAI-EXAONE--K-EXAONE-236B-A23B | friendli | — | $0.2 | $0.1 | — | 50% | +| MiniMaxAI--MiniMax-M2.5 | friendli | — | $0.3 | $0.06 | — | 80% | +| deepseek-ai--DeepSeek-V3.2 | friendli | — | $0.5 | $0.25 | — | 50% | +| zai-org--GLM-5 | friendli | — | $1 | $0.5 | — | 50% | +| zai-org--GLM-5.1 | friendli | — | $1.4 | $0.26 | — | 81% | +| gemini-1.5-flash | google | — | $0.075 | $0.01875 | $0.2625 | 75% | +| gemini-1.5-flash-8b | google | — | $0.075 | $0.01875 | $0.2625 | 75% | +| gemini-1.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | +| gemini-2.0-flash | google | — | $0.1 | $0.025 | $0.35 | 75% | +| gemini-2.0-flash-lite | google | — | $0.075 | $0.01875 | $0.2625 | 75% | +| gemini-2.5-flash | google | — | $0.15 | $0.0375 | $0.525 | 75% | +| gemini-2.5-flash-lite | google | — | $0.1 | $0.025 | $0.35 | 75% | +| gemini-2.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | +| claude-haiku-4-5 | google-vertex | — | $1 | $0.1 | — | 90% | +| claude-opus-4 | google-vertex | — | $15 | $1.5 | — | 90% | +| claude-opus-4-1 | google-vertex | — | $15 | $1.5 | — | 90% | +| claude-opus-4-5 | google-vertex | — | $5 | $0.5 | — | 90% | +| claude-opus-4-6 | google-vertex | — | $5 | $0.5 | — | 90% | +| claude-opus-4-7 | google-vertex | — | $5 | $0.5 | — | 90% | +| claude-sonnet-4 | google-vertex | — | $3 | $0.3 | — | 90% | +| claude-sonnet-4-5 | google-vertex | — | $3 | $0.3 | — | 90% | +| claude-sonnet-4-6 | google-vertex | — | $3 | $0.3 | — | 90% | +| deepseek-v3-1 | google-vertex | — | $0.6 | $0.06 | — | 90% | +| deepseek-v3-2 | google-vertex | — | $0.56 | $0.056 | — | 90% | +| gemini-2-5-flash | google-vertex | — | $0.3 | $0.03 | — | 90% | +| gemini-2-5-flash-lite | google-vertex | — | $0.1 | $0.01 | — | 90% | +| gemini-2-5-pro | google-vertex | — | $1.25 | $0.13 | — | 90% | +| gemini-3-1-flash-lite | google-vertex | — | $0.25 | $0.025 | — | 90% | +| gemini-3-flash | google-vertex | — | $0.5 | $0.05 | — | 90% | +| gemini-3-pro | google-vertex | — | $2 | $0.2 | — | 90% | +| gemma-4-26b | google-vertex | — | $0.15 | $0.015 | — | 90% | +| glm-5 | google-vertex | — | $1 | $0.1 | — | 90% | +| gpt-oss-20b | google-vertex | — | $0.07 | $0.007 | — | 90% | +| grok-4-1-fast | google-vertex | — | $0.2 | $0.05 | — | 75% | +| grok-4-20 | google-vertex | — | $2 | $0.2 | — | 90% | +| kimi-k2-thinking | google-vertex | — | $0.6 | $0.06 | — | 90% | +| minimax-m2 | google-vertex | — | $0.3 | $0.03 | — | 90% | +| qwen3-coder-480b-a35b | google-vertex | — | $0.22 | $0.022 | — | 90% | +| gpt-oss-120b | groq | — | $0.15 | $0.075 | — | 50% | +| gpt-oss-20b | groq | — | $0.075 | $0.0375 | — | 50% | +| kimi-k2-instruct-0905 | groq | — | $1 | $0.5 | — | 50% | +| deepseek--deepseek-v4-flash | hpc-ai | — | $0.14 | $0.028 | — | 80% | +| deepseek--deepseek-v4-pro | hpc-ai | — | $1.74 | $0.145 | — | 92% | +| minimax--minimax-m2.5 | hpc-ai | — | $0.3 | $0.03 | — | 90% | +| moonshotai--kimi-k2.5 | hpc-ai | — | $0.6 | $0.1 | — | 83% | +| moonshotai--kimi-k2.6 | hpc-ai | — | $0.95 | $0.16 | — | 83% | +| xiaomi--mimo-v2.5 | hpc-ai | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | hpc-ai | — | $1 | $0.2 | — | 80% | +| zai--glm-5.1 | hpc-ai | — | $1.4 | $0.26 | — | 81% | +| mercury | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-coder | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-edit | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| mercury-edit-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | +| claude-opus-4-6 | jiekou | — | $2.75 | $0.475 | — | 83% | +| claude-opus-4-6-dd | jiekou | — | $2.85 | $0.275 | — | 90% | +| claude-opus-4-7 | jiekou | — | $4.75 | $0.475 | — | 90% | +| claude-sonnet-4-6 | jiekou | — | $1.65 | $0.285 | — | 83% | +| claude-sonnet-4-6-dd | jiekou | — | $4.75 | $0.165 | — | 97% | +| doubao-seed-1-8-251228 | jiekou | — | $0.11 | $0.0221 | — | 80% | +| gemini-2.5-flash | jiekou | — | $0.095 | $0.0285 | — | 70% | +| gemini-2.5-flash-lite | jiekou | — | $1.1875 | $0.0095 | — | 99% | +| gemini-2.5-pro | jiekou | — | $0.285 | $0.1187 | — | 58% | +| gemini-3-flash-preview | jiekou | — | $0.095 | $0.0475 | — | 50% | +| gemini-3.1-flash-lite-preview | jiekou | — | $1.9 | $0.0237 | — | 99% | +| gemini-3.1-pro-preview | jiekou | — | $0.475 | $0.19 | — | 60% | +| gpt-4.1 | jiekou | — | $9.5 | $0.5 | — | 95% | +| gpt-4.1-mini | jiekou | — | $0.1 | $0.1 | — | 0% | +| gpt-4o | jiekou | — | $1.9 | $1.1875 | — | 38% | +| gpt-5-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | +| gpt-5-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | +| gpt-5-nano | jiekou | — | $0.2375 | $0.0047 | — | 98% | +| gpt-5.1-chat-latest | jiekou | — | $1.1875 | $0.1187 | — | 90% | +| gpt-5.1-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | +| gpt-5.1-codex-max | jiekou | — | $0.2375 | $0.1187 | — | 50% | +| gpt-5.1-codex-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | +| gpt-5.2-chat-latest | jiekou | — | $1.1875 | $0.1662 | — | 86% | +| gpt-5.2-codex | jiekou | — | $1.6625 | $0.175 | — | 89% | +| gpt-5.3-chat-latest | jiekou | — | $1.6625 | $0.1662 | — | 90% | +| gpt-5.3-codex | jiekou | — | $1.75 | $0.1662 | — | 91% | +| gpt-5.4 | jiekou | — | $1.6625 | $0.2375 | — | 86% | +| gpt-5.4-nano | jiekou | — | $0.7125 | $0.019 | — | 97% | +| gpt-5.5 | jiekou | — | $0.19 | $0.5 | — | -163% | +| gpt-5.5-light | jiekou | — | $5 | $0.025 | — | 100% | +| grok-code-fast-1 | jiekou | — | $0.19 | $0.019 | — | 90% | +| minimax-minimax-m2.1 | jiekou | — | $0.55 | $0.03 | — | 95% | +| moonshotai-kimi-k2.5 | jiekou | — | $0.6 | $0.1 | — | 83% | +| o3 | jiekou | — | $1.045 | $0.475 | — | 55% | +| o3-mini | jiekou | — | $1.045 | $0.5225 | — | 50% | +| zai-org-glm-4.7-flash | jiekou | — | $0.6 | $0.01 | — | 98% | +| claude-3-7-sonnet-20250219 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-3-opus | llmgateway | — | $15 | $1.5 | $18.75 | 90% | +| claude-haiku-4-5 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | +| claude-haiku-4-5-20251001 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | +| claude-opus-4-1-20250805 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-20250514 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-5-20251101 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-6 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-7 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | +| claude-sonnet-4-20250514 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5-20250929 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-6 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | +| deepseek-v3.1 | llmgateway | — | $0.56 | $0.112 | — | 80% | +| deepseek-v3.2 | llmgateway | — | $0.28 | $0.028 | — | 90% | +| deepseek-v4-flash | llmgateway | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | llmgateway | — | $0.435 | $0.003625 | — | 99% | +| gemini-2.5-flash | llmgateway | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-image | llmgateway | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-lite | llmgateway | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-pro | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gemini-3-flash-preview | llmgateway | — | $0.5 | $0.05 | — | 90% | +| gemini-3-pro-image-preview | llmgateway | — | $2 | $0.2 | — | 90% | +| gemini-3.1-flash-lite | llmgateway | — | $0.25 | $0.025 | $0.08333 | 90% | +| gemini-3.1-pro-preview | llmgateway | — | $2 | $0.2 | — | 90% | +| gemini-3.5-flash | llmgateway | — | $1.5 | $0.15 | $0.08333 | 90% | +| gemini-pro-latest | llmgateway | — | $2 | $0.2 | — | 90% | +| glm-4.5 | llmgateway | — | $0.6 | $0.11 | — | 82% | +| glm-4.5-air | llmgateway | — | $0.2 | $0.03 | — | 85% | +| glm-4.5-airx | llmgateway | — | $1.1 | $0.22 | — | 80% | +| glm-4.5-x | llmgateway | — | $2.2 | $0.45 | — | 80% | +| glm-4.5v | llmgateway | — | $0.6 | $0.11 | — | 82% | +| glm-4.6v | llmgateway | — | $0.3 | $0.05 | — | 83% | +| glm-4.6v-flashx | llmgateway | — | $0.04 | $0.004 | — | 90% | +| glm-4.7 | llmgateway | — | $0.6 | $0.11 | — | 82% | +| glm-4.7-flashx | llmgateway | — | $0.07 | $0.01 | — | 86% | +| glm-5.1 | llmgateway | — | $1.4 | $0.26 | — | 81% | +| gpt-4.1 | llmgateway | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini | llmgateway | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano | llmgateway | — | $0.1 | $0.025 | — | 75% | +| gpt-4o | llmgateway | — | $2.5 | $1.25 | — | 50% | +| gpt-5 | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gpt-5-chat-latest | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gpt-5-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | +| gpt-5-nano | llmgateway | — | $0.05 | $0.005 | — | 90% | +| gpt-5.1 | llmgateway | — | $1.25 | $0.125 | — | 90% | +| gpt-5.1-codex-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | +| gpt-5.2 | llmgateway | — | $1.75 | $0.175 | — | 90% | +| gpt-5.2-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | +| gpt-5.3-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | +| gpt-5.4 | llmgateway | — | $2.5 | $0.25 | — | 90% | +| gpt-5.4-mini | llmgateway | — | $0.75 | $0.075 | — | 90% | +| gpt-5.4-nano | llmgateway | — | $0.2 | $0.02 | — | 90% | +| gpt-5.5 | llmgateway | — | $5 | $0.5 | — | 90% | +| grok-4 | llmgateway | — | $3 | $0.75 | — | 75% | +| grok-4-20-beta-0309-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-20-beta-0309-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-20-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-20-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | +| grok-4-3 | llmgateway | — | $1.25 | $0.3125 | — | 75% | +| kimi-k2-thinking | llmgateway | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking-turbo | llmgateway | — | $1.15 | $0.15 | — | 87% | +| kimi-k2.5 | llmgateway | — | $0.6 | $0.1 | — | 83% | +| kimi-k2.6 | llmgateway | — | $0.95 | $0.16 | — | 83% | +| mimo-v2-flash | llmgateway | — | $0.1 | $0.01 | — | 90% | +| mimo-v2-omni | llmgateway | — | $0.4 | $0.08 | — | 80% | +| mimo-v2-pro | llmgateway | — | $1 | $0.2 | — | 80% | +| mimo-v2.5 | llmgateway | — | $0.4 | $0.08 | — | 80% | +| mimo-v2.5-pro | llmgateway | — | $1 | $0.2 | — | 80% | +| minimax-m2 | llmgateway | — | $0.2 | $0.03 | — | 85% | +| minimax-m2.5-highspeed | llmgateway | — | $0.6 | $0.03 | — | 95% | +| minimax-m2.7 | llmgateway | — | $0.3 | $0.06 | — | 80% | +| minimax-m2.7-highspeed | llmgateway | — | $0.6 | $0.06 | — | 90% | +| o1 | llmgateway | — | $15 | $7.5 | — | 50% | +| o3 | llmgateway | — | $2 | $0.5 | — | 75% | +| o3-mini | llmgateway | — | $1.1 | $0.55 | — | 50% | +| o4-mini | llmgateway | — | $1.1 | $0.275 | — | 75% | +| qwen-flash | llmgateway | — | $0.05 | $0.01 | $0.0625 | 80% | +| qwen-plus | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | +| qwen-plus-latest | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | +| qwen3-coder-flash | llmgateway | — | $0.3 | $0.06 | $0.375 | 80% | +| qwen3-coder-next | llmgateway | — | $0.108 | $0.06 | — | 44% | +| qwen3-coder-plus | llmgateway | — | $6 | $1.2 | $7.5 | 80% | +| qwen3-max | llmgateway | — | $3 | $0.6 | $3.75 | 80% | +| qwen3-max-2026-01-23 | llmgateway | — | $1.2 | $0.24 | $1.5 | 80% | +| qwen3-vl-flash | llmgateway | — | $0.05 | $0.01 | — | 80% | +| qwen3-vl-plus | llmgateway | — | $0.2 | $0.04 | $0.25 | 80% | +| qwen3.6-max-preview | llmgateway | — | $1.3 | $0.13 | — | 90% | +| qwen3.6-plus | llmgateway | — | $0.5 | $0.05 | — | 90% | +| seed-1-6-250615 | llmgateway | — | $0.25 | $0.05 | — | 80% | +| seed-1-6-250915 | llmgateway | — | $0.25 | $0.05 | — | 80% | +| seed-1-6-flash-250715 | llmgateway | — | $0.07 | $0.015 | — | 79% | +| seed-1-8-251228 | llmgateway | — | $0.25 | $0.05 | — | 80% | +| aion-labs--aion-2.0 | martian | — | $0.8 | $0.2 | — | 75% | +| alibaba--tongyi-deepresearch-30b-a3b | martian | — | $0.09 | $0.09 | — | 0% | +| amazon--nova-premier-v1 | martian | — | $2.5 | $0.625 | — | 75% | +| anthropic--claude-haiku-4-5 | martian | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-haiku-4-5-20251001 | martian | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4-0 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-1 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-1-20250805 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-20250514 | martian | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-5 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-5-20251101 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-6 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-7 | martian | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-sonnet-4-0 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-20250514 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-5 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-5-20250929 | martian | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-6 | martian | — | $3 | $0.3 | $3.75 | 90% | +| arcee-ai--trinity-large-thinking | martian | — | $0.22 | $0.06 | — | 73% | +| bytedance--ui-tars-1.5-7b | martian | — | $0.1 | $0.1 | — | 0% | +| deepinfra--deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | +| deepinfra--deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | +| deepinfra--deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | +| deepinfra--deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | +| deepinfra--deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | +| deepinfra--google--gemini-2.5-flash | martian | — | $0.3 | $0.03 | $0.083333 | 90% | +| deepinfra--google--gemini-2.5-pro | martian | — | $1.25 | $0.125 | $0.375 | 90% | +| deepinfra--z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | +| deepinfra--z-ai--glm-4.7-flash | martian | — | $0.06 | $0.01 | — | 83% | +| deepinfra--z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | +| deepinfra--z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | +| deepseek--deepseek-chat-v3-0324 | martian | — | $0.2 | $0.135 | — | 32% | +| deepseek--deepseek-chat-v3.1 | martian | — | $0.21 | $0.13 | — | 38% | +| deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | +| deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | +| deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | +| deepseek--deepseek-v3.2-speciale | martian | — | $0.287 | $0.058 | — | 80% | +| deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | +| deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | +| ibm-granite--granite-4.1-8b | martian | — | $0.05 | $0.05 | — | 0% | +| inception--mercury-2 | martian | — | $0.25 | $0.025 | — | 90% | +| inclusionai--ling-2.6-1t | martian | — | $0.3 | $0.06 | — | 80% | +| inclusionai--ling-2.6-flash | martian | — | $0.08 | $0.016 | — | 80% | +| kwaipilot--kat-coder-pro-v2 | martian | — | $0.3 | $0.06 | — | 80% | +| microsoft--phi-4-mini-instruct | martian | — | $0.08 | $0.08 | — | 0% | +| minimax--minimax-m2 | martian | — | $0.255 | $0.03 | — | 88% | +| minimax--minimax-m2-her | martian | — | $0.3 | $0.03 | — | 90% | +| minimax--minimax-m2.1 | martian | — | $0.29 | $0.03 | — | 90% | +| mistralai--codestral-2508 | martian | — | $0.3 | $0.03 | — | 90% | +| mistralai--devstral-medium | martian | — | $0.4 | $0.04 | — | 90% | +| mistralai--devstral-small | martian | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-14b-2512 | martian | — | $0.2 | $0.02 | — | 90% | +| mistralai--ministral-3b-2512 | martian | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-8b-2512 | martian | — | $0.15 | $0.015 | — | 90% | +| mistralai--mistral-large | martian | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2407 | martian | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2411 | martian | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2512 | martian | — | $0.5 | $0.05 | — | 90% | +| mistralai--mistral-medium-3 | martian | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-medium-3.1 | martian | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-saba | martian | — | $0.2 | $0.02 | — | 90% | +| mistralai--mistral-small-2603 | martian | — | $0.15 | $0.015 | — | 90% | +| mistralai--mixtral-8x22b-instruct | martian | — | $2 | $0.2 | — | 90% | +| mistralai--pixtral-large-2411 | martian | — | $2 | $0.2 | — | 90% | +| mistralai--voxtral-small-24b-2507 | martian | — | $0.1 | $0.01 | — | 90% | +| moonshotai--kimi-k2-thinking | martian | — | $0.6 | $0.15 | — | 75% | +| moonshotai--kimi-k2.5 | martian | — | $0.4 | $0.05 | — | 88% | +| moonshotai--kimi-k2.6 | martian | — | $0.74 | $0.25 | — | 66% | +| openai--gpt-4.1 | martian | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-2025-04-14 | martian | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-mini | martian | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-mini-2025-04-14 | martian | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-nano | martian | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4.1-nano-2025-04-14 | martian | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4o-2024-08-06 | martian | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-11-20 | martian | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-mini | martian | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-mini-2024-07-18 | martian | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-5 | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-2025-08-07 | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-codex | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-mini | martian | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-mini-2025-08-07 | martian | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-nano | martian | — | $0.05 | $0.01 | — | 80% | +| openai--gpt-5-nano-2025-08-07 | martian | — | $0.05 | $0.01 | — | 80% | +| openai--gpt-5.1 | martian | — | $1.25 | $0.13 | — | 90% | +| openai--gpt-5.1-2025-11-13 | martian | — | $1.25 | $0.13 | — | 90% | +| openai--gpt-5.1-codex | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-max | martian | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-mini | martian | — | $0.25 | $0.03 | — | 88% | +| openai--gpt-5.2 | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-2025-12-11 | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-codex | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-codex | martian | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.4 | martian | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-2026-03-05 | martian | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | martian | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-mini-2026-03-17 | martian | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | martian | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.4-nano-2026-03-17 | martian | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | martian | — | $5 | $0.5 | — | 90% | +| openai--gpt-5.5-2026-04-23 | martian | — | $5 | $0.5 | — | 90% | +| openai--o1 | martian | — | $15 | $7.5 | — | 50% | +| openai--o1-2024-12-17 | martian | — | $15 | $7.5 | — | 50% | +| openai--o3 | martian | — | $2 | $0.5 | — | 75% | +| openai--o3-2025-04-16 | martian | — | $2 | $0.5 | — | 75% | +| openai--o3-mini | martian | — | $1.1 | $0.55 | — | 50% | +| openai--o3-mini-2025-01-31 | martian | — | $1.1 | $0.55 | — | 50% | +| openai--o4-mini | martian | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-2025-04-16 | martian | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-deep-research | martian | — | $2 | $0.5 | — | 75% | +| openai--o4-mini-deep-research-2025-06-26 | martian | — | $2 | $0.5 | — | 75% | +| qwen--qwen-plus | martian | — | $0.26 | $0.052 | $0.325 | 80% | +| qwen--qwen-plus-2025-07-28 | martian | — | $0.26 | — | $0.325 | — | +| qwen--qwen-plus-2025-07-28--thinking | martian | — | $0.26 | — | $0.325 | — | +| qwen--qwen3-30b-a3b-thinking-2507 | martian | — | $0.08 | $0.08 | — | 0% | +| qwen--qwen3-8b | martian | — | $0.05 | $0.05 | — | 0% | +| qwen--qwen3-coder-flash | martian | — | $0.195 | $0.039 | $0.24375 | 80% | +| qwen--qwen3-coder-next | martian | — | $0.11 | $0.07 | — | 36% | +| qwen--qwen3-coder-plus | martian | — | $0.65 | $0.13 | $0.8125 | 80% | +| qwen--qwen3-max | martian | — | $0.78 | $0.156 | $0.975 | 80% | +| qwen--qwen3-vl-235b-a22b-instruct | martian | — | $0.2 | $0.11 | — | 45% | +| qwen--qwen3.5-35b-a3b | martian | — | $0.14 | $0.05 | — | 64% | +| qwen--qwen3.5-flash-02-23 | martian | — | $0.065 | — | $0.08125 | — | +| qwen--qwen3.5-plus-02-15 | martian | — | $0.26 | — | $0.325 | — | +| qwen--qwen3.6-35b-a3b | martian | — | $0.15 | $0.05 | — | 67% | +| qwen--qwen3.6-flash | martian | — | $0.25 | — | $0.3125 | — | +| qwen--qwen3.6-max-preview | martian | — | $1.04 | — | $1.3 | — | +| qwen--qwen3.6-plus | martian | — | $0.325 | — | $0.40625 | — | +| tencent--hy3-preview | martian | — | $0.066 | $0.029 | — | 56% | +| thedrummer--cydonia-24b-v4.1 | martian | — | $0.3 | $0.15 | — | 50% | +| thedrummer--skyfall-36b-v2 | martian | — | $0.55 | $0.25 | — | 55% | +| x-ai--grok-3 | martian | — | $3 | $0.75 | — | 75% | +| x-ai--grok-3-mini | martian | — | $0.3 | $0.075 | — | 75% | +| x-ai--grok-4 | martian | — | $3 | $0.75 | — | 75% | +| x-ai--grok-4.20-multi-agent | martian | — | $2 | $0.2 | — | 90% | +| x-ai--grok-4.3 | martian | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-code-fast-1 | martian | — | $0.2 | $0.02 | — | 90% | +| xiaomi--mimo-v2-flash | martian | — | $0.1 | $0.01 | — | 90% | +| xiaomi--mimo-v2-omni | martian | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2-pro | martian | — | $1 | $0.2 | — | 80% | +| xiaomi--mimo-v2.5 | martian | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | martian | — | $1 | $0.2 | — | 80% | +| z-ai--glm-4.5 | martian | — | $0.6 | $0.11 | — | 82% | +| z-ai--glm-4.5-air | martian | — | $0.13 | $0.025 | — | 81% | +| z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | +| z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | +| z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | +| MiniMax-M2 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | +| MiniMax-M2.1 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | +| MiniMax-M2.1-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | +| MiniMax-M2.5 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | +| MiniMax-M2.5-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | +| MiniMax-M2.7 | minimax | — | $2.1 | $0.42 | $2.625 | 80% | +| MiniMax-M2.7-highspeed | minimax | — | $4.2 | $0.42 | $2.625 | 90% | +| kimi-k2-0711-preview | moonshotai | — | $4 | $1 | — | 75% | +| kimi-k2-0905-preview | moonshotai | — | $4 | $1 | — | 75% | +| kimi-k2-thinking | moonshotai | — | $4 | $1 | — | 75% | +| kimi-k2-thinking-turbo | moonshotai | — | $8 | $1 | — | 88% | +| kimi-k2-turbo-preview | moonshotai | — | $8 | $1 | — | 88% | +| kimi-k2.5 | moonshotai | — | $4 | $0.7 | — | 82% | +| kimi-k2.6 | moonshotai | — | $6.5 | $1.1 | — | 83% | +| kimi-k2.6-long | moonshotai | — | $6.5 | $1.1 | — | 83% | +| kimi-vl-a3b | moonshotai | — | $4 | $0.7 | — | 82% | +| kimi-vl-a3b-thinking | moonshotai | — | $4 | $0.7 | — | 82% | +| aion-labs--aion-2.5 | nanogpt | — | $1 | $0.35 | — | 65% | +| alibaba--qwen3.6-flash | nanogpt | — | $0.19 | $0.02 | $0.24 | 89% | +| deepseek--deepseek-latest | nanogpt | — | $1.1 | $0.11 | — | 90% | +| deepseek--deepseek-v4-flash | nanogpt | — | $0.14 | $0.028 | — | 80% | +| deepseek--deepseek-v4-flash--thinking | nanogpt | — | $0.14 | $0.028 | — | 80% | +| deepseek--deepseek-v4-pro | nanogpt | — | $1.1 | $0.11 | — | 90% | +| deepseek--deepseek-v4-pro--thinking | nanogpt | — | $1.1 | $0.11 | — | 90% | +| deepseek--deepseek-v4-pro-cheaper | nanogpt | — | $0.435 | $0.0435 | — | 90% | +| deepseek--deepseek-v4-pro-cheaper--thinking | nanogpt | — | $0.435 | $0.0435 | — | 90% | +| ernie-5.1 | nanogpt | — | $0.75 | $0.75 | — | 0% | +| ernie-5.1--thinking | nanogpt | — | $0.75 | $0.75 | — | 0% | +| google--gemini-3-flash-preview | nanogpt | — | $0.5 | — | $0.08333 | — | +| google--gemini-3.1-flash-lite | nanogpt | — | $0.25 | — | $0.08333 | — | +| google--gemini-3.1-pro-preview | nanogpt | — | $2 | — | $0.375 | — | +| google--gemini-3.1-pro-preview-customtools | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-pro-preview-high | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-pro-preview-low | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.5-flash | nanogpt | — | $1.5 | — | $0.08333 | — | +| google--gemini-flash-lite-latest | nanogpt | — | $0.25 | $0.025 | $0.08333 | 90% | +| google--gemini-pro-latest | nanogpt | — | $2 | $0.2 | $0.375 | 90% | +| ibm-granite--granite-4.1-8b | nanogpt | — | $0.05 | $0.05 | — | 0% | +| inclusionai--ling-2.6-1t | nanogpt | — | $0.3 | $0.06 | — | 80% | +| mercury-2 | nanogpt | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5.4 | nanogpt | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | nanogpt | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | nanogpt | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | nanogpt | — | $5 | $0.5 | — | 90% | +| openai--gpt-chat-latest | nanogpt | — | $5 | $0.5 | — | 90% | +| openai--gpt-latest | nanogpt | — | $5 | $0.5 | — | 90% | +| qwen-3.6-plus | nanogpt | — | $0.325 | $0.0325 | $0.40625 | 90% | +| sarvam-105b | nanogpt | — | $0.045 | $0.028 | — | 38% | +| sarvam-30b | nanogpt | — | $0.028 | $0.017 | — | 39% | +| tencent--hy3-preview | nanogpt | — | $0.066 | $0.029 | — | 56% | +| upstage--solar-pro-3 | nanogpt | — | $0.15 | $0.015 | — | 90% | +| x-ai--grok-4.3 | nanogpt | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-latest | nanogpt | — | $1.25 | $0.2 | — | 84% | +| xiaomi--mimo-v2-omni | nanogpt | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2-pro | nanogpt | — | $1 | $0.2 | — | 80% | +| xiaomi--mimo-v2.5 | nanogpt | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | nanogpt | — | $0.6 | $0.12 | — | 80% | +| z-ai--glm-5-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | +| z-ai--glm-5v-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | +| z-ai--glm-5v-turbo--thinking | nanogpt | — | $1.2 | $0.24 | — | 80% | +| zai-org--glm-4.7-original | nanogpt | — | $0.6 | $0.11 | — | 82% | +| zai-org--glm-4.7-original--thinking | nanogpt | — | $0.6 | $0.11 | — | 82% | +| gpt-4.1 | openai | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini | openai | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano | openai | — | $0.1 | $0.025 | — | 75% | +| gpt-4o | openai | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-mini | openai | — | $0.15 | $0.075 | — | 50% | +| o1 | openai | — | $15 | $7.5 | — | 50% | +| o1-mini | openai | — | $1.5 | $0.75 | — | 50% | +| o3 | openai | — | $10 | $2.5 | — | 75% | +| o3-mini | openai | — | $1.1 | $0.55 | — | 50% | +| o4-mini | openai | — | $1.1 | $0.275 | — | 75% | +| aion-labs--aion-2.0 | openrouter | — | $0.8 | $0.2 | — | 75% | +| alibaba--tongyi-deepresearch-30b-a3b | openrouter | — | $0.09 | $0.09 | — | 0% | +| amazon--nova-premier-v1 | openrouter | — | $2.5 | $0.625 | — | 75% | +| anthropic--claude-3-haiku | openrouter | — | $0.25 | $0.03 | $0.3 | 88% | +| anthropic--claude-3.5-haiku | openrouter | — | $0.8 | $0.08 | $1 | 90% | +| anthropic--claude-haiku-4.5 | openrouter | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4 | openrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.1 | openrouter | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4.5 | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.6 | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.6-fast | openrouter | — | $30 | $3 | $37.5 | 90% | +| anthropic--claude-opus-4.7 | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4.7-fast | openrouter | — | $30 | $3 | $37.5 | 90% | +| anthropic--claude-sonnet-4 | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4.5 | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4.6 | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| arcee-ai--trinity-large-thinking | openrouter | — | $0.22 | $0.06 | — | 73% | +| bytedance--ui-tars-1.5-7b | openrouter | — | $0.1 | $0.1 | — | 0% | +| deepseek--deepseek-chat-v3-0324 | openrouter | — | $0.2 | $0.135 | — | 32% | +| deepseek--deepseek-chat-v3.1 | openrouter | — | $0.21 | $0.13 | — | 38% | +| deepseek--deepseek-r1-0528 | openrouter | — | $0.5 | $0.35 | — | 30% | +| deepseek--deepseek-v3.1-terminus | openrouter | — | $0.27 | $0.13 | — | 52% | +| deepseek--deepseek-v3.2 | openrouter | — | $0.252 | $0.0252 | — | 90% | +| deepseek--deepseek-v3.2-speciale | openrouter | — | $0.287 | $0.058 | — | 80% | +| deepseek--deepseek-v4-flash | openrouter | — | $0.112 | $0.022 | — | 80% | +| deepseek--deepseek-v4-pro | openrouter | — | $0.435 | $0.003625 | — | 99% | +| google--gemini-2.0-flash-001 | openrouter | — | $0.1 | $0.025 | $0.083333 | 75% | +| google--gemini-2.5-flash | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | +| google--gemini-2.5-flash-image | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | +| google--gemini-2.5-flash-lite | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | +| google--gemini-2.5-flash-lite-preview-09-2025 | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | +| google--gemini-2.5-pro | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | +| google--gemini-2.5-pro-preview | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | +| google--gemini-2.5-pro-preview-05-06 | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | +| google--gemini-3-flash-preview | openrouter | — | $0.5 | $0.05 | $0.083333 | 90% | +| google--gemini-3-pro-image-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-flash-lite | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | +| google--gemini-3.1-flash-lite-preview | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | +| google--gemini-3.1-pro-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.1-pro-preview-customtools | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| google--gemini-3.5-flash | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | +| ibm-granite--granite-4.1-8b | openrouter | — | $0.05 | $0.05 | — | 0% | +| inception--mercury-2 | openrouter | — | $0.25 | $0.025 | — | 90% | +| inclusionai--ling-2.6-1t | openrouter | — | $0.3 | $0.06 | — | 80% | +| inclusionai--ling-2.6-flash | openrouter | — | $0.01 | $0.002 | — | 80% | +| inclusionai--ring-2.6-1t | openrouter | — | $0.075 | $0.015 | — | 80% | +| kwaipilot--kat-coder-pro-v2 | openrouter | — | $0.3 | $0.06 | — | 80% | +| microsoft--phi-4-mini-instruct | openrouter | — | $0.08 | $0.08 | — | 0% | +| minimax--minimax-m2 | openrouter | — | $0.255 | $0.03 | — | 88% | +| minimax--minimax-m2-her | openrouter | — | $0.3 | $0.03 | — | 90% | +| minimax--minimax-m2.1 | openrouter | — | $0.29 | $0.03 | — | 90% | +| mistralai--codestral-2508 | openrouter | — | $0.3 | $0.03 | — | 90% | +| mistralai--devstral-2512 | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--devstral-medium | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--devstral-small | openrouter | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-14b-2512 | openrouter | — | $0.2 | $0.02 | — | 90% | +| mistralai--ministral-3b-2512 | openrouter | — | $0.1 | $0.01 | — | 90% | +| mistralai--ministral-8b-2512 | openrouter | — | $0.15 | $0.015 | — | 90% | +| mistralai--mistral-large | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2407 | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--mistral-large-2512 | openrouter | — | $0.5 | $0.05 | — | 90% | +| mistralai--mistral-medium-3 | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-medium-3.1 | openrouter | — | $0.4 | $0.04 | — | 90% | +| mistralai--mistral-saba | openrouter | — | $0.2 | $0.02 | — | 90% | +| mistralai--mistral-small-2603 | openrouter | — | $0.15 | $0.015 | — | 90% | +| mistralai--mixtral-8x22b-instruct | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--pixtral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | +| mistralai--voxtral-small-24b-2507 | openrouter | — | $0.1 | $0.01 | — | 90% | +| moonshotai--kimi-k2.5 | openrouter | — | $0.4 | $0.09 | — | 78% | +| moonshotai--kimi-k2.6 | openrouter | — | $0.73 | $0.25 | — | 66% | +| openai--gpt-4.1 | openrouter | — | $2 | $0.5 | — | 75% | +| openai--gpt-4.1-mini | openrouter | — | $0.4 | $0.1 | — | 75% | +| openai--gpt-4.1-nano | openrouter | — | $0.1 | $0.025 | — | 75% | +| openai--gpt-4o-2024-08-06 | openrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-2024-11-20 | openrouter | — | $2.5 | $1.25 | — | 50% | +| openai--gpt-4o-mini | openrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-4o-mini-2024-07-18 | openrouter | — | $0.15 | $0.075 | — | 50% | +| openai--gpt-5 | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-chat | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-codex | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5-image | openrouter | — | $10 | $1.25 | — | 88% | +| openai--gpt-5-image-mini | openrouter | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5-mini | openrouter | — | $0.25 | $0.025 | — | 90% | +| openai--gpt-5-nano | openrouter | — | $0.05 | $0.01 | — | 80% | +| openai--gpt-5.1 | openrouter | — | $1.25 | $0.13 | — | 90% | +| openai--gpt-5.1-chat | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-max | openrouter | — | $1.25 | $0.125 | — | 90% | +| openai--gpt-5.1-codex-mini | openrouter | — | $0.25 | $0.03 | — | 88% | +| openai--gpt-5.2 | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-chat | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.2-codex | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-chat | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.3-codex | openrouter | — | $1.75 | $0.175 | — | 90% | +| openai--gpt-5.4 | openrouter | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-image-2 | openrouter | — | $8 | $2 | — | 75% | +| openai--gpt-5.4-mini | openrouter | — | $0.75 | $0.075 | — | 90% | +| openai--gpt-5.4-nano | openrouter | — | $0.2 | $0.02 | — | 90% | +| openai--gpt-5.5 | openrouter | — | $5 | $0.5 | — | 90% | +| openai--gpt-chat-latest | openrouter | — | $5 | $0.5 | — | 90% | +| openai--gpt-oss-safeguard-20b | openrouter | — | $0.075 | $0.037 | — | 51% | +| openai--o1 | openrouter | — | $15 | $7.5 | — | 50% | +| openai--o3 | openrouter | — | $2 | $0.5 | — | 75% | +| openai--o3-deep-research | openrouter | — | $10 | $2.5 | — | 75% | +| openai--o3-mini | openrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o3-mini-high | openrouter | — | $1.1 | $0.55 | — | 50% | +| openai--o4-mini | openrouter | — | $1.1 | $0.275 | — | 75% | +| openai--o4-mini-deep-research | openrouter | — | $2 | $0.5 | — | 75% | +| openai--o4-mini-high | openrouter | — | $1.1 | $0.275 | — | 75% | +| qwen--qwen-plus | openrouter | — | $0.26 | $0.052 | $0.325 | 80% | +| qwen--qwen-plus-2025-07-28 | openrouter | — | $0.26 | — | $0.325 | — | +| qwen--qwen-plus-2025-07-28--thinking | openrouter | — | $0.26 | — | $0.325 | — | +| qwen--qwen3-30b-a3b-thinking-2507 | openrouter | — | $0.08 | $0.08 | — | 0% | +| qwen--qwen3-8b | openrouter | — | $0.05 | $0.05 | — | 0% | +| qwen--qwen3-coder-flash | openrouter | — | $0.195 | $0.039 | $0.24375 | 80% | +| qwen--qwen3-coder-next | openrouter | — | $0.11 | $0.07 | — | 36% | +| qwen--qwen3-coder-plus | openrouter | — | $0.65 | $0.13 | $0.8125 | 80% | +| qwen--qwen3-max | openrouter | — | $0.78 | $0.156 | $0.975 | 80% | +| qwen--qwen3-vl-235b-a22b-instruct | openrouter | — | $0.2 | $0.11 | — | 45% | +| qwen--qwen3.5-397b-a17b | openrouter | — | $0.39 | $0.195 | — | 50% | +| qwen--qwen3.5-flash-02-23 | openrouter | — | $0.065 | — | $0.08125 | — | +| qwen--qwen3.5-plus-02-15 | openrouter | — | $0.26 | — | $0.325 | — | +| qwen--qwen3.6-35b-a3b | openrouter | — | $0.15 | $0.05 | — | 67% | +| qwen--qwen3.6-flash | openrouter | — | $0.1875 | — | $0.234375 | — | +| qwen--qwen3.6-max-preview | openrouter | — | $1.04 | — | $1.3 | — | +| qwen--qwen3.6-plus | openrouter | — | $0.325 | — | $0.40625 | — | +| tencent--hy3-preview | openrouter | — | $0.066 | $0.029 | — | 56% | +| thedrummer--cydonia-24b-v4.1 | openrouter | — | $0.3 | $0.15 | — | 50% | +| thedrummer--skyfall-36b-v2 | openrouter | — | $0.55 | $0.25 | — | 55% | +| upstage--solar-pro-3 | openrouter | — | $0.15 | $0.015 | — | 90% | +| x-ai--grok-4.20 | openrouter | — | $1.25 | $0.2 | — | 84% | +| x-ai--grok-4.20-multi-agent | openrouter | — | $2 | $0.2 | — | 90% | +| x-ai--grok-4.3 | openrouter | — | $1.25 | $0.2 | — | 84% | +| xiaomi--mimo-v2-flash | openrouter | — | $0.1 | $0.01 | — | 90% | +| xiaomi--mimo-v2-omni | openrouter | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2-pro | openrouter | — | $1 | $0.2 | — | 80% | +| xiaomi--mimo-v2.5 | openrouter | — | $0.4 | $0.08 | — | 80% | +| xiaomi--mimo-v2.5-pro | openrouter | — | $1 | $0.2 | — | 80% | +| z-ai--glm-4.5 | openrouter | — | $0.6 | $0.11 | — | 82% | +| z-ai--glm-4.5-air | openrouter | — | $0.13 | $0.025 | — | 81% | +| z-ai--glm-4.5v | openrouter | — | $0.6 | $0.11 | — | 82% | +| z-ai--glm-4.6 | openrouter | — | $0.43 | $0.08 | — | 81% | +| z-ai--glm-4.6v | openrouter | — | $0.3 | $0.05 | — | 83% | +| z-ai--glm-4.7 | openrouter | — | $0.4 | $0.08 | — | 80% | +| z-ai--glm-4.7-flash | openrouter | — | $0.06 | $0.01 | — | 83% | +| z-ai--glm-5 | openrouter | — | $0.6 | $0.12 | — | 80% | +| z-ai--glm-5-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | +| z-ai--glm-5v-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | +| ~anthropic--claude-haiku-latest | openrouter | — | $1 | $0.1 | $1.25 | 90% | +| ~anthropic--claude-opus-latest | openrouter | — | $5 | $0.5 | $6.25 | 90% | +| ~anthropic--claude-sonnet-latest | openrouter | — | $3 | $0.3 | $3.75 | 90% | +| ~google--gemini-flash-latest | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | +| ~google--gemini-pro-latest | openrouter | — | $2 | $0.2 | $0.375 | 90% | +| ~moonshotai--kimi-latest | openrouter | — | $0.73 | $0.25 | — | 66% | +| ~openai--gpt-latest | openrouter | — | $5 | $0.5 | — | 90% | +| ~openai--gpt-mini-latest | openrouter | — | $0.75 | $0.075 | — | 90% | +| deepseek--deepseek-v3-0324 | ppio | — | $2 | $0.6 | — | 70% | +| deepseek--deepseek-v3.1 | ppio | — | $4 | $2 | — | 50% | +| deepseek--deepseek-v3.1-terminus | ppio | — | $4 | $2 | — | 50% | +| deepseek--deepseek-v3.2 | ppio | — | $2 | $0.2 | — | 90% | +| deepseek--deepseek-v4-flash | ppio | — | $1 | $0.2 | — | 80% | +| deepseek--deepseek-v4-pro | ppio | — | $12 | $1 | — | 92% | +| minimax--minimax-m2 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | +| minimax--minimax-m2.1 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | +| minimax--minimax-m2.5 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | +| minimax--minimax-m2.5-highspeed | ppio | — | $4.2 | $0.21 | $2.625 | 95% | +| minimax--minimax-m2.7 | ppio | — | $2.1 | $0.42 | $2.625 | 80% | +| minimax--minimax-m2.7-highspeed | ppio | — | $4.2 | $0.42 | $2.625 | 90% | +| moonshotai--kimi-k2.5 | ppio | — | $4 | $0.7 | — | 82% | +| moonshotai--kimi-k2.6 | ppio | — | $6.5 | $1.1 | — | 83% | +| xiaomimimo--mimo-v2-flash | ppio | — | $0.7 | $0.07 | — | 90% | +| zai-org--glm-4.5 | ppio | — | $4 | $0.8 | — | 80% | +| zai-org--glm-4.5-air | ppio | — | $1.2 | $0.24 | — | 80% | +| zai-org--glm-4.5v | ppio | — | $4 | $0.8 | — | 80% | +| zai-org--glm-4.6 | ppio | — | $4 | $0.8 | $0.8 | 80% | +| zai-org--glm-4.7-flash | ppio | — | $0.5 | $0.1 | — | 80% | +| gemma-3-27b | privatemode | — | $0.77 | $0.08 | — | 90% | +| gemma-4-31b | privatemode | — | $0.77 | $0.08 | — | 90% | +| gpt-oss-120b | privatemode | — | $0.43 | $0.04 | — | 91% | +| kimi-k2-5 | privatemode | — | $1.55 | $0.15 | — | 90% | +| qwen3-coder-30b-a3b | privatemode | — | $0.43 | $0.04 | — | 91% | +| alibaba--qwen-max | requesty | — | $1.6 | $1.6 | $1.6 | 0% | +| alibaba--qwen-plus | requesty | — | $0.4 | $0.4 | $0.4 | 0% | +| alibaba--qwen-turbo | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| alibaba--qwen3-30b-a3b-instruct-2507 | requesty | — | $0.2 | $0.2 | $0.8 | 0% | +| alibaba--qwen3-coder-flash | requesty | — | $0.3 | $0.08 | $0.3 | 73% | +| alibaba--qwen3.5 | requesty | — | $0.6 | $0.6 | $3.6 | 0% | +| anthropic--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | +| anthropic--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| anthropic--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| anthropic--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| anthropic--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| azure--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| azure--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| azure--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| azure--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| azure--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | +| azure--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | +| azure--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| azure--gpt-5.2 | requesty | — | $1.75 | $0.175 | $14 | 90% | +| azure--openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| azure--openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| azure--openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| bedrock--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | +| bedrock--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | +| bedrock--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| bedrock--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| bedrock--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| bedrock--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| bedrock--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| bedrock--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| deepinfra--deepseek-ai--deepseek-r1 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | +| deepinfra--deepseek-ai--deepseek-r1-distill-llama-70b | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--deepseek-ai--deepseek-v3 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | +| deepinfra--deepseek-ai--deepseek-v3.1 | requesty | — | $0.3 | $0.3 | $1 | 0% | +| deepinfra--kimi k2.5 | requesty | — | $0.45 | $0.07 | — | 84% | +| deepinfra--meta-llama--llama-3.2-90b-vision-instruct | requesty | — | $0.35 | $0.35 | $0.35 | 0% | +| deepinfra--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.12 | $0.12 | $0.12 | 0% | +| deepinfra--meta-llama--meta-llama-3.1-405b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | +| deepinfra--meta-llama--meta-llama-3.1-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.02 | $0.02 | $0.02 | 0% | +| deepinfra--microsoft--phi-4 | requesty | — | $0.07 | $0.07 | $0.07 | 0% | +| deepinfra--qwen--qwen2.5-72b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | +| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | — | $0.07 | $0.07 | $0.07 | 0% | +| deepinfra--qwen--qwen3-235b-a22b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| deepinfra--qwen--qwen3-32b | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| deepinfra--qwen--qwen3-coder-480b-a35b-instruct | requesty | — | $0.4 | $0.4 | $1.6 | 0% | +| deepinfra--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | +| deepinfra--zai-org--glm-4.5-air | requesty | — | $0.2 | $0.2 | $1.1 | 0% | +| deepseek--deepseek-chat | requesty | — | $0.14 | $0.028 | $0.14 | 80% | +| deepseek--deepseek-reasoner | requesty | — | $0.14 | $0.028 | $0.14 | 80% | +| deepseek--deepseek-v4-flash | requesty | — | $0.14 | $0.028 | $0.14 | 80% | +| deepseek--deepseek-v4-pro | requesty | — | $1.74 | $0.145 | $1.74 | 92% | +| fireworks--deepseek-v3.2 | requesty | — | $0.56 | $0.28 | — | 50% | +| fireworks--deepseek-v4-pro | requesty | — | $1.74 | $0.15 | — | 91% | +| fireworks--glm-5 | requesty | — | $1 | $0.2 | — | 80% | +| fireworks--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | +| fireworks--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $3 | 83% | +| fireworks--kimi-k2.6 | requesty | — | $0.95 | $0.16 | — | 83% | +| fireworks--minimax-m2.5 | requesty | — | $0.3 | $0.03 | — | 90% | +| fireworks--minimax-m2.7 | requesty | — | $0.3 | $0.06 | — | 80% | +| fireworks--qwen3.6-plus | requesty | — | $0.5 | $0.1 | — | 80% | +| google--gemini-2.0-flash-001 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| google--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | +| google--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | +| google--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | +| google--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | +| google--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| google--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | +| google--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| groq--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.75 | 0% | +| groq--openai--gpt-oss-20b | requesty | — | $0.1 | $0.1 | $0.5 | 0% | +| inceptron--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | +| inceptron--kimi-k2.6 | requesty | — | $0.8 | $0.2 | — | 75% | +| inceptron--minimax-m2.5 | requesty | — | $0.28 | $0.03 | — | 89% | +| minimaxi--minimax-m2 | requesty | — | $0.3 | $0.3 | $1.2 | 0% | +| minimaxi--minimax-m2.5 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | +| minimaxi--minimax-m2.5-highspeed | requesty | — | $0.6 | $0.06 | $2.4 | 90% | +| minimaxi--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | +| minimaxi--minimax-m2.7-highspeed | requesty | — | $0.6 | $0.06 | $1.2 | 90% | +| mistral--codestral-latest | requesty | — | $0.3 | $0.3 | $0.9 | 0% | +| mistral--devstral-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | +| mistral--devstral-medium-2507 | requesty | — | $0.4 | $0.4 | $2 | 0% | +| mistral--devstral-small-2507 | requesty | — | $0.1 | $0.1 | $0.3 | 0% | +| mistral--devstral-small-latest | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| mistral--mistral-large-latest | requesty | — | $0.5 | $0.5 | $0.5 | 0% | +| mistral--mistral-medium-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | +| mistral--mistral-small-2503 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| mistral--mistral-small-2603 | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| mistral--mistral-small-latest | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| mistral--open-mistral-7b | requesty | — | $0.25 | $0.25 | $0.25 | 0% | +| mistral--pixtral-large-latest | requesty | — | $2 | $2 | $5 | 0% | +| moonshot--kimi-k2-0711-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-0905-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-thinking | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-thinking-turbo | requesty | — | $0.6 | $0.15 | $0.6 | 75% | +| moonshot--kimi-k2-turbo-preview | requesty | — | $1.2 | $0.3 | $5 | 75% | +| moonshot--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $0.6 | 83% | +| moonshot--kimi-k2.6 | requesty | — | $0.95 | $0.16 | $0.96 | 83% | +| nebius--deepseek-ai--deepseek-v3.2 | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| nebius--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.13 | $0.13 | $0.13 | 0% | +| nebius--moonshotai--kimi-k2.5 | requesty | — | $0.5 | $0.5 | $2.5 | 0% | +| nebius--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| novita--deepseek--deepseek-prover-v2-671b | requesty | — | $0.7 | $0.7 | $0.7 | 0% | +| novita--deepseek--deepseek-r1 | requesty | — | $4 | $4 | $4 | 0% | +| novita--deepseek--deepseek-r1-distill-llama-70b | requesty | — | $0.8 | $0.8 | $0.8 | 0% | +| novita--deepseek--deepseek-r1-distill-qwen-14b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | +| novita--deepseek--deepseek-r1-distill-qwen-32b | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| novita--deepseek--deepseek-r1-turbo | requesty | — | $0.7 | $0.7 | $2.5 | 0% | +| novita--deepseek--deepseek-v3-0324 | requesty | — | $0.4 | $0.4 | $0.4 | 0% | +| novita--deepseek--deepseek-v3-turbo | requesty | — | $0.4 | $0.4 | $0.4 | 0% | +| novita--deepseek--deepseek-v3.2 | requesty | — | $0.269 | $0.1345 | $0.269 | 50% | +| novita--deepseek--deepseek_v3 | requesty | — | $0.89 | $0.89 | $0.89 | 0% | +| novita--glm-5 | requesty | — | $1 | $0.2 | — | 80% | +| novita--gryphe--mythomax-l2-13b | requesty | — | $0.09 | $0.09 | $0.09 | 0% | +| novita--meta-llama--llama-3-70b-instruct | requesty | — | $0.51 | $0.51 | $0.51 | 0% | +| novita--meta-llama--llama-3-8b-instruct | requesty | — | $0.04 | $0.04 | $0.04 | 0% | +| novita--meta-llama--llama-3.1-8b-instruct | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| novita--meta-llama--llama-3.2-1b-instruct | requesty | — | $0.02 | $0.02 | $0.02 | 0% | +| novita--meta-llama--llama-3.2-3b-instruct | requesty | — | $0.03 | $0.03 | $0.03 | 0% | +| novita--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.39 | $0.39 | $0.39 | 0% | +| novita--meta-llama--llama-4-maverick-17b-128e-instruct-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| novita--microsoft--wizardlm-2-8x22b | requesty | — | $0.62 | $0.62 | $0.62 | 0% | +| novita--minimax--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | +| novita--mistralai--mistral-nemo | requesty | — | $0.17 | $0.17 | $0.17 | 0% | +| novita--moonshotai--kimi-k2-instruct | requesty | — | $0.57 | $0.57 | $0.57 | 0% | +| novita--nousresearch--hermes-2-pro-llama-3-8b | requesty | — | $0.14 | $0.14 | $0.14 | 0% | +| novita--qwen--qwen-2.5-72b-instruct | requesty | — | $0.38 | $0.38 | $0.38 | 0% | +| novita--qwen--qwen2.5-vl-72b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | +| novita--qwen--qwen3-235b-a22b-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| novita--qwen--qwen3.5-397b-a17b | requesty | — | $0.6 | $0.6 | $3.6 | 0% | +| novita--sao10k--l3-70b-euryale-v2.1 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | +| novita--sao10k--l3-8b-lunaris | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| novita--sao10k--l3-8b-stheno-v3.2 | requesty | — | $0.05 | $0.05 | $0.05 | 0% | +| novita--sao10k--l31-70b-euryale-v2.2 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | +| novita--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | +| novita--zai-org--glm-4.6 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | +| openai--chatgpt-4o | requesty | — | $5 | $5 | $5 | 0% | +| openai--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| openai--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| openai--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| openai--gpt-4o | requesty | — | $2.5 | $1.25 | $2.5 | 50% | +| openai--gpt-4o-2024-05-13 | requesty | — | $2.5 | $2.5 | $2.5 | 0% | +| openai--gpt-4o-2024-08-06 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | +| openai--gpt-4o-2024-11-20 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | +| openai--gpt-4o-mini | requesty | — | $0.15 | $0.075 | $0.15 | 50% | +| openai--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | +| openai--gpt-5-mini:flex | requesty | — | $0.125 | $0.0125 | $0.125 | 90% | +| openai--gpt-5-mini:priority | requesty | — | $0.45 | $0.045 | $3.6 | 90% | +| openai--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | +| openai--gpt-5-nano:flex | requesty | — | $0.025 | $0.0025 | $0.025 | 90% | +| openai--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5.1-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | +| openai--gpt-5.2-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai--gpt-5.4 | requesty | — | $2.5 | $0.25 | — | 90% | +| openai--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | +| openai--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | +| openai--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | +| openai--gpt-5:flex | requesty | — | $0.625 | $0.0625 | $0.625 | 90% | +| openai--gpt-5:priority | requesty | — | $2.5 | $0.25 | $20 | 90% | +| openai--o1 | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o1-pro | requesty | — | $150 | $150 | $150 | 0% | +| openai--o1:high | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o1:low | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o1:medium | requesty | — | $15 | $7.5 | $15 | 50% | +| openai--o3 | requesty | — | $2 | $0.5 | $2 | 75% | +| openai--o3-deep-research | requesty | — | $10 | $2.5 | $10 | 75% | +| openai--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3-mini:high | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3-mini:low | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3-mini:medium | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai--o3:flex | requesty | — | $1 | $0.25 | $1 | 75% | +| openai--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai--o4-mini-deep-research | requesty | — | $2 | $0.5 | $2 | 75% | +| openai--o4-mini:flex | requesty | — | $0.55 | $0.138 | $0.55 | 75% | +| openai--o4-mini:high | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai--o4-mini:low | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai--o4-mini:medium | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | +| openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | +| openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | +| openai-responses--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai-responses--gpt-5-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | +| openai-responses--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | +| openai-responses--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | +| openai-responses--gpt-5-pro | requesty | — | $15 | $15 | $120 | 0% | +| openai-responses--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | +| openai-responses--gpt-5.1-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | +| openai-responses--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | +| openai-responses--gpt-5.2-codex | requesty | — | $1.75 | $0.175 | $1.75 | 90% | +| openai-responses--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai-responses--gpt-5.3-codex | requesty | — | $1.75 | $0.175 | $14 | 90% | +| openai-responses--gpt-5.4 | requesty | — | $2.5 | $0.25 | $2.5 | 90% | +| openai-responses--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | +| openai-responses--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | +| openai-responses--gpt-5.4-pro | requesty | — | $30 | $30 | $180 | 0% | +| openai-responses--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | +| openai-responses--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | +| openai-responses--o3-pro | requesty | — | $20 | $20 | $20 | 0% | +| openai-responses--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | +| parasail--parasail-gemma3-27b-it | requesty | — | $0.3 | $0.3 | $0.5 | 0% | +| parasail--parasail-kimi-k2-instruct | requesty | — | $0.99 | $0.99 | $2.99 | 0% | +| parasail--parasail-qwen25-vl-72b-instruct | requesty | — | $0.7 | $0.7 | $0.7 | 0% | +| parasail--parasail-qwen3-235b-a22b-instruct-2507 | requesty | — | $0.15 | $0.15 | $0.85 | 0% | +| perplexity--sonar | requesty | — | $1 | $1 | $1 | 0% | +| perplexity--sonar-pro | requesty | — | $3 | $3 | $3 | 0% | +| perplexity--sonar-reasoning-pro | requesty | — | $2 | $2 | $2 | 0% | +| together--deepseek-ai--deepseek-r1 | requesty | — | $3 | $3 | $7 | 0% | +| together--deepseek-ai--deepseek-v3 | requesty | — | $1.25 | $1.25 | $1.25 | 0% | +| together--meta-llama--llama-3.2-3b-instruct-turbo | requesty | — | $0.06 | $0.06 | $0.06 | 0% | +| together--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | +| together--meta-llama--llamaguard-2-8b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | +| together--meta-llama--meta-llama-3-8b-instruct-lite | requesty | — | $0.1 | $0.1 | $0.1 | 0% | +| together--meta-llama--meta-llama-3.1-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | +| together--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.18 | $0.18 | $0.18 | 0% | +| together--qwen--qwen2.5-72b-instruct-turbo | requesty | — | $1.2 | $1.2 | $1.2 | 0% | +| together--qwen--qwen2.5-7b-instruct-turbo | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| vertex--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | +| vertex--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| vertex--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | +| vertex--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| vertex--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| vertex--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | +| vertex--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | +| vertex--deepseek-v3.2 | requesty | — | $0.56 | $0.056 | $1.68 | 90% | +| vertex--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | +| vertex--gemini-2.5-flash-image | requesty | — | $0.3 | $0.3 | $2.5 | 0% | +| vertex--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | +| vertex--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | +| vertex--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | +| vertex--gemini-3-flash-preview:flex | requesty | — | $0.25 | $0.025 | $0.5 | 90% | +| vertex--gemini-3-pro-image-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| vertex--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| vertex--gemini-3.1-flash-lite | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | +| vertex--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | +| vertex--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | +| vertex--gemini-3.1-pro-preview:flex | requesty | — | $1 | $0.1 | $2.75 | 90% | +| vertex--gemini-3.5-flash | requesty | — | $1.5 | $0.15 | $1.583 | 90% | +| vertex--kimi-k2 | requesty | — | $0.6 | $0.06 | $2.5 | 90% | +| xai--grok-2-1212 | requesty | — | $2 | $2 | $2 | 0% | +| xai--grok-3 | requesty | — | $5 | $5 | $5 | 0% | +| xai--grok-3-mini | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| xai--grok-3-mini:high | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| xai--grok-3-mini:low | requesty | — | $0.3 | $0.3 | $0.3 | 0% | +| xai--grok-4 | requesty | — | $3 | $0.75 | $3 | 75% | +| xai--grok-4-1-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4-1-fast-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4-fast | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | +| xai--grok-4.2-beta | requesty | — | $2 | $0.2 | $2 | 90% | +| xai--grok-4.3 | requesty | — | $1.25 | $0.2 | $1.25 | 84% | +| xai--grok-code-fast-1 | requesty | — | $0.2 | $0.02 | $0.2 | 90% | +| zai--glm-4.5 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | +| zai--glm-4.6 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | +| zai--glm-4.7 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | +| zai--glm-5 | requesty | — | $1 | $0.2 | — | 80% | +| zai--glm-5.1 | requesty | — | $1.4 | $0.26 | $4.4 | 81% | +| deepseek-v3.2 | siliconflow | — | $0.27 | $0.135 | — | 50% | +| deepseek-v4-flash | siliconflow | — | $0.14 | $0.028 | — | 80% | +| deepseek-v4-pro | siliconflow | — | $1.74 | $0.145 | — | 92% | +| glm-4.7 | siliconflow | — | $0.42 | $0.11 | — | 74% | +| glm-5 | siliconflow | — | $0.95 | $0.2 | — | 79% | +| glm-5.1 | siliconflow | — | $1.4 | $0.26 | — | 81% | +| hy3-preview | siliconflow | — | $0.066 | $0.029 | — | 56% | +| kimi-k2.5 | siliconflow | — | $0.45 | $0.07 | — | 84% | +| kimi-k2.6 | siliconflow | — | $0.9 | $0.2 | — | 78% | +| minimax-m2.5 | siliconflow | — | $0.3 | $0.03 | — | 90% | +| step-1-32k | stepfun | — | $15 | $3 | — | 80% | +| step-1-8k | stepfun | — | $5 | $1 | — | 80% | +| step-1o-audio | stepfun | — | $25 | $5 | — | 80% | +| step-1o-turbo-vision | stepfun | — | $2.5 | $0.5 | — | 80% | +| step-1o-vision-32k | stepfun | — | $15 | $3 | — | 80% | +| step-1v-32k | stepfun | — | $15 | $3 | — | 80% | +| step-1v-8k | stepfun | — | $5 | $1 | — | 80% | +| step-2-16k | stepfun | — | $38 | $7.6 | — | 80% | +| step-2-16k-exp | stepfun | — | $38 | $7.6 | — | 80% | +| step-2-mini | stepfun | — | $1 | $0.2 | — | 80% | +| step-3 | stepfun | — | $1.5 | $0.3 | — | 80% | +| step-3.5-flash | stepfun | — | $0.7 | $0.14 | — | 80% | +| step-3.5-flash-2603 | stepfun | — | $0.7 | $0.14 | — | 80% | +| step-audio-2 | stepfun | — | $10 | $2 | — | 80% | +| step-r1-v-mini | stepfun | — | $2.5 | $0.5 | — | 80% | +| stepaudio-2.5-chat | stepfun | — | $10 | $2 | — | 80% | +| stepaudio-2.5-realtime | stepfun | — | $10 | $2 | — | 80% | +| deepseek-v4-flash | tencent-tokenhub | — | $1 | $0.2 | — | 80% | +| deepseek-v4-pro | tencent-tokenhub | — | $12 | $1 | — | 92% | +| glm-5 | tencent-tokenhub | — | $4 | $1 | — | 75% | +| glm-5-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | +| glm-5.1 | tencent-tokenhub | — | $6 | $1.3 | — | 78% | +| glm-5v-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | +| hy3-preview | tencent-tokenhub | — | $1.2 | $0.4 | — | 67% | +| kimi-k2.5 | tencent-tokenhub | — | $4 | $0.7 | — | 82% | +| kimi-k2.6 | tencent-tokenhub | — | $6.5 | $1.1 | — | 83% | +| minimax-m2.5 | tencent-tokenhub | — | $2.1 | $0.21 | — | 90% | +| minimax-m2.7 | tencent-tokenhub | — | $2.1 | $0.42 | — | 80% | +| MiniMaxAI--MiniMax-M2.5 | togetherai | — | $0.3 | $0.06 | — | 80% | +| MiniMaxAI--MiniMax-M2.7 | togetherai | — | $0.3 | $0.06 | — | 80% | +| deepseek-ai--DeepSeek-V4-Pro | togetherai | — | $2.1 | $0.2 | — | 90% | +| moonshotai--Kimi-K2.6 | togetherai | — | $1.2 | $0.2 | — | 83% | +| solar-pro2 | upstage | — | $0.15 | $0.015 | — | 90% | +| solar-pro3 | upstage | — | $0.15 | $0.015 | — | 90% | +| aion-labs-aion-2-0 | venice | — | $1 | $0.25 | — | 75% | +| arcee-trinity-large-thinking | venice | — | $0.3125 | $0.075 | — | 76% | +| claude-opus-4-5 | venice | — | $6 | $0.6 | — | 90% | +| claude-opus-4-6 | venice | — | $6 | $0.6 | — | 90% | +| claude-opus-4-6-fast | venice | — | $36 | $3.6 | — | 90% | +| claude-opus-4-7 | venice | — | $6 | $0.6 | — | 90% | +| claude-opus-4-7-fast | venice | — | $36 | $3.6 | — | 90% | +| claude-sonnet-4-5 | venice | — | $3.75 | $0.375 | — | 90% | +| claude-sonnet-4-6 | venice | — | $3.6 | $0.36 | — | 90% | +| deepseek-v3.2 | venice | — | $0.33 | $0.16 | — | 52% | +| deepseek-v4-flash | venice | — | $0.17 | $0.028 | — | 84% | +| deepseek-v4-pro | venice | — | $1.73 | $0.33 | — | 81% | +| gemini-3-1-pro-preview | venice | — | $2.5 | $0.5 | — | 80% | +| gemini-3-flash-preview | venice | — | $0.7 | $0.07 | — | 90% | +| grok-4-20 | venice | — | $1.42 | $0.23 | — | 84% | +| grok-4-20-multi-agent | venice | — | $1.42 | $0.23 | — | 84% | +| grok-4-3 | venice | — | $1.42 | $0.23 | — | 84% | +| kimi-k2-5 | venice | — | $0.56 | $0.22 | — | 61% | +| kimi-k2-6 | venice | — | $0.85 | $0.22 | — | 74% | +| mercury-2 | venice | — | $0.3125 | $0.03125 | — | 90% | +| minimax-m25 | venice | — | $0.34 | $0.04 | — | 88% | +| minimax-m27 | venice | — | $0.375 | $0.075 | — | 80% | +| openai-gpt-4o-mini-2024-07-18 | venice | — | $0.1875 | $0.09375 | — | 50% | +| openai-gpt-52 | venice | — | $2.19 | $0.219 | — | 90% | +| openai-gpt-52-codex | venice | — | $2.19 | $0.219 | — | 90% | +| openai-gpt-53-codex | venice | — | $2.19 | $0.219 | — | 90% | +| openai-gpt-54 | venice | — | $3.13 | $0.313 | — | 90% | +| openai-gpt-54-mini | venice | — | $0.9375 | $0.09375 | — | 90% | +| openai-gpt-55 | venice | — | $6.25 | $0.625 | — | 90% | +| qwen-3-6-plus | venice | — | $0.625 | $0.0625 | — | 90% | +| qwen3-5-35b-a3b | venice | — | $0.3125 | $0.15625 | — | 50% | +| qwen3-coder-480b-a35b-instruct-turbo | venice | — | $0.35 | $0.04 | — | 89% | +| z-ai-glm-5-turbo | venice | — | $1.2 | $0.24 | — | 80% | +| z-ai-glm-5v-turbo | venice | — | $1.5 | $0.3 | — | 80% | +| zai-org-glm-4.6 | venice | — | $0.85 | $0.3 | — | 65% | +| zai-org-glm-4.7 | venice | — | $0.55 | $0.11 | — | 80% | +| zai-org-glm-5 | venice | — | $1 | $0.2 | — | 80% | +| zai-org-glm-5-1 | venice | — | $1.75 | $0.325 | — | 81% | +| GLM-5.1 | wafer | — | $1.5 | $0.15 | — | 90% | +| Qwen3.5-397B-A17B | wafer | — | $0.6 | $0.06 | — | 90% | + +> 💡 使用[交互式目录](https://i-need-token.github.io/ai-models/)搜索和筛选更多条件的模型。 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 From ddd7b13abc7c0f43f749437e177cbd7ec091e441 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:45:47 +0800 Subject: [PATCH 137/311] Add Cached Pricing links to llms.txt and llms-full.txt --- llms-full.txt | 1 + llms.txt | 1 + 2 files changed, 2 insertions(+) diff --git a/llms-full.txt b/llms-full.txt index e7a20ee0..70bf4d12 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -320,6 +320,7 @@ by_context = sorted( - [Glossary](docs/glossary.md) — key terms and definitions for AI model terminology - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers +- [Cached Pricing](docs/cached-pricing.md) — 75 models with prompt caching, 50-90% input cost savings - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models diff --git a/llms.txt b/llms.txt index a7be3c74..75a3cc2c 100644 --- a/llms.txt +++ b/llms.txt @@ -75,6 +75,7 @@ modalities: - [Glossary](docs/glossary.md) — key terms and definitions for AI model terminology - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers +- [Cached Pricing](docs/cached-pricing.md) — 75 models with prompt caching, 50-90% input cost savings - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models From 828185a5fca34eb137d3f2602de6528a8054d7cb Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:48:24 +0800 Subject: [PATCH 138/311] Add Video Models and Cached Pricing links to README documentation table (EN + ZH) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Video Models: 167 video input + 4 video output models - Cached Pricing: 1,353 models with prompt caching — 50-90% input cost savings - Both EN and ZH documentation tables now cover all 23 docs --- README.md | 48 ++++++++++++++++++++++++++---------------------- 1 file changed, 26 insertions(+), 22 deletions(-) diff --git a/README.md b/README.md index ecece18c..9eaea776 100644 --- a/README.md +++ b/README.md @@ -411,6 +411,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | | [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | | [Audio Models](docs/audio-models.md) | 118 audio input + 34 audio output models | +| [Video Models](docs/video-models.md) | 167 video input + 4 video output models | | [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | @@ -420,6 +421,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Glossary](docs/glossary.md) | Key terms and definitions for AI model terminology | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | +| [Cached Pricing](docs/cached-pricing.md) | 1,353 models with prompt caching — 50-90% input cost savings | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | | [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | @@ -432,28 +434,30 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin **中文文档：** -| 文档 | 描述 | -| -------------------------------------------- | ----------------------------------------------- | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | -| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| 文档 | 描述 | +| -------------------------------------------- | ------------------------------------------------ | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | +| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | +| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [缓存定价](docs/zh/cached-pricing.md) | 1,353 个支持提示缓存的模型 — 输入成本节省 50-90% | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles From 8cdfb660d78ea7460c957b4b86eb840f7bba469a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:48:59 +0800 Subject: [PATCH 139/311] Fix cached pricing model count: 75 (not 1,353) --- README.md | 50 +++++++++++++++++++++++++------------------------- 1 file changed, 25 insertions(+), 25 deletions(-) diff --git a/README.md b/README.md index 9eaea776..d6cfd93c 100644 --- a/README.md +++ b/README.md @@ -421,7 +421,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Glossary](docs/glossary.md) | Key terms and definitions for AI model terminology | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | -| [Cached Pricing](docs/cached-pricing.md) | 1,353 models with prompt caching — 50-90% input cost savings | +| [Cached Pricing](docs/cached-pricing.md) | 75 models with prompt caching — 50-90% input cost savings | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | | [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | @@ -434,30 +434,30 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin **中文文档：** -| 文档 | 描述 | -| -------------------------------------------- | ------------------------------------------------ | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | -| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | -| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [缓存定价](docs/zh/cached-pricing.md) | 1,353 个支持提示缓存的模型 — 输入成本节省 50-90% | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| 文档 | 描述 | +| -------------------------------------------- | ----------------------------------------------- | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | +| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | +| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [缓存定价](docs/zh/cached-pricing.md) | 75 个支持提示缓存的模型 — 输入成本节省 50-90% | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles From b0d319ffa4933a77f755af1149a1d97b9beec49c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:51:59 +0800 Subject: [PATCH 140/311] Fix missing doc links in AGENTS.md, llms.txt, llms-full.txt - Add cached-pricing.md to AGENTS.md - Add video-models.md to llms.txt and llms-full.txt - All 23 docs now covered in all index files --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index 8bb9a3c8..ac5cd37a 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -24,6 +24,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) - [`docs/model-comparison.md`](docs/model-comparison.md) — Model comparison tables ([中文](docs/zh/model-comparison.md)) - [`docs/pricing-comparison.md`](docs/pricing-comparison.md) — Pricing comparison across providers ([中文](docs/zh/pricing-comparison.md)) +- [`docs/cached-pricing.md`](docs/cached-pricing.md) — Models with prompt caching, 50-90% input cost savings ([中文](docs/zh/cached-pricing.md)) - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) - [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) - [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) From 2e1fd67a69c6d6aff67f10607d893a888b174a28 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 15:57:30 +0800 Subject: [PATCH 141/311] =?UTF-8?q?Fix=20cached=20pricing=20model=20count:?= =?UTF-8?q?=2075=20=E2=86=92=201,374=20across=20README,=20llms.txt,=20llms?= =?UTF-8?q?-full.txt?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 50 +- docs/cached-pricing.md | 1580 +++++-------------------------------- docs/zh/cached-pricing.md | 1580 +++++-------------------------------- llms-full.txt | 2 +- llms.txt | 2 +- 5 files changed, 433 insertions(+), 2781 deletions(-) diff --git a/README.md b/README.md index d6cfd93c..08503a5f 100644 --- a/README.md +++ b/README.md @@ -421,7 +421,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Glossary](docs/glossary.md) | Key terms and definitions for AI model terminology | | [Model Comparison](docs/model-comparison.md) | Compare flagship, cost-effective, free, and open-weight models | | [Pricing Comparison](docs/pricing-comparison.md) | Side-by-side pricing across providers and platforms | -| [Cached Pricing](docs/cached-pricing.md) | 75 models with prompt caching — 50-90% input cost savings | +| [Cached Pricing](docs/cached-pricing.md) | 1,374 models with prompt caching — 50-90% input cost savings | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | | [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | @@ -434,30 +434,30 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin **中文文档：** -| 文档 | 描述 | -| -------------------------------------------- | ----------------------------------------------- | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | -| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | -| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [缓存定价](docs/zh/cached-pricing.md) | 75 个支持提示缓存的模型 — 输入成本节省 50-90% | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| 文档 | 描述 | +| -------------------------------------------- | ------------------------------------------------ | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | +| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | +| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles diff --git a/docs/cached-pricing.md b/docs/cached-pricing.md index 8cd3a6e1..0714acab 100644 --- a/docs/cached-pricing.md +++ b/docs/cached-pricing.md @@ -27,1384 +27,210 @@ Prompt caching lets you store repeated prompt prefixes (system prompts, few-shot ## Model Pricing -| Model | Provider | Context | Input $/M | Cache Read $/M | Cache Write $/M | Savings | -| ---------------------------------------------------------- | ---------------- | ------- | ------------------- | --------------------- | --------------- | ------- | -| aistudio_gemini-2.0-flash | aihubmix | — | $0.05 | $0.125 | — | -150% | -| aistudio_gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | -| anthropic-opus-4-6 | aihubmix | — | $2.5 | $0.25 | $3.125 | 90% | -| claude-haiku-4-5 | aihubmix | — | $0.55 | $0.055 | $0.6875 | 90% | -| claude-sonnet-4-0 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | -| claude-sonnet-4-5 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | -| claude-sonnet-4-5-think | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | -| codex-mini-latest | aihubmix | — | $0.75 | $0.1875 | — | 75% | -| deepseek-v3.2 | aihubmix | — | $0.151 | $0.0151 | — | 90% | -| deepseek-v3.2-exp | aihubmix | — | $0.137 | $0.0137 | — | 90% | -| deepseek-v3.2-exp-think | aihubmix | — | $0.137 | $0.0137 | — | 90% | -| deepseek-v3.2-think | aihubmix | — | $0.151 | $0.0151 | — | 90% | -| doubao-1.5-lite-32k | aihubmix | — | $0.025 | $0.005 | — | 80% | -| doubao-1.5-pro-32k | aihubmix | — | $0.067 | $0.0134 | — | 80% | -| doubao-lite-32k | aihubmix | — | $0.03 | $0.006 | — | 80% | -| doubao-pro-32k | aihubmix | — | $0.07 | $0.014 | — | 80% | -| doubao-seed-1-6 | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-flash | aihubmix | — | $0.022 | $0.0044 | — | 80% | -| doubao-seed-1-6-flash-250615 | aihubmix | — | $0.022 | $0.0044 | — | 80% | -| doubao-seed-1-6-lite | aihubmix | — | $0.041 | $0.0082 | — | 80% | -| doubao-seed-1-6-thinking | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-thinking-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-vision-250815 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | -| doubao-seed-1-8 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | -| gemini-2.0-flash | aihubmix | — | $0.05 | $0.0125 | — | 75% | -| gemini-2.0-flash-001 | aihubmix | — | $0.05 | $0.125 | — | -150% | -| gemini-2.0-flash-search | aihubmix | — | $0.05 | $0.0125 | — | 75% | -| gemini-2.5-flash | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-lite | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-search | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-pro | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| glm-4.5-airx | aihubmix | — | $0.55 | $0.11 | — | 80% | -| glm-4.5-x | aihubmix | — | $1.1 | $0.22 | — | 80% | -| glm-4.6 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | -| glm-4.6v | aihubmix | — | $0.0685 | $0.0137 | — | 80% | -| glm-4.7 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | -| gpt-4.1 | aihubmix | — | $1 | $0.25 | — | 75% | -| gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | -| gpt-4.1-nano | aihubmix | — | $0.05 | $0.0125 | — | 75% | -| gpt-4o | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-2024-08-06 | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-2024-08-06-global | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-2024-11-20 | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-mini | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-mini-2024-07-18 | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-mini-global | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-mini-search-preview | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-search-preview | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-5 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | -| gpt-5-nano | aihubmix | — | $0.025 | $0.0025 | — | 90% | -| gpt-5.1 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-codex-max | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-codex-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | -| gpt-5.2 | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-chat-latest | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-codex | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-high | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-low | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-pro | aihubmix | — | $10.5 | $1.05 | — | 90% | -| grok-4 | aihubmix | — | $1.65 | $0.4125 | — | 75% | -| grok-4-1-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4-1-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4.20-beta-0309-non-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-4.20-beta-0309-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-4.20-multi-agent-0309 | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-4.20-multi-agent-beta-0309 | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-code-fast-1 | aihubmix | — | $0.1 | $0.025 | — | 75% | -| kimi-k2-thinking | aihubmix | — | $0.274 | $0.0685 | — | 75% | -| kimi-k2-turbo-preview | aihubmix | — | $0.6 | $0.15 | — | 75% | -| kimi-k2.5 | aihubmix | — | $0.3 | $0.0525 | — | 82% | -| mimo-v2-flash | aihubmix | — | $0.0959 | $0.01918 | — | 80% | -| mimo-v2-omni | aihubmix | — | $0.22 | $0.044 | — | 80% | -| mimo-v2-pro | aihubmix | — | $0.55 | $0.11 | — | 80% | -| nvidia-nemotron-3-super-120b-a12b | aihubmix | — | $0.055 | $0.01375 | — | 75% | -| o1 | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-2024-12-17 | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-global | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-mini | aihubmix | — | $1.5 | $0.75 | — | 50% | -| o1-mini-2024-09-12 | aihubmix | — | $1.5 | $0.75 | — | 50% | -| o1-preview | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-preview-2024-09-12 | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o3 | aihubmix | — | $1 | $0.25 | — | 75% | -| o3-deep-research | aihubmix | — | $5 | $1.25 | — | 75% | -| o3-global | aihubmix | — | $1 | $0.25 | — | 75% | -| o3-mini | aihubmix | — | $0.55 | $0.275 | — | 50% | -| o3-mini-global | aihubmix | — | $0.55 | $0.275 | — | 50% | -| o4-mini | aihubmix | — | $0.55 | $0.1375 | — | 75% | -| qwen-plus | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-plus-2025-04-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-plus-2025-07-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-plus-latest | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-turbo | aihubmix | — | $0.023 | $0.0046 | — | 80% | -| qwen-turbo-latest | aihubmix | — | $0.023 | $0.0046 | — | 80% | -| qwen3-coder-plus | aihubmix | — | $0.27 | $0.054 | — | 80% | -| qwen3-max | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | -| qwen3-max-2026-01-23 | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | -| qwen3-max-preview | aihubmix | — | $0.423 | $0.0846 | — | 80% | -| qwen3-vl-flash | aihubmix | — | $0.0103 | $0.00206 | — | 80% | -| qwen3-vl-plus | aihubmix | — | $0.0685 | $0.0137 | — | 80% | -| zai-glm-5-turbo | aihubmix | — | $0.6 | $0.12 | — | 80% | -| aion-2.0 | aion | — | $0.7999999999999999 | $0.19999999999999998 | — | 75% | -| aion-2.5 | aion | — | $1 | $0.35 | — | 65% | -| amazon-nova-2-lite | amazon-bedrock | — | $0.33 | $0.0825 | — | 75% | -| amazon-nova-lite | amazon-bedrock | — | $0.06 | $0.015 | — | 75% | -| amazon-nova-micro | amazon-bedrock | — | $0.035 | $0.00875 | — | 75% | -| amazon-nova-premier | amazon-bedrock | — | $2.5 | $0.625 | — | 75% | -| amazon-nova-pro | amazon-bedrock | — | $0.8 | $0.2 | — | 75% | -| claude-haiku-4-5-20251001 | auriko | — | $1 | $0.1 | $1.25 | 90% | -| claude-opus-4-1-20250805 | auriko | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-20250514 | auriko | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-5-20251101 | auriko | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-6 | auriko | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-7 | auriko | — | $5 | $0.5 | $6.25 | 90% | -| claude-sonnet-4-20250514 | auriko | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-5-20250929 | auriko | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-6 | auriko | — | $3 | $0.3 | $3.75 | 90% | -| deepseek-r1-0528 | auriko | — | $0.5 | $0.35 | — | 30% | -| deepseek-v3-0324 | auriko | — | $0.2 | $0.135 | — | 32% | -| deepseek-v3.1 | auriko | — | $0.21 | $0.13 | — | 38% | -| deepseek-v3.1-terminus | auriko | — | $0.27 | $0.13 | — | 52% | -| deepseek-v3.2 | auriko | — | $0.26 | $0.13 | — | 50% | -| deepseek-v4-flash | auriko | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-pro | auriko | — | $0.435 | $0.003625 | — | 99% | -| gemini-2.5-flash | auriko | — | $0.3 | $0.03 | — | 90% | -| gemini-2.5-flash-lite | auriko | — | $0.1 | $0.01 | — | 90% | -| gemini-2.5-pro | auriko | — | $1.25 | $0.125 | — | 90% | -| gemini-3-flash-preview | auriko | — | $0.5 | $0.05 | — | 90% | -| gemini-3.1-flash-lite | auriko | — | $0.25 | $0.025 | — | 90% | -| gemini-3.1-flash-lite-preview | auriko | — | $0.25 | $0.025 | — | 90% | -| gemini-3.1-pro-preview | auriko | — | $2 | $0.2 | — | 90% | -| gemini-3.1-pro-preview-customtools | auriko | — | $2 | $0.2 | — | 90% | -| gemini-flash-latest | auriko | — | $0.5 | $0.05 | — | 90% | -| gemini-flash-lite-latest | auriko | — | $0.1 | $0.01 | — | 90% | -| gemini-pro-latest | auriko | — | $2 | $0.2 | — | 90% | -| glm-4.5 | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.5-air | auriko | — | $0.2 | $0.03 | — | 85% | -| glm-4.5-airx | auriko | — | $1.1 | $0.22 | — | 80% | -| glm-4.5-x | auriko | — | $2.2 | $0.45 | — | 80% | -| glm-4.5v | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.6 | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.6v | auriko | — | $0.3 | $0.05 | — | 83% | -| glm-4.6v-flashx | auriko | — | $0.04 | $0.004 | — | 90% | -| glm-4.7 | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.7-flashx | auriko | — | $0.07 | $0.01 | — | 86% | -| glm-5 | auriko | — | $1 | $0.2 | — | 80% | -| glm-5-turbo | auriko | — | $1.2 | $0.24 | — | 80% | -| glm-5.1 | auriko | — | $1.4 | $0.26 | — | 81% | -| glm-5v-turbo | auriko | — | $1.2 | $0.24 | — | 80% | -| gpt-4.1-2025-04-14 | auriko | — | $2 | $0.5 | — | 75% | -| gpt-4.1-mini-2025-04-14 | auriko | — | $0.4 | $0.1 | — | 75% | -| gpt-4.1-nano-2025-04-14 | auriko | — | $0.1 | $0.025 | — | 75% | -| gpt-4o-2024-08-06 | auriko | — | $2.5 | $1.25 | — | 50% | -| gpt-4o-2024-11-20 | auriko | — | $2.5 | $1.25 | — | 50% | -| gpt-4o-mini-2024-07-18 | auriko | — | $0.15 | $0.075 | — | 50% | -| gpt-5-2025-08-07 | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5-mini-2025-08-07 | auriko | — | $0.25 | $0.025 | — | 90% | -| gpt-5-nano-2025-08-07 | auriko | — | $0.05 | $0.005 | — | 90% | -| gpt-5.1-2025-11-13 | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5.1-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5.2-2025-12-11 | auriko | — | $1.75 | $0.175 | — | 90% | -| gpt-5.2-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | -| gpt-5.3-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | -| gpt-5.4-2026-03-05 | auriko | — | $2.5 | $0.25 | — | 90% | -| gpt-5.4-mini-2026-03-17 | auriko | — | $0.75 | $0.075 | — | 90% | -| gpt-5.4-nano-2026-03-17 | auriko | — | $0.2 | $0.02 | — | 90% | -| gpt-5.5-2026-04-23 | auriko | — | $5 | $0.5 | — | 90% | -| gpt-oss-120b | auriko | — | $0.15 | $0.01 | — | 93% | -| gpt-oss-20b | auriko | — | $0.07 | $0.04 | — | 43% | -| grok-4.20-0309-non-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | -| grok-4.20-0309-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | -| grok-4.3 | auriko | — | $1.25 | $0.2 | — | 84% | -| hy3-preview | auriko | — | $0.066 | $0.029 | — | 56% | -| kimi-k2-0711-preview | auriko | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-0905-preview | auriko | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-thinking | auriko | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-thinking-turbo | auriko | — | $1.15 | $0.15 | — | 87% | -| kimi-k2-turbo-preview | auriko | — | $1.15 | $0.15 | — | 87% | -| kimi-k2.5 | auriko | — | $0.6 | $0.1 | — | 83% | -| kimi-k2.6 | auriko | — | $0.95 | $0.16 | — | 83% | -| mimo-v2.5 | auriko | — | $0.4 | $0.08 | — | 80% | -| mimo-v2.5-pro | auriko | — | $1 | $0.2 | — | 80% | -| minimax-m2 | auriko | — | $0.3 | — | $0.375 | — | -| minimax-m2-1 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | -| minimax-m2-1-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | -| minimax-m2-5 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | -| minimax-m2-5-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | -| minimax-m2-7 | auriko | — | $0.3 | — | $0.375 | — | -| minimax-m2-7-highspeed | auriko | — | $0.6 | — | $0.375 | — | -| o3-2025-04-16 | auriko | — | $2 | $0.5 | — | 75% | -| o3-mini-2025-01-31 | auriko | — | $1.1 | $0.55 | — | 50% | -| o4-mini-2025-04-16 | auriko | — | $1.1 | $0.275 | — | 75% | -| qwen-3-coder-480b-a35b-instruct | auriko | — | $0.3 | $0.1 | — | 67% | -| qwen-3-max | auriko | — | $1.2 | $0.24 | — | 80% | -| qwen-3-max-thinking | auriko | — | $1.2 | $0.24 | — | 80% | -| qwen-3-vl-235b-a22b-instruct | auriko | — | $0.2 | $0.11 | — | 45% | -| qwen-3.5-35b-a3b | auriko | — | $0.14 | $0.05 | — | 64% | -| seed-1.8 | auriko | — | $0.25 | $0.05 | — | 80% | -| seed-2.0-code | auriko | — | $0.5 | $0.1 | — | 80% | -| seed-2.0-mini | auriko | — | $0.1 | $0.02 | — | 80% | -| seed-2.0-pro | auriko | — | $0.5 | $0.1 | — | 80% | -| step-3.5-flash | auriko | — | $0.1 | $0.02 | — | 80% | -| deepseek-v3.2 | baidu | — | $0.252 | $0.0252 | — | 90% | -| deepseek-v4-flash | baidu | — | $0.126 | $0.0252 | — | 80% | -| deepseek-v4-pro | baidu | — | $1.521 | $0.126 | — | 92% | -| glm-5 | baidu | — | $0.7 | $0.14 | — | 80% | -| glm-5.1 | baidu | — | $0.98 | $0.182 | — | 81% | -| minimax-m2.5 | baidu | — | $0.27 | $0.027 | — | 90% | -| deepseek-v3-1 | baseten | — | $0.5 | $0.25 | — | 50% | -| deepseek-v4 | baseten | — | $1.74 | $0.145 | — | 92% | -| glm-4-7 | baseten | — | $0.6 | $0.12 | — | 80% | -| glm-5 | baseten | — | $0.95 | $0.2 | — | 79% | -| kimi-k2-5 | baseten | — | $0.6 | $0.12 | — | 80% | -| kimi-k2-6 | baseten | — | $1 | $0.2 | — | 80% | -| minimax-m2-5 | baseten | — | $0.3 | $0.06 | — | 80% | -| nvidia-nemotron-3-super | baseten | — | $0.3 | $0.06 | — | 80% | -| MiniMaxAI--MiniMax-M2.5-TEE | chutes | — | $0.15 | $0.075 | — | 50% | -| Qwen--Qwen3-235B-A22B-Thinking-2507 | chutes | — | $0.11 | $0.055 | — | 50% | -| Qwen--Qwen3-32B-TEE | chutes | — | $0.08 | $0.04 | — | 50% | -| Qwen--Qwen3.5-397B-A17B-TEE | chutes | — | $0.39 | $0.195 | — | 50% | -| Qwen--Qwen3.6-27B-TEE | chutes | — | $0.5 | $0.25 | — | 50% | -| deepseek-ai--DeepSeek-V3.2-TEE | chutes | — | $0.28 | $0.14 | — | 50% | -| google--gemma-4-31B-turbo-TEE | chutes | — | $0.13 | $0.065 | — | 50% | -| moonshotai--Kimi-K2.5-TEE | chutes | — | $0.44 | $0.22 | — | 50% | -| moonshotai--Kimi-K2.6-TEE | chutes | — | $0.74 | $0.37 | — | 50% | -| zai-org--GLM-5-TEE | chutes | — | $0.95 | $0.475 | — | 50% | -| zai-org--GLM-5-Turbo | chutes | — | $0.4891 | $0.24455 | — | 50% | -| zai-org--GLM-5.1-TEE | chutes | — | $1.05 | $0.525 | — | 50% | -| kimi-k2-thinking | clarifai | — | $1.5 | $0.15 | — | 90% | -| kimi-k2.5 | cloudflare | — | $0.6 | $0.1 | — | 83% | -| kimi-k2.6 | cloudflare | — | $0.95 | $0.16 | — | 83% | -| claude-4-5-sonnet | cortecs | — | $2.683 | $0.293 | — | 89% | -| claude-4-6-sonnet | cortecs | — | $2.869 | $0.287 | — | 90% | -| claude-haiku-4-5 | cortecs | — | $0.894 | $0.089 | — | 90% | -| claude-opus4-5 | cortecs | — | $4.769 | $0.477 | — | 90% | -| claude-opus4-6 | cortecs | — | $4.769 | $0.477 | — | 90% | -| claude-opus4-7 | cortecs | — | $4.769 | $0.477 | — | 90% | -| claude-sonnet-4 | cortecs | — | $2.601 | $0.26 | — | 90% | -| codestral-2508 | cortecs | — | $0.3 | $0.03 | — | 90% | -| deepseek-chat-v3.1 | cortecs | — | $0.177 | $0.044 | — | 75% | -| deepseek-r1-0528 | cortecs | — | $0.585 | $0.146 | — | 75% | -| deepseek-v3-0324 | cortecs | — | $0.266 | $0.067 | — | 75% | -| deepseek-v3.2 | cortecs | — | $0.266 | $0.067 | — | 75% | -| deepseek-v4-flash | cortecs | — | $0.133 | $0.033 | — | 75% | -| deepseek-v4-pro | cortecs | — | $1.553 | $0.388 | — | 75% | -| devstral-2512 | cortecs | — | $0.4 | $0.04 | — | 90% | -| devstral-medium-2507 | cortecs | — | $0.4 | $0.04 | — | 90% | -| devstral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | -| gemini-2.5-flash | cortecs | — | $0.268 | $0.026 | — | 90% | -| gemini-2.5-pro | cortecs | — | $1.342 | $0.217 | — | 84% | -| gemini-3.1-flash-lite | cortecs | — | $0.244 | $0.022 | — | 91% | -| glm-4.6 | cortecs | — | $0.355 | $0.089 | — | 75% | -| glm-4.7 | cortecs | — | $0.532 | $0.133 | — | 75% | -| glm-5 | cortecs | — | $0.887 | $0.222 | — | 75% | -| glm-5-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | -| glm-5.1 | cortecs | — | $1.242 | $0.311 | — | 75% | -| glm-5v-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | -| gpt-4.1 | cortecs | — | $1.968 | $0.49 | — | 75% | -| gpt-4.1-mini | cortecs | — | $0.39 | $0.12 | — | 69% | -| gpt-4.1-nano | cortecs | — | $0.1 | $0.05 | — | 50% | -| gpt-4o | cortecs | — | $2.387 | $1.194 | — | 50% | -| gpt-4o-mini | cortecs | — | $0.143 | $0.073 | — | 49% | -| gpt-5 | cortecs | — | $1.234 | $0.14 | — | 89% | -| gpt-5-mini | cortecs | — | $0.25 | $0.05 | — | 80% | -| gpt-5-nano | cortecs | — | $0.054 | $0.017 | — | 69% | -| gpt-5.1 | cortecs | — | $1.234 | $0.14 | — | 89% | -| gpt-5.4 | cortecs | — | $2.601 | $0.217 | — | 92% | -| gpt-oss-120b | cortecs | — | $0.035 | $0.009 | — | 74% | -| gpt-oss-20b | cortecs | — | $0.027 | $0.007 | — | 74% | -| kimi-k2.5 | cortecs | — | $0.444 | $0.111 | — | 75% | -| kimi-k2.6 | cortecs | — | $0.694 | $0.173 | — | 75% | -| llama-3.3-70b-instruct | cortecs | — | $0.089 | $0.023 | — | 74% | -| llama-4-maverick | cortecs | — | $0.124 | $0.03 | — | 76% | -| magistral-medium-2509 | cortecs | — | $2 | $0.2 | — | 90% | -| magistral-small-2509 | cortecs | — | $0.5 | $0.05 | — | 90% | -| minimax-m2 | cortecs | — | $0.222 | $0.055 | — | 75% | -| minimax-m2.5 | cortecs | — | $0.266 | $0.027 | — | 90% | -| minimax-m2.7 | cortecs | — | $0.444 | $0.111 | — | 75% | -| ministral-14b-2512 | cortecs | — | $0.2 | $0.02 | — | 90% | -| ministral-3b-2512 | cortecs | — | $0.1 | $0.01 | — | 90% | -| ministral-8b-2512 | cortecs | — | $0.15 | $0.015 | — | 90% | -| mistral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | -| mistral-large-2512 | cortecs | — | $0.5 | $0.05 | — | 90% | -| mistral-medium-2508 | cortecs | — | $0.4 | $0.04 | — | 90% | -| mistral-medium-3.5 | cortecs | — | $1.25 | $0.125 | — | 90% | -| mistral-nemo-instruct-2407 | cortecs | — | $0.13 | $0.013 | — | 90% | -| mistral-small-2506 | cortecs | — | $0.1 | $0.01 | — | 90% | -| mistral-small-2603 | cortecs | — | $0.128 | $0.013 | — | 90% | -| nemotron-3-super-120b-a12b | cortecs | — | $0.266 | $0.067 | — | 75% | -| pixtral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | -| qwen-2.5-72b-instruct | cortecs | — | $0.062 | $0.016 | — | 74% | -| qwen3-235b-a22b-instruct-2507 | cortecs | — | $0.062 | $0.016 | — | 74% | -| qwen3-coder-30b-a3b-instruct | cortecs | — | $0.053 | $0.013 | — | 75% | -| qwen3-vl-235b-a22b | cortecs | — | $0.186 | $0.047 | — | 75% | -| qwen3.5-122b-a10b | cortecs | — | $0.444 | $0.111 | — | 75% | -| voxtral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | -| databricks-claude-haiku-4-5 | databricks | — | $1 | $0.1 | $1.25 | 90% | -| databricks-claude-opus-4-1 | databricks | — | $15 | $1.5 | $18.75 | 90% | -| databricks-claude-opus-4-5 | databricks | — | $5 | $0.5 | $6.25 | 90% | -| databricks-claude-sonnet-4 | databricks | — | $3 | $0.3 | $3.75 | 90% | -| databricks-claude-sonnet-4-5 | databricks | — | $3 | $0.3 | $3.75 | 90% | -| databricks-gemini-3-1-flash-lite | databricks | — | $0.25 | $0.02 | $0.25 | 92% | -| databricks-gemini-3-1-pro | databricks | — | $2.5 | $0.25 | $2.5 | 90% | -| databricks-gemini-3-flash | databricks | — | $0.63 | $0.06 | $0.63 | 90% | -| databricks-gpt-5 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | -| databricks-gpt-5-1 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | -| databricks-gpt-5-1-codex-max | databricks | — | $1.25 | $0.13 | $1.25 | 90% | -| databricks-gpt-5-1-codex-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | -| databricks-gpt-5-2 | databricks | — | $1.75 | $0.18 | $1.75 | 90% | -| databricks-gpt-5-2-codex | databricks | — | $1.75 | $0.18 | $1.75 | 90% | -| databricks-gpt-5-4 | databricks | — | $2.5 | $0.25 | $2.5 | 90% | -| databricks-gpt-5-4-mini | databricks | — | $0.75 | $0.07 | $0.75 | 91% | -| databricks-gpt-5-4-nano | databricks | — | $0.2 | $0.02 | $0.2 | 90% | -| databricks-gpt-5-5 | databricks | — | $5 | $0.5 | $5 | 90% | -| databricks-gpt-5-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | -| databricks-gpt-5-nano | databricks | — | $0.05 | Free | $0.05 | — | -| deepseek-r1-0528 | deepinfra | — | $0.5 | $0.35 | — | 30% | -| deepseek-v3-0324 | deepinfra | — | $0.2 | $0.135 | — | 32% | -| deepseek-v3.1 | deepinfra | — | $0.21 | $0.13 | — | 38% | -| deepseek-v3.1-terminus | deepinfra | — | $0.27 | $0.13 | — | 52% | -| deepseek-v3.2 | deepinfra | — | $0.26 | $0.13 | — | 50% | -| deepseek-v4-flash | deepinfra | — | $0.14 | $0.028 | — | 80% | -| deepseek-v4-pro | deepinfra | — | $1.74 | $0.145 | — | 92% | -| glm-4.6 | deepinfra | — | $0.43 | $0.08 | — | 81% | -| glm-4.7 | deepinfra | — | $0.4 | $0.08 | — | 80% | -| glm-4.7-flash | deepinfra | — | $0.06 | $0.01 | — | 83% | -| glm-5 | deepinfra | — | $0.6 | $0.12 | — | 80% | -| glm-5.1 | deepinfra | — | $1.05 | $0.205 | — | 80% | -| kimi-k2.5 | deepinfra | — | $0.45 | $0.07 | — | 84% | -| kimi-k2.6 | deepinfra | — | $0.75 | $0.15 | — | 80% | -| mimo-v2.5 | deepinfra | — | $0.4 | $0.08 | — | 80% | -| mimo-v2.5-pro | deepinfra | — | $1 | $0.2 | — | 80% | -| minimax-m2.5 | deepinfra | — | $0.15 | $0.03 | — | 80% | -| qwen3-235b-a22b-thinking-2507 | deepinfra | — | $0.23 | $0.2 | — | 13% | -| qwen3-coder-480b-a35b-instruct-turbo | deepinfra | — | $0.3 | $0.1 | — | 67% | -| qwen3-max | deepinfra | — | $1.2 | $0.24 | — | 80% | -| qwen3-max-thinking | deepinfra | — | $1.2 | $0.24 | — | 80% | -| qwen3-vl-235b-a22b-instruct | deepinfra | — | $0.2 | $0.11 | — | 45% | -| qwen3.5-35b-a3b | deepinfra | — | $0.14 | $0.05 | — | 64% | -| qwen3.5-397b-a17b | deepinfra | — | $0.49 | $0.3 | — | 39% | -| seed-1.8 | deepinfra | — | $0.25 | $0.05 | — | 80% | -| seed-2.0-code | deepinfra | — | $0.5 | $0.1 | — | 80% | -| seed-2.0-mini | deepinfra | — | $0.1 | $0.02 | — | 80% | -| seed-2.0-pro | deepinfra | — | $0.5 | $0.1 | — | 80% | -| step-3.5-flash | deepinfra | — | $0.1 | $0.02 | — | 80% | -| deepseek-chat | deepseek | — | $0.14 | $0.0028 | — | 98% | -| deepseek-reasoner | deepseek | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-flash | deepseek | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-pro | deepseek | — | $0.435 | $0.003625 | — | 99% | -| arcee-trinity-large-thinking | digitalocean | — | $0.25 | $0.06 | — | 76% | -| anthropic--claude-3-5-haiku-20241022 | fastrouter | — | $0.8 | $0.08 | $1 | 90% | -| anthropic--claude-4.5-sonnet | fastrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-haiku-4.5 | fastrouter | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4-20250514 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.1 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.5 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.6 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.7 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-sonnet-4-20250514 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4.6 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | -| deepseek--deepseek-v3.2 | fastrouter | — | $0.27 | $0.216 | — | 20% | -| deepseek--deepseek-v4-pro | fastrouter | — | $2.1 | $0.2 | — | 90% | -| google--gemini-2.0-flash-001 | fastrouter | — | $0.1 | $0.025 | $0.1833 | 75% | -| google--gemini-2.5-flash | fastrouter | — | $0.3 | $0.075 | $0.3833 | 75% | -| google--gemini-2.5-pro | fastrouter | — | $1.25 | $0.31 | $1.625 | 75% | -| google--gemini-3-flash-preview | fastrouter | — | $0.5 | $0.05 | — | 90% | -| google--gemini-3.1-flash-lite-preview | fastrouter | — | $0.25 | $0.025 | $0.083333 | 90% | -| google--gemini-3.1-pro-preview | fastrouter | — | $2 | $0.31 | $1.625 | 84% | -| minimax--minimax-m2.1 | fastrouter | — | $0.27 | $0.03 | — | 89% | -| minimax--minimax-m2.5 | fastrouter | — | $0.3 | $0.03 | — | 90% | -| minimax--minimax-m2.5-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | -| minimax--minimax-m2.7 | fastrouter | — | $0.3 | $0.06 | — | 80% | -| minimax--minimax-m2.7-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | -| moonshotai--kimi-k2-thinking | fastrouter | — | $0.6 | $0.15 | — | 75% | -| moonshotai--kimi-k2.6 | fastrouter | — | $0.55 | $0.15 | — | 73% | -| openai--gpt-4.1 | fastrouter | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-mini | fastrouter | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-nano | fastrouter | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4o | fastrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-08-06 | fastrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-11-20 | fastrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-mini | fastrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-mini-2024-07-18 | fastrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-realtime-preview-2024-12-17 | fastrouter | — | $5 | $2.5 | — | 50% | -| openai--gpt-4o-realtime-preview-2025-06-03 | fastrouter | — | $5 | $2.5 | — | 50% | -| openai--gpt-5 | fastrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-mini | fastrouter | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-nano | fastrouter | — | $0.05 | $0.005 | — | 90% | -| openai--gpt-5.1 | fastrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.2 | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.4 | fastrouter | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | fastrouter | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | fastrouter | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | fastrouter | — | $5 | $0.5 | — | 90% | -| openai--gpt-realtime | fastrouter | — | $4 | $0.5 | — | 88% | -| openai--gpt-realtime-1.5 | fastrouter | — | $4 | $0.4 | — | 90% | -| openai--gpt-realtime-mini | fastrouter | — | $0.6 | $0.06 | — | 90% | -| openai--o1 | fastrouter | — | $15 | $7.5 | — | 50% | -| openai--o3 | fastrouter | — | $2 | $0.5 | — | 75% | -| openai--o3-mini | fastrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o3-mini-high | fastrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o4-mini | fastrouter | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-high | fastrouter | — | $1.1 | $0.275 | — | 75% | -| x-ai--grok-3-beta | fastrouter | — | $5 | $1.25 | — | 75% | -| x-ai--grok-3-mini-beta | fastrouter | — | $0.6 | $0.15 | — | 75% | -| x-ai--grok-4 | fastrouter | — | $3 | $0.75 | — | 75% | -| x-ai--grok-4.20-beta | fastrouter | — | $2 | $0.2 | — | 90% | -| x-ai--grok-4.3 | fastrouter | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-code-fast-1 | fastrouter | — | $0.2 | $0.02 | — | 90% | -| z-ai--glm-4.5-air | fastrouter | — | $0.2 | $0.03 | — | 85% | -| LGAI-EXAONE--K-EXAONE-236B-A23B | friendli | — | $0.2 | $0.1 | — | 50% | -| MiniMaxAI--MiniMax-M2.5 | friendli | — | $0.3 | $0.06 | — | 80% | -| deepseek-ai--DeepSeek-V3.2 | friendli | — | $0.5 | $0.25 | — | 50% | -| zai-org--GLM-5 | friendli | — | $1 | $0.5 | — | 50% | -| zai-org--GLM-5.1 | friendli | — | $1.4 | $0.26 | — | 81% | -| gemini-1.5-flash | google | — | $0.075 | $0.01875 | $0.2625 | 75% | -| gemini-1.5-flash-8b | google | — | $0.075 | $0.01875 | $0.2625 | 75% | -| gemini-1.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | -| gemini-2.0-flash | google | — | $0.1 | $0.025 | $0.35 | 75% | -| gemini-2.0-flash-lite | google | — | $0.075 | $0.01875 | $0.2625 | 75% | -| gemini-2.5-flash | google | — | $0.15 | $0.0375 | $0.525 | 75% | -| gemini-2.5-flash-lite | google | — | $0.1 | $0.025 | $0.35 | 75% | -| gemini-2.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | -| claude-haiku-4-5 | google-vertex | — | $1 | $0.1 | — | 90% | -| claude-opus-4 | google-vertex | — | $15 | $1.5 | — | 90% | -| claude-opus-4-1 | google-vertex | — | $15 | $1.5 | — | 90% | -| claude-opus-4-5 | google-vertex | — | $5 | $0.5 | — | 90% | -| claude-opus-4-6 | google-vertex | — | $5 | $0.5 | — | 90% | -| claude-opus-4-7 | google-vertex | — | $5 | $0.5 | — | 90% | -| claude-sonnet-4 | google-vertex | — | $3 | $0.3 | — | 90% | -| claude-sonnet-4-5 | google-vertex | — | $3 | $0.3 | — | 90% | -| claude-sonnet-4-6 | google-vertex | — | $3 | $0.3 | — | 90% | -| deepseek-v3-1 | google-vertex | — | $0.6 | $0.06 | — | 90% | -| deepseek-v3-2 | google-vertex | — | $0.56 | $0.056 | — | 90% | -| gemini-2-5-flash | google-vertex | — | $0.3 | $0.03 | — | 90% | -| gemini-2-5-flash-lite | google-vertex | — | $0.1 | $0.01 | — | 90% | -| gemini-2-5-pro | google-vertex | — | $1.25 | $0.13 | — | 90% | -| gemini-3-1-flash-lite | google-vertex | — | $0.25 | $0.025 | — | 90% | -| gemini-3-flash | google-vertex | — | $0.5 | $0.05 | — | 90% | -| gemini-3-pro | google-vertex | — | $2 | $0.2 | — | 90% | -| gemma-4-26b | google-vertex | — | $0.15 | $0.015 | — | 90% | -| glm-5 | google-vertex | — | $1 | $0.1 | — | 90% | -| gpt-oss-20b | google-vertex | — | $0.07 | $0.007 | — | 90% | -| grok-4-1-fast | google-vertex | — | $0.2 | $0.05 | — | 75% | -| grok-4-20 | google-vertex | — | $2 | $0.2 | — | 90% | -| kimi-k2-thinking | google-vertex | — | $0.6 | $0.06 | — | 90% | -| minimax-m2 | google-vertex | — | $0.3 | $0.03 | — | 90% | -| qwen3-coder-480b-a35b | google-vertex | — | $0.22 | $0.022 | — | 90% | -| gpt-oss-120b | groq | — | $0.15 | $0.075 | — | 50% | -| gpt-oss-20b | groq | — | $0.075 | $0.0375 | — | 50% | -| kimi-k2-instruct-0905 | groq | — | $1 | $0.5 | — | 50% | -| deepseek--deepseek-v4-flash | hpc-ai | — | $0.14 | $0.028 | — | 80% | -| deepseek--deepseek-v4-pro | hpc-ai | — | $1.74 | $0.145 | — | 92% | -| minimax--minimax-m2.5 | hpc-ai | — | $0.3 | $0.03 | — | 90% | -| moonshotai--kimi-k2.5 | hpc-ai | — | $0.6 | $0.1 | — | 83% | -| moonshotai--kimi-k2.6 | hpc-ai | — | $0.95 | $0.16 | — | 83% | -| xiaomi--mimo-v2.5 | hpc-ai | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | hpc-ai | — | $1 | $0.2 | — | 80% | -| zai--glm-5.1 | hpc-ai | — | $1.4 | $0.26 | — | 81% | -| mercury | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-coder | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-edit | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-edit-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| claude-opus-4-6 | jiekou | — | $2.75 | $0.475 | — | 83% | -| claude-opus-4-6-dd | jiekou | — | $2.85 | $0.275 | — | 90% | -| claude-opus-4-7 | jiekou | — | $4.75 | $0.475 | — | 90% | -| claude-sonnet-4-6 | jiekou | — | $1.65 | $0.285 | — | 83% | -| claude-sonnet-4-6-dd | jiekou | — | $4.75 | $0.165 | — | 97% | -| doubao-seed-1-8-251228 | jiekou | — | $0.11 | $0.0221 | — | 80% | -| gemini-2.5-flash | jiekou | — | $0.095 | $0.0285 | — | 70% | -| gemini-2.5-flash-lite | jiekou | — | $1.1875 | $0.0095 | — | 99% | -| gemini-2.5-pro | jiekou | — | $0.285 | $0.1187 | — | 58% | -| gemini-3-flash-preview | jiekou | — | $0.095 | $0.0475 | — | 50% | -| gemini-3.1-flash-lite-preview | jiekou | — | $1.9 | $0.0237 | — | 99% | -| gemini-3.1-pro-preview | jiekou | — | $0.475 | $0.19 | — | 60% | -| gpt-4.1 | jiekou | — | $9.5 | $0.5 | — | 95% | -| gpt-4.1-mini | jiekou | — | $0.1 | $0.1 | — | 0% | -| gpt-4o | jiekou | — | $1.9 | $1.1875 | — | 38% | -| gpt-5-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | -| gpt-5-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | -| gpt-5-nano | jiekou | — | $0.2375 | $0.0047 | — | 98% | -| gpt-5.1-chat-latest | jiekou | — | $1.1875 | $0.1187 | — | 90% | -| gpt-5.1-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | -| gpt-5.1-codex-max | jiekou | — | $0.2375 | $0.1187 | — | 50% | -| gpt-5.1-codex-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | -| gpt-5.2-chat-latest | jiekou | — | $1.1875 | $0.1662 | — | 86% | -| gpt-5.2-codex | jiekou | — | $1.6625 | $0.175 | — | 89% | -| gpt-5.3-chat-latest | jiekou | — | $1.6625 | $0.1662 | — | 90% | -| gpt-5.3-codex | jiekou | — | $1.75 | $0.1662 | — | 91% | -| gpt-5.4 | jiekou | — | $1.6625 | $0.2375 | — | 86% | -| gpt-5.4-nano | jiekou | — | $0.7125 | $0.019 | — | 97% | -| gpt-5.5 | jiekou | — | $0.19 | $0.5 | — | -163% | -| gpt-5.5-light | jiekou | — | $5 | $0.025 | — | 100% | -| grok-code-fast-1 | jiekou | — | $0.19 | $0.019 | — | 90% | -| minimax-minimax-m2.1 | jiekou | — | $0.55 | $0.03 | — | 95% | -| moonshotai-kimi-k2.5 | jiekou | — | $0.6 | $0.1 | — | 83% | -| o3 | jiekou | — | $1.045 | $0.475 | — | 55% | -| o3-mini | jiekou | — | $1.045 | $0.5225 | — | 50% | -| zai-org-glm-4.7-flash | jiekou | — | $0.6 | $0.01 | — | 98% | -| claude-3-7-sonnet-20250219 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-3-opus | llmgateway | — | $15 | $1.5 | $18.75 | 90% | -| claude-haiku-4-5 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | -| claude-haiku-4-5-20251001 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | -| claude-opus-4-1-20250805 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-20250514 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-5-20251101 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-6 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-7 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | -| claude-sonnet-4-20250514 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-5 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-5-20250929 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-6 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| deepseek-v3.1 | llmgateway | — | $0.56 | $0.112 | — | 80% | -| deepseek-v3.2 | llmgateway | — | $0.28 | $0.028 | — | 90% | -| deepseek-v4-flash | llmgateway | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-pro | llmgateway | — | $0.435 | $0.003625 | — | 99% | -| gemini-2.5-flash | llmgateway | — | $0.3 | $0.03 | — | 90% | -| gemini-2.5-flash-image | llmgateway | — | $0.3 | $0.03 | — | 90% | -| gemini-2.5-flash-lite | llmgateway | — | $0.1 | $0.01 | — | 90% | -| gemini-2.5-pro | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gemini-3-flash-preview | llmgateway | — | $0.5 | $0.05 | — | 90% | -| gemini-3-pro-image-preview | llmgateway | — | $2 | $0.2 | — | 90% | -| gemini-3.1-flash-lite | llmgateway | — | $0.25 | $0.025 | $0.08333 | 90% | -| gemini-3.1-pro-preview | llmgateway | — | $2 | $0.2 | — | 90% | -| gemini-3.5-flash | llmgateway | — | $1.5 | $0.15 | $0.08333 | 90% | -| gemini-pro-latest | llmgateway | — | $2 | $0.2 | — | 90% | -| glm-4.5 | llmgateway | — | $0.6 | $0.11 | — | 82% | -| glm-4.5-air | llmgateway | — | $0.2 | $0.03 | — | 85% | -| glm-4.5-airx | llmgateway | — | $1.1 | $0.22 | — | 80% | -| glm-4.5-x | llmgateway | — | $2.2 | $0.45 | — | 80% | -| glm-4.5v | llmgateway | — | $0.6 | $0.11 | — | 82% | -| glm-4.6v | llmgateway | — | $0.3 | $0.05 | — | 83% | -| glm-4.6v-flashx | llmgateway | — | $0.04 | $0.004 | — | 90% | -| glm-4.7 | llmgateway | — | $0.6 | $0.11 | — | 82% | -| glm-4.7-flashx | llmgateway | — | $0.07 | $0.01 | — | 86% | -| glm-5.1 | llmgateway | — | $1.4 | $0.26 | — | 81% | -| gpt-4.1 | llmgateway | — | $2 | $0.5 | — | 75% | -| gpt-4.1-mini | llmgateway | — | $0.4 | $0.1 | — | 75% | -| gpt-4.1-nano | llmgateway | — | $0.1 | $0.025 | — | 75% | -| gpt-4o | llmgateway | — | $2.5 | $1.25 | — | 50% | -| gpt-5 | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gpt-5-chat-latest | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gpt-5-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | -| gpt-5-nano | llmgateway | — | $0.05 | $0.005 | — | 90% | -| gpt-5.1 | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gpt-5.1-codex-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | -| gpt-5.2 | llmgateway | — | $1.75 | $0.175 | — | 90% | -| gpt-5.2-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | -| gpt-5.3-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | -| gpt-5.4 | llmgateway | — | $2.5 | $0.25 | — | 90% | -| gpt-5.4-mini | llmgateway | — | $0.75 | $0.075 | — | 90% | -| gpt-5.4-nano | llmgateway | — | $0.2 | $0.02 | — | 90% | -| gpt-5.5 | llmgateway | — | $5 | $0.5 | — | 90% | -| grok-4 | llmgateway | — | $3 | $0.75 | — | 75% | -| grok-4-20-beta-0309-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-20-beta-0309-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-20-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-20-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-3 | llmgateway | — | $1.25 | $0.3125 | — | 75% | -| kimi-k2-thinking | llmgateway | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-thinking-turbo | llmgateway | — | $1.15 | $0.15 | — | 87% | -| kimi-k2.5 | llmgateway | — | $0.6 | $0.1 | — | 83% | -| kimi-k2.6 | llmgateway | — | $0.95 | $0.16 | — | 83% | -| mimo-v2-flash | llmgateway | — | $0.1 | $0.01 | — | 90% | -| mimo-v2-omni | llmgateway | — | $0.4 | $0.08 | — | 80% | -| mimo-v2-pro | llmgateway | — | $1 | $0.2 | — | 80% | -| mimo-v2.5 | llmgateway | — | $0.4 | $0.08 | — | 80% | -| mimo-v2.5-pro | llmgateway | — | $1 | $0.2 | — | 80% | -| minimax-m2 | llmgateway | — | $0.2 | $0.03 | — | 85% | -| minimax-m2.5-highspeed | llmgateway | — | $0.6 | $0.03 | — | 95% | -| minimax-m2.7 | llmgateway | — | $0.3 | $0.06 | — | 80% | -| minimax-m2.7-highspeed | llmgateway | — | $0.6 | $0.06 | — | 90% | -| o1 | llmgateway | — | $15 | $7.5 | — | 50% | -| o3 | llmgateway | — | $2 | $0.5 | — | 75% | -| o3-mini | llmgateway | — | $1.1 | $0.55 | — | 50% | -| o4-mini | llmgateway | — | $1.1 | $0.275 | — | 75% | -| qwen-flash | llmgateway | — | $0.05 | $0.01 | $0.0625 | 80% | -| qwen-plus | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | -| qwen-plus-latest | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | -| qwen3-coder-flash | llmgateway | — | $0.3 | $0.06 | $0.375 | 80% | -| qwen3-coder-next | llmgateway | — | $0.108 | $0.06 | — | 44% | -| qwen3-coder-plus | llmgateway | — | $6 | $1.2 | $7.5 | 80% | -| qwen3-max | llmgateway | — | $3 | $0.6 | $3.75 | 80% | -| qwen3-max-2026-01-23 | llmgateway | — | $1.2 | $0.24 | $1.5 | 80% | -| qwen3-vl-flash | llmgateway | — | $0.05 | $0.01 | — | 80% | -| qwen3-vl-plus | llmgateway | — | $0.2 | $0.04 | $0.25 | 80% | -| qwen3.6-max-preview | llmgateway | — | $1.3 | $0.13 | — | 90% | -| qwen3.6-plus | llmgateway | — | $0.5 | $0.05 | — | 90% | -| seed-1-6-250615 | llmgateway | — | $0.25 | $0.05 | — | 80% | -| seed-1-6-250915 | llmgateway | — | $0.25 | $0.05 | — | 80% | -| seed-1-6-flash-250715 | llmgateway | — | $0.07 | $0.015 | — | 79% | -| seed-1-8-251228 | llmgateway | — | $0.25 | $0.05 | — | 80% | -| aion-labs--aion-2.0 | martian | — | $0.8 | $0.2 | — | 75% | -| alibaba--tongyi-deepresearch-30b-a3b | martian | — | $0.09 | $0.09 | — | 0% | -| amazon--nova-premier-v1 | martian | — | $2.5 | $0.625 | — | 75% | -| anthropic--claude-haiku-4-5 | martian | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-haiku-4-5-20251001 | martian | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4-0 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-1 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-1-20250805 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-20250514 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-5 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-5-20251101 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-6 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-7 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-sonnet-4-0 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-20250514 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-5 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-5-20250929 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-6 | martian | — | $3 | $0.3 | $3.75 | 90% | -| arcee-ai--trinity-large-thinking | martian | — | $0.22 | $0.06 | — | 73% | -| bytedance--ui-tars-1.5-7b | martian | — | $0.1 | $0.1 | — | 0% | -| deepinfra--deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | -| deepinfra--deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | -| deepinfra--deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | -| deepinfra--deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | -| deepinfra--deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | -| deepinfra--google--gemini-2.5-flash | martian | — | $0.3 | $0.03 | $0.083333 | 90% | -| deepinfra--google--gemini-2.5-pro | martian | — | $1.25 | $0.125 | $0.375 | 90% | -| deepinfra--z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | -| deepinfra--z-ai--glm-4.7-flash | martian | — | $0.06 | $0.01 | — | 83% | -| deepinfra--z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | -| deepinfra--z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | -| deepseek--deepseek-chat-v3-0324 | martian | — | $0.2 | $0.135 | — | 32% | -| deepseek--deepseek-chat-v3.1 | martian | — | $0.21 | $0.13 | — | 38% | -| deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | -| deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | -| deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | -| deepseek--deepseek-v3.2-speciale | martian | — | $0.287 | $0.058 | — | 80% | -| deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | -| deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | -| ibm-granite--granite-4.1-8b | martian | — | $0.05 | $0.05 | — | 0% | -| inception--mercury-2 | martian | — | $0.25 | $0.025 | — | 90% | -| inclusionai--ling-2.6-1t | martian | — | $0.3 | $0.06 | — | 80% | -| inclusionai--ling-2.6-flash | martian | — | $0.08 | $0.016 | — | 80% | -| kwaipilot--kat-coder-pro-v2 | martian | — | $0.3 | $0.06 | — | 80% | -| microsoft--phi-4-mini-instruct | martian | — | $0.08 | $0.08 | — | 0% | -| minimax--minimax-m2 | martian | — | $0.255 | $0.03 | — | 88% | -| minimax--minimax-m2-her | martian | — | $0.3 | $0.03 | — | 90% | -| minimax--minimax-m2.1 | martian | — | $0.29 | $0.03 | — | 90% | -| mistralai--codestral-2508 | martian | — | $0.3 | $0.03 | — | 90% | -| mistralai--devstral-medium | martian | — | $0.4 | $0.04 | — | 90% | -| mistralai--devstral-small | martian | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-14b-2512 | martian | — | $0.2 | $0.02 | — | 90% | -| mistralai--ministral-3b-2512 | martian | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-8b-2512 | martian | — | $0.15 | $0.015 | — | 90% | -| mistralai--mistral-large | martian | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2407 | martian | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2411 | martian | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2512 | martian | — | $0.5 | $0.05 | — | 90% | -| mistralai--mistral-medium-3 | martian | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-medium-3.1 | martian | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-saba | martian | — | $0.2 | $0.02 | — | 90% | -| mistralai--mistral-small-2603 | martian | — | $0.15 | $0.015 | — | 90% | -| mistralai--mixtral-8x22b-instruct | martian | — | $2 | $0.2 | — | 90% | -| mistralai--pixtral-large-2411 | martian | — | $2 | $0.2 | — | 90% | -| mistralai--voxtral-small-24b-2507 | martian | — | $0.1 | $0.01 | — | 90% | -| moonshotai--kimi-k2-thinking | martian | — | $0.6 | $0.15 | — | 75% | -| moonshotai--kimi-k2.5 | martian | — | $0.4 | $0.05 | — | 88% | -| moonshotai--kimi-k2.6 | martian | — | $0.74 | $0.25 | — | 66% | -| openai--gpt-4.1 | martian | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-2025-04-14 | martian | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-mini | martian | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-mini-2025-04-14 | martian | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-nano | martian | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4.1-nano-2025-04-14 | martian | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4o-2024-08-06 | martian | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-11-20 | martian | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-mini | martian | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-mini-2024-07-18 | martian | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-5 | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-2025-08-07 | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-codex | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-mini | martian | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-mini-2025-08-07 | martian | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-nano | martian | — | $0.05 | $0.01 | — | 80% | -| openai--gpt-5-nano-2025-08-07 | martian | — | $0.05 | $0.01 | — | 80% | -| openai--gpt-5.1 | martian | — | $1.25 | $0.13 | — | 90% | -| openai--gpt-5.1-2025-11-13 | martian | — | $1.25 | $0.13 | — | 90% | -| openai--gpt-5.1-codex | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-max | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-mini | martian | — | $0.25 | $0.03 | — | 88% | -| openai--gpt-5.2 | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-2025-12-11 | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-codex | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-codex | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.4 | martian | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-2026-03-05 | martian | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | martian | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-mini-2026-03-17 | martian | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | martian | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.4-nano-2026-03-17 | martian | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | martian | — | $5 | $0.5 | — | 90% | -| openai--gpt-5.5-2026-04-23 | martian | — | $5 | $0.5 | — | 90% | -| openai--o1 | martian | — | $15 | $7.5 | — | 50% | -| openai--o1-2024-12-17 | martian | — | $15 | $7.5 | — | 50% | -| openai--o3 | martian | — | $2 | $0.5 | — | 75% | -| openai--o3-2025-04-16 | martian | — | $2 | $0.5 | — | 75% | -| openai--o3-mini | martian | — | $1.1 | $0.55 | — | 50% | -| openai--o3-mini-2025-01-31 | martian | — | $1.1 | $0.55 | — | 50% | -| openai--o4-mini | martian | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-2025-04-16 | martian | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-deep-research | martian | — | $2 | $0.5 | — | 75% | -| openai--o4-mini-deep-research-2025-06-26 | martian | — | $2 | $0.5 | — | 75% | -| qwen--qwen-plus | martian | — | $0.26 | $0.052 | $0.325 | 80% | -| qwen--qwen-plus-2025-07-28 | martian | — | $0.26 | — | $0.325 | — | -| qwen--qwen-plus-2025-07-28--thinking | martian | — | $0.26 | — | $0.325 | — | -| qwen--qwen3-30b-a3b-thinking-2507 | martian | — | $0.08 | $0.08 | — | 0% | -| qwen--qwen3-8b | martian | — | $0.05 | $0.05 | — | 0% | -| qwen--qwen3-coder-flash | martian | — | $0.195 | $0.039 | $0.24375 | 80% | -| qwen--qwen3-coder-next | martian | — | $0.11 | $0.07 | — | 36% | -| qwen--qwen3-coder-plus | martian | — | $0.65 | $0.13 | $0.8125 | 80% | -| qwen--qwen3-max | martian | — | $0.78 | $0.156 | $0.975 | 80% | -| qwen--qwen3-vl-235b-a22b-instruct | martian | — | $0.2 | $0.11 | — | 45% | -| qwen--qwen3.5-35b-a3b | martian | — | $0.14 | $0.05 | — | 64% | -| qwen--qwen3.5-flash-02-23 | martian | — | $0.065 | — | $0.08125 | — | -| qwen--qwen3.5-plus-02-15 | martian | — | $0.26 | — | $0.325 | — | -| qwen--qwen3.6-35b-a3b | martian | — | $0.15 | $0.05 | — | 67% | -| qwen--qwen3.6-flash | martian | — | $0.25 | — | $0.3125 | — | -| qwen--qwen3.6-max-preview | martian | — | $1.04 | — | $1.3 | — | -| qwen--qwen3.6-plus | martian | — | $0.325 | — | $0.40625 | — | -| tencent--hy3-preview | martian | — | $0.066 | $0.029 | — | 56% | -| thedrummer--cydonia-24b-v4.1 | martian | — | $0.3 | $0.15 | — | 50% | -| thedrummer--skyfall-36b-v2 | martian | — | $0.55 | $0.25 | — | 55% | -| x-ai--grok-3 | martian | — | $3 | $0.75 | — | 75% | -| x-ai--grok-3-mini | martian | — | $0.3 | $0.075 | — | 75% | -| x-ai--grok-4 | martian | — | $3 | $0.75 | — | 75% | -| x-ai--grok-4.20-multi-agent | martian | — | $2 | $0.2 | — | 90% | -| x-ai--grok-4.3 | martian | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-code-fast-1 | martian | — | $0.2 | $0.02 | — | 90% | -| xiaomi--mimo-v2-flash | martian | — | $0.1 | $0.01 | — | 90% | -| xiaomi--mimo-v2-omni | martian | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2-pro | martian | — | $1 | $0.2 | — | 80% | -| xiaomi--mimo-v2.5 | martian | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | martian | — | $1 | $0.2 | — | 80% | -| z-ai--glm-4.5 | martian | — | $0.6 | $0.11 | — | 82% | -| z-ai--glm-4.5-air | martian | — | $0.13 | $0.025 | — | 81% | -| z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | -| z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | -| z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | -| MiniMax-M2 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | -| MiniMax-M2.1 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | -| MiniMax-M2.1-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | -| MiniMax-M2.5 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | -| MiniMax-M2.5-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | -| MiniMax-M2.7 | minimax | — | $2.1 | $0.42 | $2.625 | 80% | -| MiniMax-M2.7-highspeed | minimax | — | $4.2 | $0.42 | $2.625 | 90% | -| kimi-k2-0711-preview | moonshotai | — | $4 | $1 | — | 75% | -| kimi-k2-0905-preview | moonshotai | — | $4 | $1 | — | 75% | -| kimi-k2-thinking | moonshotai | — | $4 | $1 | — | 75% | -| kimi-k2-thinking-turbo | moonshotai | — | $8 | $1 | — | 88% | -| kimi-k2-turbo-preview | moonshotai | — | $8 | $1 | — | 88% | -| kimi-k2.5 | moonshotai | — | $4 | $0.7 | — | 82% | -| kimi-k2.6 | moonshotai | — | $6.5 | $1.1 | — | 83% | -| kimi-k2.6-long | moonshotai | — | $6.5 | $1.1 | — | 83% | -| kimi-vl-a3b | moonshotai | — | $4 | $0.7 | — | 82% | -| kimi-vl-a3b-thinking | moonshotai | — | $4 | $0.7 | — | 82% | -| aion-labs--aion-2.5 | nanogpt | — | $1 | $0.35 | — | 65% | -| alibaba--qwen3.6-flash | nanogpt | — | $0.19 | $0.02 | $0.24 | 89% | -| deepseek--deepseek-latest | nanogpt | — | $1.1 | $0.11 | — | 90% | -| deepseek--deepseek-v4-flash | nanogpt | — | $0.14 | $0.028 | — | 80% | -| deepseek--deepseek-v4-flash--thinking | nanogpt | — | $0.14 | $0.028 | — | 80% | -| deepseek--deepseek-v4-pro | nanogpt | — | $1.1 | $0.11 | — | 90% | -| deepseek--deepseek-v4-pro--thinking | nanogpt | — | $1.1 | $0.11 | — | 90% | -| deepseek--deepseek-v4-pro-cheaper | nanogpt | — | $0.435 | $0.0435 | — | 90% | -| deepseek--deepseek-v4-pro-cheaper--thinking | nanogpt | — | $0.435 | $0.0435 | — | 90% | -| ernie-5.1 | nanogpt | — | $0.75 | $0.75 | — | 0% | -| ernie-5.1--thinking | nanogpt | — | $0.75 | $0.75 | — | 0% | -| google--gemini-3-flash-preview | nanogpt | — | $0.5 | — | $0.08333 | — | -| google--gemini-3.1-flash-lite | nanogpt | — | $0.25 | — | $0.08333 | — | -| google--gemini-3.1-pro-preview | nanogpt | — | $2 | — | $0.375 | — | -| google--gemini-3.1-pro-preview-customtools | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-pro-preview-high | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-pro-preview-low | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.5-flash | nanogpt | — | $1.5 | — | $0.08333 | — | -| google--gemini-flash-lite-latest | nanogpt | — | $0.25 | $0.025 | $0.08333 | 90% | -| google--gemini-pro-latest | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| ibm-granite--granite-4.1-8b | nanogpt | — | $0.05 | $0.05 | — | 0% | -| inclusionai--ling-2.6-1t | nanogpt | — | $0.3 | $0.06 | — | 80% | -| mercury-2 | nanogpt | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5.4 | nanogpt | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | nanogpt | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | nanogpt | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | nanogpt | — | $5 | $0.5 | — | 90% | -| openai--gpt-chat-latest | nanogpt | — | $5 | $0.5 | — | 90% | -| openai--gpt-latest | nanogpt | — | $5 | $0.5 | — | 90% | -| qwen-3.6-plus | nanogpt | — | $0.325 | $0.0325 | $0.40625 | 90% | -| sarvam-105b | nanogpt | — | $0.045 | $0.028 | — | 38% | -| sarvam-30b | nanogpt | — | $0.028 | $0.017 | — | 39% | -| tencent--hy3-preview | nanogpt | — | $0.066 | $0.029 | — | 56% | -| upstage--solar-pro-3 | nanogpt | — | $0.15 | $0.015 | — | 90% | -| x-ai--grok-4.3 | nanogpt | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-latest | nanogpt | — | $1.25 | $0.2 | — | 84% | -| xiaomi--mimo-v2-omni | nanogpt | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2-pro | nanogpt | — | $1 | $0.2 | — | 80% | -| xiaomi--mimo-v2.5 | nanogpt | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | nanogpt | — | $0.6 | $0.12 | — | 80% | -| z-ai--glm-5-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | -| z-ai--glm-5v-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | -| z-ai--glm-5v-turbo--thinking | nanogpt | — | $1.2 | $0.24 | — | 80% | -| zai-org--glm-4.7-original | nanogpt | — | $0.6 | $0.11 | — | 82% | -| zai-org--glm-4.7-original--thinking | nanogpt | — | $0.6 | $0.11 | — | 82% | -| gpt-4.1 | openai | — | $2 | $0.5 | — | 75% | -| gpt-4.1-mini | openai | — | $0.4 | $0.1 | — | 75% | -| gpt-4.1-nano | openai | — | $0.1 | $0.025 | — | 75% | -| gpt-4o | openai | — | $2.5 | $1.25 | — | 50% | -| gpt-4o-mini | openai | — | $0.15 | $0.075 | — | 50% | -| o1 | openai | — | $15 | $7.5 | — | 50% | -| o1-mini | openai | — | $1.5 | $0.75 | — | 50% | -| o3 | openai | — | $10 | $2.5 | — | 75% | -| o3-mini | openai | — | $1.1 | $0.55 | — | 50% | -| o4-mini | openai | — | $1.1 | $0.275 | — | 75% | -| aion-labs--aion-2.0 | openrouter | — | $0.8 | $0.2 | — | 75% | -| alibaba--tongyi-deepresearch-30b-a3b | openrouter | — | $0.09 | $0.09 | — | 0% | -| amazon--nova-premier-v1 | openrouter | — | $2.5 | $0.625 | — | 75% | -| anthropic--claude-3-haiku | openrouter | — | $0.25 | $0.03 | $0.3 | 88% | -| anthropic--claude-3.5-haiku | openrouter | — | $0.8 | $0.08 | $1 | 90% | -| anthropic--claude-haiku-4.5 | openrouter | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4 | openrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.1 | openrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.5 | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.6 | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.6-fast | openrouter | — | $30 | $3 | $37.5 | 90% | -| anthropic--claude-opus-4.7 | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.7-fast | openrouter | — | $30 | $3 | $37.5 | 90% | -| anthropic--claude-sonnet-4 | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4.5 | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4.6 | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| arcee-ai--trinity-large-thinking | openrouter | — | $0.22 | $0.06 | — | 73% | -| bytedance--ui-tars-1.5-7b | openrouter | — | $0.1 | $0.1 | — | 0% | -| deepseek--deepseek-chat-v3-0324 | openrouter | — | $0.2 | $0.135 | — | 32% | -| deepseek--deepseek-chat-v3.1 | openrouter | — | $0.21 | $0.13 | — | 38% | -| deepseek--deepseek-r1-0528 | openrouter | — | $0.5 | $0.35 | — | 30% | -| deepseek--deepseek-v3.1-terminus | openrouter | — | $0.27 | $0.13 | — | 52% | -| deepseek--deepseek-v3.2 | openrouter | — | $0.252 | $0.0252 | — | 90% | -| deepseek--deepseek-v3.2-speciale | openrouter | — | $0.287 | $0.058 | — | 80% | -| deepseek--deepseek-v4-flash | openrouter | — | $0.112 | $0.022 | — | 80% | -| deepseek--deepseek-v4-pro | openrouter | — | $0.435 | $0.003625 | — | 99% | -| google--gemini-2.0-flash-001 | openrouter | — | $0.1 | $0.025 | $0.083333 | 75% | -| google--gemini-2.5-flash | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | -| google--gemini-2.5-flash-image | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | -| google--gemini-2.5-flash-lite | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | -| google--gemini-2.5-flash-lite-preview-09-2025 | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | -| google--gemini-2.5-pro | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | -| google--gemini-2.5-pro-preview | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | -| google--gemini-2.5-pro-preview-05-06 | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | -| google--gemini-3-flash-preview | openrouter | — | $0.5 | $0.05 | $0.083333 | 90% | -| google--gemini-3-pro-image-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-flash-lite | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | -| google--gemini-3.1-flash-lite-preview | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | -| google--gemini-3.1-pro-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-pro-preview-customtools | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.5-flash | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | -| ibm-granite--granite-4.1-8b | openrouter | — | $0.05 | $0.05 | — | 0% | -| inception--mercury-2 | openrouter | — | $0.25 | $0.025 | — | 90% | -| inclusionai--ling-2.6-1t | openrouter | — | $0.3 | $0.06 | — | 80% | -| inclusionai--ling-2.6-flash | openrouter | — | $0.01 | $0.002 | — | 80% | -| inclusionai--ring-2.6-1t | openrouter | — | $0.075 | $0.015 | — | 80% | -| kwaipilot--kat-coder-pro-v2 | openrouter | — | $0.3 | $0.06 | — | 80% | -| microsoft--phi-4-mini-instruct | openrouter | — | $0.08 | $0.08 | — | 0% | -| minimax--minimax-m2 | openrouter | — | $0.255 | $0.03 | — | 88% | -| minimax--minimax-m2-her | openrouter | — | $0.3 | $0.03 | — | 90% | -| minimax--minimax-m2.1 | openrouter | — | $0.29 | $0.03 | — | 90% | -| mistralai--codestral-2508 | openrouter | — | $0.3 | $0.03 | — | 90% | -| mistralai--devstral-2512 | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--devstral-medium | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--devstral-small | openrouter | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-14b-2512 | openrouter | — | $0.2 | $0.02 | — | 90% | -| mistralai--ministral-3b-2512 | openrouter | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-8b-2512 | openrouter | — | $0.15 | $0.015 | — | 90% | -| mistralai--mistral-large | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2407 | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2512 | openrouter | — | $0.5 | $0.05 | — | 90% | -| mistralai--mistral-medium-3 | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-medium-3.1 | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-saba | openrouter | — | $0.2 | $0.02 | — | 90% | -| mistralai--mistral-small-2603 | openrouter | — | $0.15 | $0.015 | — | 90% | -| mistralai--mixtral-8x22b-instruct | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--pixtral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--voxtral-small-24b-2507 | openrouter | — | $0.1 | $0.01 | — | 90% | -| moonshotai--kimi-k2.5 | openrouter | — | $0.4 | $0.09 | — | 78% | -| moonshotai--kimi-k2.6 | openrouter | — | $0.73 | $0.25 | — | 66% | -| openai--gpt-4.1 | openrouter | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-mini | openrouter | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-nano | openrouter | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4o-2024-08-06 | openrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-11-20 | openrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-mini | openrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-mini-2024-07-18 | openrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-5 | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-chat | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-codex | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-image | openrouter | — | $10 | $1.25 | — | 88% | -| openai--gpt-5-image-mini | openrouter | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5-mini | openrouter | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-nano | openrouter | — | $0.05 | $0.01 | — | 80% | -| openai--gpt-5.1 | openrouter | — | $1.25 | $0.13 | — | 90% | -| openai--gpt-5.1-chat | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-max | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-mini | openrouter | — | $0.25 | $0.03 | — | 88% | -| openai--gpt-5.2 | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-chat | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-codex | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-chat | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-codex | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.4 | openrouter | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-image-2 | openrouter | — | $8 | $2 | — | 75% | -| openai--gpt-5.4-mini | openrouter | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | openrouter | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | openrouter | — | $5 | $0.5 | — | 90% | -| openai--gpt-chat-latest | openrouter | — | $5 | $0.5 | — | 90% | -| openai--gpt-oss-safeguard-20b | openrouter | — | $0.075 | $0.037 | — | 51% | -| openai--o1 | openrouter | — | $15 | $7.5 | — | 50% | -| openai--o3 | openrouter | — | $2 | $0.5 | — | 75% | -| openai--o3-deep-research | openrouter | — | $10 | $2.5 | — | 75% | -| openai--o3-mini | openrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o3-mini-high | openrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o4-mini | openrouter | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-deep-research | openrouter | — | $2 | $0.5 | — | 75% | -| openai--o4-mini-high | openrouter | — | $1.1 | $0.275 | — | 75% | -| qwen--qwen-plus | openrouter | — | $0.26 | $0.052 | $0.325 | 80% | -| qwen--qwen-plus-2025-07-28 | openrouter | — | $0.26 | — | $0.325 | — | -| qwen--qwen-plus-2025-07-28--thinking | openrouter | — | $0.26 | — | $0.325 | — | -| qwen--qwen3-30b-a3b-thinking-2507 | openrouter | — | $0.08 | $0.08 | — | 0% | -| qwen--qwen3-8b | openrouter | — | $0.05 | $0.05 | — | 0% | -| qwen--qwen3-coder-flash | openrouter | — | $0.195 | $0.039 | $0.24375 | 80% | -| qwen--qwen3-coder-next | openrouter | — | $0.11 | $0.07 | — | 36% | -| qwen--qwen3-coder-plus | openrouter | — | $0.65 | $0.13 | $0.8125 | 80% | -| qwen--qwen3-max | openrouter | — | $0.78 | $0.156 | $0.975 | 80% | -| qwen--qwen3-vl-235b-a22b-instruct | openrouter | — | $0.2 | $0.11 | — | 45% | -| qwen--qwen3.5-397b-a17b | openrouter | — | $0.39 | $0.195 | — | 50% | -| qwen--qwen3.5-flash-02-23 | openrouter | — | $0.065 | — | $0.08125 | — | -| qwen--qwen3.5-plus-02-15 | openrouter | — | $0.26 | — | $0.325 | — | -| qwen--qwen3.6-35b-a3b | openrouter | — | $0.15 | $0.05 | — | 67% | -| qwen--qwen3.6-flash | openrouter | — | $0.1875 | — | $0.234375 | — | -| qwen--qwen3.6-max-preview | openrouter | — | $1.04 | — | $1.3 | — | -| qwen--qwen3.6-plus | openrouter | — | $0.325 | — | $0.40625 | — | -| tencent--hy3-preview | openrouter | — | $0.066 | $0.029 | — | 56% | -| thedrummer--cydonia-24b-v4.1 | openrouter | — | $0.3 | $0.15 | — | 50% | -| thedrummer--skyfall-36b-v2 | openrouter | — | $0.55 | $0.25 | — | 55% | -| upstage--solar-pro-3 | openrouter | — | $0.15 | $0.015 | — | 90% | -| x-ai--grok-4.20 | openrouter | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-4.20-multi-agent | openrouter | — | $2 | $0.2 | — | 90% | -| x-ai--grok-4.3 | openrouter | — | $1.25 | $0.2 | — | 84% | -| xiaomi--mimo-v2-flash | openrouter | — | $0.1 | $0.01 | — | 90% | -| xiaomi--mimo-v2-omni | openrouter | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2-pro | openrouter | — | $1 | $0.2 | — | 80% | -| xiaomi--mimo-v2.5 | openrouter | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | openrouter | — | $1 | $0.2 | — | 80% | -| z-ai--glm-4.5 | openrouter | — | $0.6 | $0.11 | — | 82% | -| z-ai--glm-4.5-air | openrouter | — | $0.13 | $0.025 | — | 81% | -| z-ai--glm-4.5v | openrouter | — | $0.6 | $0.11 | — | 82% | -| z-ai--glm-4.6 | openrouter | — | $0.43 | $0.08 | — | 81% | -| z-ai--glm-4.6v | openrouter | — | $0.3 | $0.05 | — | 83% | -| z-ai--glm-4.7 | openrouter | — | $0.4 | $0.08 | — | 80% | -| z-ai--glm-4.7-flash | openrouter | — | $0.06 | $0.01 | — | 83% | -| z-ai--glm-5 | openrouter | — | $0.6 | $0.12 | — | 80% | -| z-ai--glm-5-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | -| z-ai--glm-5v-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | -| ~anthropic--claude-haiku-latest | openrouter | — | $1 | $0.1 | $1.25 | 90% | -| ~anthropic--claude-opus-latest | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| ~anthropic--claude-sonnet-latest | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| ~google--gemini-flash-latest | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | -| ~google--gemini-pro-latest | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| ~moonshotai--kimi-latest | openrouter | — | $0.73 | $0.25 | — | 66% | -| ~openai--gpt-latest | openrouter | — | $5 | $0.5 | — | 90% | -| ~openai--gpt-mini-latest | openrouter | — | $0.75 | $0.075 | — | 90% | -| deepseek--deepseek-v3-0324 | ppio | — | $2 | $0.6 | — | 70% | -| deepseek--deepseek-v3.1 | ppio | — | $4 | $2 | — | 50% | -| deepseek--deepseek-v3.1-terminus | ppio | — | $4 | $2 | — | 50% | -| deepseek--deepseek-v3.2 | ppio | — | $2 | $0.2 | — | 90% | -| deepseek--deepseek-v4-flash | ppio | — | $1 | $0.2 | — | 80% | -| deepseek--deepseek-v4-pro | ppio | — | $12 | $1 | — | 92% | -| minimax--minimax-m2 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | -| minimax--minimax-m2.1 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | -| minimax--minimax-m2.5 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | -| minimax--minimax-m2.5-highspeed | ppio | — | $4.2 | $0.21 | $2.625 | 95% | -| minimax--minimax-m2.7 | ppio | — | $2.1 | $0.42 | $2.625 | 80% | -| minimax--minimax-m2.7-highspeed | ppio | — | $4.2 | $0.42 | $2.625 | 90% | -| moonshotai--kimi-k2.5 | ppio | — | $4 | $0.7 | — | 82% | -| moonshotai--kimi-k2.6 | ppio | — | $6.5 | $1.1 | — | 83% | -| xiaomimimo--mimo-v2-flash | ppio | — | $0.7 | $0.07 | — | 90% | -| zai-org--glm-4.5 | ppio | — | $4 | $0.8 | — | 80% | -| zai-org--glm-4.5-air | ppio | — | $1.2 | $0.24 | — | 80% | -| zai-org--glm-4.5v | ppio | — | $4 | $0.8 | — | 80% | -| zai-org--glm-4.6 | ppio | — | $4 | $0.8 | $0.8 | 80% | -| zai-org--glm-4.7-flash | ppio | — | $0.5 | $0.1 | — | 80% | -| gemma-3-27b | privatemode | — | $0.77 | $0.08 | — | 90% | -| gemma-4-31b | privatemode | — | $0.77 | $0.08 | — | 90% | -| gpt-oss-120b | privatemode | — | $0.43 | $0.04 | — | 91% | -| kimi-k2-5 | privatemode | — | $1.55 | $0.15 | — | 90% | -| qwen3-coder-30b-a3b | privatemode | — | $0.43 | $0.04 | — | 91% | -| alibaba--qwen-max | requesty | — | $1.6 | $1.6 | $1.6 | 0% | -| alibaba--qwen-plus | requesty | — | $0.4 | $0.4 | $0.4 | 0% | -| alibaba--qwen-turbo | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| alibaba--qwen3-30b-a3b-instruct-2507 | requesty | — | $0.2 | $0.2 | $0.8 | 0% | -| alibaba--qwen3-coder-flash | requesty | — | $0.3 | $0.08 | $0.3 | 73% | -| alibaba--qwen3.5 | requesty | — | $0.6 | $0.6 | $3.6 | 0% | -| anthropic--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| azure--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| azure--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| azure--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| azure--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| azure--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | -| azure--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | -| azure--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| azure--gpt-5.2 | requesty | — | $1.75 | $0.175 | $14 | 90% | -| azure--openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| azure--openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| azure--openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| bedrock--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | -| bedrock--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | -| bedrock--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| bedrock--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| bedrock--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| bedrock--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| bedrock--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| bedrock--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| deepinfra--deepseek-ai--deepseek-r1 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | -| deepinfra--deepseek-ai--deepseek-r1-distill-llama-70b | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--deepseek-ai--deepseek-v3 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | -| deepinfra--deepseek-ai--deepseek-v3.1 | requesty | — | $0.3 | $0.3 | $1 | 0% | -| deepinfra--kimi k2.5 | requesty | — | $0.45 | $0.07 | — | 84% | -| deepinfra--meta-llama--llama-3.2-90b-vision-instruct | requesty | — | $0.35 | $0.35 | $0.35 | 0% | -| deepinfra--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.12 | $0.12 | $0.12 | 0% | -| deepinfra--meta-llama--meta-llama-3.1-405b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | -| deepinfra--meta-llama--meta-llama-3.1-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.02 | $0.02 | $0.02 | 0% | -| deepinfra--microsoft--phi-4 | requesty | — | $0.07 | $0.07 | $0.07 | 0% | -| deepinfra--qwen--qwen2.5-72b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | — | $0.07 | $0.07 | $0.07 | 0% | -| deepinfra--qwen--qwen3-235b-a22b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| deepinfra--qwen--qwen3-32b | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| deepinfra--qwen--qwen3-coder-480b-a35b-instruct | requesty | — | $0.4 | $0.4 | $1.6 | 0% | -| deepinfra--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | -| deepinfra--zai-org--glm-4.5-air | requesty | — | $0.2 | $0.2 | $1.1 | 0% | -| deepseek--deepseek-chat | requesty | — | $0.14 | $0.028 | $0.14 | 80% | -| deepseek--deepseek-reasoner | requesty | — | $0.14 | $0.028 | $0.14 | 80% | -| deepseek--deepseek-v4-flash | requesty | — | $0.14 | $0.028 | $0.14 | 80% | -| deepseek--deepseek-v4-pro | requesty | — | $1.74 | $0.145 | $1.74 | 92% | -| fireworks--deepseek-v3.2 | requesty | — | $0.56 | $0.28 | — | 50% | -| fireworks--deepseek-v4-pro | requesty | — | $1.74 | $0.15 | — | 91% | -| fireworks--glm-5 | requesty | — | $1 | $0.2 | — | 80% | -| fireworks--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | -| fireworks--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $3 | 83% | -| fireworks--kimi-k2.6 | requesty | — | $0.95 | $0.16 | — | 83% | -| fireworks--minimax-m2.5 | requesty | — | $0.3 | $0.03 | — | 90% | -| fireworks--minimax-m2.7 | requesty | — | $0.3 | $0.06 | — | 80% | -| fireworks--qwen3.6-plus | requesty | — | $0.5 | $0.1 | — | 80% | -| google--gemini-2.0-flash-001 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| google--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | -| google--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | -| google--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | -| google--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | -| google--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| google--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | -| google--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| groq--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.75 | 0% | -| groq--openai--gpt-oss-20b | requesty | — | $0.1 | $0.1 | $0.5 | 0% | -| inceptron--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | -| inceptron--kimi-k2.6 | requesty | — | $0.8 | $0.2 | — | 75% | -| inceptron--minimax-m2.5 | requesty | — | $0.28 | $0.03 | — | 89% | -| minimaxi--minimax-m2 | requesty | — | $0.3 | $0.3 | $1.2 | 0% | -| minimaxi--minimax-m2.5 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | -| minimaxi--minimax-m2.5-highspeed | requesty | — | $0.6 | $0.06 | $2.4 | 90% | -| minimaxi--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | -| minimaxi--minimax-m2.7-highspeed | requesty | — | $0.6 | $0.06 | $1.2 | 90% | -| mistral--codestral-latest | requesty | — | $0.3 | $0.3 | $0.9 | 0% | -| mistral--devstral-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | -| mistral--devstral-medium-2507 | requesty | — | $0.4 | $0.4 | $2 | 0% | -| mistral--devstral-small-2507 | requesty | — | $0.1 | $0.1 | $0.3 | 0% | -| mistral--devstral-small-latest | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| mistral--mistral-large-latest | requesty | — | $0.5 | $0.5 | $0.5 | 0% | -| mistral--mistral-medium-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | -| mistral--mistral-small-2503 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| mistral--mistral-small-2603 | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| mistral--mistral-small-latest | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| mistral--open-mistral-7b | requesty | — | $0.25 | $0.25 | $0.25 | 0% | -| mistral--pixtral-large-latest | requesty | — | $2 | $2 | $5 | 0% | -| moonshot--kimi-k2-0711-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-0905-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-thinking | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-thinking-turbo | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-turbo-preview | requesty | — | $1.2 | $0.3 | $5 | 75% | -| moonshot--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $0.6 | 83% | -| moonshot--kimi-k2.6 | requesty | — | $0.95 | $0.16 | $0.96 | 83% | -| nebius--deepseek-ai--deepseek-v3.2 | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| nebius--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.13 | $0.13 | $0.13 | 0% | -| nebius--moonshotai--kimi-k2.5 | requesty | — | $0.5 | $0.5 | $2.5 | 0% | -| nebius--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| novita--deepseek--deepseek-prover-v2-671b | requesty | — | $0.7 | $0.7 | $0.7 | 0% | -| novita--deepseek--deepseek-r1 | requesty | — | $4 | $4 | $4 | 0% | -| novita--deepseek--deepseek-r1-distill-llama-70b | requesty | — | $0.8 | $0.8 | $0.8 | 0% | -| novita--deepseek--deepseek-r1-distill-qwen-14b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| novita--deepseek--deepseek-r1-distill-qwen-32b | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| novita--deepseek--deepseek-r1-turbo | requesty | — | $0.7 | $0.7 | $2.5 | 0% | -| novita--deepseek--deepseek-v3-0324 | requesty | — | $0.4 | $0.4 | $0.4 | 0% | -| novita--deepseek--deepseek-v3-turbo | requesty | — | $0.4 | $0.4 | $0.4 | 0% | -| novita--deepseek--deepseek-v3.2 | requesty | — | $0.269 | $0.1345 | $0.269 | 50% | -| novita--deepseek--deepseek_v3 | requesty | — | $0.89 | $0.89 | $0.89 | 0% | -| novita--glm-5 | requesty | — | $1 | $0.2 | — | 80% | -| novita--gryphe--mythomax-l2-13b | requesty | — | $0.09 | $0.09 | $0.09 | 0% | -| novita--meta-llama--llama-3-70b-instruct | requesty | — | $0.51 | $0.51 | $0.51 | 0% | -| novita--meta-llama--llama-3-8b-instruct | requesty | — | $0.04 | $0.04 | $0.04 | 0% | -| novita--meta-llama--llama-3.1-8b-instruct | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| novita--meta-llama--llama-3.2-1b-instruct | requesty | — | $0.02 | $0.02 | $0.02 | 0% | -| novita--meta-llama--llama-3.2-3b-instruct | requesty | — | $0.03 | $0.03 | $0.03 | 0% | -| novita--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.39 | $0.39 | $0.39 | 0% | -| novita--meta-llama--llama-4-maverick-17b-128e-instruct-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| novita--microsoft--wizardlm-2-8x22b | requesty | — | $0.62 | $0.62 | $0.62 | 0% | -| novita--minimax--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | -| novita--mistralai--mistral-nemo | requesty | — | $0.17 | $0.17 | $0.17 | 0% | -| novita--moonshotai--kimi-k2-instruct | requesty | — | $0.57 | $0.57 | $0.57 | 0% | -| novita--nousresearch--hermes-2-pro-llama-3-8b | requesty | — | $0.14 | $0.14 | $0.14 | 0% | -| novita--qwen--qwen-2.5-72b-instruct | requesty | — | $0.38 | $0.38 | $0.38 | 0% | -| novita--qwen--qwen2.5-vl-72b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | -| novita--qwen--qwen3-235b-a22b-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| novita--qwen--qwen3.5-397b-a17b | requesty | — | $0.6 | $0.6 | $3.6 | 0% | -| novita--sao10k--l3-70b-euryale-v2.1 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | -| novita--sao10k--l3-8b-lunaris | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| novita--sao10k--l3-8b-stheno-v3.2 | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| novita--sao10k--l31-70b-euryale-v2.2 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | -| novita--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | -| novita--zai-org--glm-4.6 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | -| openai--chatgpt-4o | requesty | — | $5 | $5 | $5 | 0% | -| openai--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| openai--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| openai--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| openai--gpt-4o | requesty | — | $2.5 | $1.25 | $2.5 | 50% | -| openai--gpt-4o-2024-05-13 | requesty | — | $2.5 | $2.5 | $2.5 | 0% | -| openai--gpt-4o-2024-08-06 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | -| openai--gpt-4o-2024-11-20 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | -| openai--gpt-4o-mini | requesty | — | $0.15 | $0.075 | $0.15 | 50% | -| openai--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | -| openai--gpt-5-mini:flex | requesty | — | $0.125 | $0.0125 | $0.125 | 90% | -| openai--gpt-5-mini:priority | requesty | — | $0.45 | $0.045 | $3.6 | 90% | -| openai--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | -| openai--gpt-5-nano:flex | requesty | — | $0.025 | $0.0025 | $0.025 | 90% | -| openai--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5.1-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | -| openai--gpt-5.2-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai--gpt-5.4 | requesty | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | -| openai--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | -| openai--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | -| openai--gpt-5:flex | requesty | — | $0.625 | $0.0625 | $0.625 | 90% | -| openai--gpt-5:priority | requesty | — | $2.5 | $0.25 | $20 | 90% | -| openai--o1 | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o1-pro | requesty | — | $150 | $150 | $150 | 0% | -| openai--o1:high | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o1:low | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o1:medium | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o3 | requesty | — | $2 | $0.5 | $2 | 75% | -| openai--o3-deep-research | requesty | — | $10 | $2.5 | $10 | 75% | -| openai--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3-mini:high | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3-mini:low | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3-mini:medium | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3:flex | requesty | — | $1 | $0.25 | $1 | 75% | -| openai--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai--o4-mini-deep-research | requesty | — | $2 | $0.5 | $2 | 75% | -| openai--o4-mini:flex | requesty | — | $0.55 | $0.138 | $0.55 | 75% | -| openai--o4-mini:high | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai--o4-mini:low | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai--o4-mini:medium | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| openai-responses--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai-responses--gpt-5-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | -| openai-responses--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | -| openai-responses--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | -| openai-responses--gpt-5-pro | requesty | — | $15 | $15 | $120 | 0% | -| openai-responses--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai-responses--gpt-5.1-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | -| openai-responses--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | -| openai-responses--gpt-5.2-codex | requesty | — | $1.75 | $0.175 | $1.75 | 90% | -| openai-responses--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai-responses--gpt-5.3-codex | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai-responses--gpt-5.4 | requesty | — | $2.5 | $0.25 | $2.5 | 90% | -| openai-responses--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | -| openai-responses--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | -| openai-responses--gpt-5.4-pro | requesty | — | $30 | $30 | $180 | 0% | -| openai-responses--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | -| openai-responses--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai-responses--o3-pro | requesty | — | $20 | $20 | $20 | 0% | -| openai-responses--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| parasail--parasail-gemma3-27b-it | requesty | — | $0.3 | $0.3 | $0.5 | 0% | -| parasail--parasail-kimi-k2-instruct | requesty | — | $0.99 | $0.99 | $2.99 | 0% | -| parasail--parasail-qwen25-vl-72b-instruct | requesty | — | $0.7 | $0.7 | $0.7 | 0% | -| parasail--parasail-qwen3-235b-a22b-instruct-2507 | requesty | — | $0.15 | $0.15 | $0.85 | 0% | -| perplexity--sonar | requesty | — | $1 | $1 | $1 | 0% | -| perplexity--sonar-pro | requesty | — | $3 | $3 | $3 | 0% | -| perplexity--sonar-reasoning-pro | requesty | — | $2 | $2 | $2 | 0% | -| together--deepseek-ai--deepseek-r1 | requesty | — | $3 | $3 | $7 | 0% | -| together--deepseek-ai--deepseek-v3 | requesty | — | $1.25 | $1.25 | $1.25 | 0% | -| together--meta-llama--llama-3.2-3b-instruct-turbo | requesty | — | $0.06 | $0.06 | $0.06 | 0% | -| together--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | -| together--meta-llama--llamaguard-2-8b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| together--meta-llama--meta-llama-3-8b-instruct-lite | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| together--meta-llama--meta-llama-3.1-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | -| together--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.18 | $0.18 | $0.18 | 0% | -| together--qwen--qwen2.5-72b-instruct-turbo | requesty | — | $1.2 | $1.2 | $1.2 | 0% | -| together--qwen--qwen2.5-7b-instruct-turbo | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| vertex--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | -| vertex--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| vertex--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| vertex--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| vertex--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| vertex--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| vertex--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--deepseek-v3.2 | requesty | — | $0.56 | $0.056 | $1.68 | 90% | -| vertex--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | -| vertex--gemini-2.5-flash-image | requesty | — | $0.3 | $0.3 | $2.5 | 0% | -| vertex--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | -| vertex--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | -| vertex--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | -| vertex--gemini-3-flash-preview:flex | requesty | — | $0.25 | $0.025 | $0.5 | 90% | -| vertex--gemini-3-pro-image-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| vertex--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| vertex--gemini-3.1-flash-lite | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | -| vertex--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | -| vertex--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| vertex--gemini-3.1-pro-preview:flex | requesty | — | $1 | $0.1 | $2.75 | 90% | -| vertex--gemini-3.5-flash | requesty | — | $1.5 | $0.15 | $1.583 | 90% | -| vertex--kimi-k2 | requesty | — | $0.6 | $0.06 | $2.5 | 90% | -| xai--grok-2-1212 | requesty | — | $2 | $2 | $2 | 0% | -| xai--grok-3 | requesty | — | $5 | $5 | $5 | 0% | -| xai--grok-3-mini | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| xai--grok-3-mini:high | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| xai--grok-3-mini:low | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| xai--grok-4 | requesty | — | $3 | $0.75 | $3 | 75% | -| xai--grok-4-1-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4-1-fast-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4-fast | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4.2-beta | requesty | — | $2 | $0.2 | $2 | 90% | -| xai--grok-4.3 | requesty | — | $1.25 | $0.2 | $1.25 | 84% | -| xai--grok-code-fast-1 | requesty | — | $0.2 | $0.02 | $0.2 | 90% | -| zai--glm-4.5 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | -| zai--glm-4.6 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | -| zai--glm-4.7 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | -| zai--glm-5 | requesty | — | $1 | $0.2 | — | 80% | -| zai--glm-5.1 | requesty | — | $1.4 | $0.26 | $4.4 | 81% | -| deepseek-v3.2 | siliconflow | — | $0.27 | $0.135 | — | 50% | -| deepseek-v4-flash | siliconflow | — | $0.14 | $0.028 | — | 80% | -| deepseek-v4-pro | siliconflow | — | $1.74 | $0.145 | — | 92% | -| glm-4.7 | siliconflow | — | $0.42 | $0.11 | — | 74% | -| glm-5 | siliconflow | — | $0.95 | $0.2 | — | 79% | -| glm-5.1 | siliconflow | — | $1.4 | $0.26 | — | 81% | -| hy3-preview | siliconflow | — | $0.066 | $0.029 | — | 56% | -| kimi-k2.5 | siliconflow | — | $0.45 | $0.07 | — | 84% | -| kimi-k2.6 | siliconflow | — | $0.9 | $0.2 | — | 78% | -| minimax-m2.5 | siliconflow | — | $0.3 | $0.03 | — | 90% | -| step-1-32k | stepfun | — | $15 | $3 | — | 80% | -| step-1-8k | stepfun | — | $5 | $1 | — | 80% | -| step-1o-audio | stepfun | — | $25 | $5 | — | 80% | -| step-1o-turbo-vision | stepfun | — | $2.5 | $0.5 | — | 80% | -| step-1o-vision-32k | stepfun | — | $15 | $3 | — | 80% | -| step-1v-32k | stepfun | — | $15 | $3 | — | 80% | -| step-1v-8k | stepfun | — | $5 | $1 | — | 80% | -| step-2-16k | stepfun | — | $38 | $7.6 | — | 80% | -| step-2-16k-exp | stepfun | — | $38 | $7.6 | — | 80% | -| step-2-mini | stepfun | — | $1 | $0.2 | — | 80% | -| step-3 | stepfun | — | $1.5 | $0.3 | — | 80% | -| step-3.5-flash | stepfun | — | $0.7 | $0.14 | — | 80% | -| step-3.5-flash-2603 | stepfun | — | $0.7 | $0.14 | — | 80% | -| step-audio-2 | stepfun | — | $10 | $2 | — | 80% | -| step-r1-v-mini | stepfun | — | $2.5 | $0.5 | — | 80% | -| stepaudio-2.5-chat | stepfun | — | $10 | $2 | — | 80% | -| stepaudio-2.5-realtime | stepfun | — | $10 | $2 | — | 80% | -| deepseek-v4-flash | tencent-tokenhub | — | $1 | $0.2 | — | 80% | -| deepseek-v4-pro | tencent-tokenhub | — | $12 | $1 | — | 92% | -| glm-5 | tencent-tokenhub | — | $4 | $1 | — | 75% | -| glm-5-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | -| glm-5.1 | tencent-tokenhub | — | $6 | $1.3 | — | 78% | -| glm-5v-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | -| hy3-preview | tencent-tokenhub | — | $1.2 | $0.4 | — | 67% | -| kimi-k2.5 | tencent-tokenhub | — | $4 | $0.7 | — | 82% | -| kimi-k2.6 | tencent-tokenhub | — | $6.5 | $1.1 | — | 83% | -| minimax-m2.5 | tencent-tokenhub | — | $2.1 | $0.21 | — | 90% | -| minimax-m2.7 | tencent-tokenhub | — | $2.1 | $0.42 | — | 80% | -| MiniMaxAI--MiniMax-M2.5 | togetherai | — | $0.3 | $0.06 | — | 80% | -| MiniMaxAI--MiniMax-M2.7 | togetherai | — | $0.3 | $0.06 | — | 80% | -| deepseek-ai--DeepSeek-V4-Pro | togetherai | — | $2.1 | $0.2 | — | 90% | -| moonshotai--Kimi-K2.6 | togetherai | — | $1.2 | $0.2 | — | 83% | -| solar-pro2 | upstage | — | $0.15 | $0.015 | — | 90% | -| solar-pro3 | upstage | — | $0.15 | $0.015 | — | 90% | -| aion-labs-aion-2-0 | venice | — | $1 | $0.25 | — | 75% | -| arcee-trinity-large-thinking | venice | — | $0.3125 | $0.075 | — | 76% | -| claude-opus-4-5 | venice | — | $6 | $0.6 | — | 90% | -| claude-opus-4-6 | venice | — | $6 | $0.6 | — | 90% | -| claude-opus-4-6-fast | venice | — | $36 | $3.6 | — | 90% | -| claude-opus-4-7 | venice | — | $6 | $0.6 | — | 90% | -| claude-opus-4-7-fast | venice | — | $36 | $3.6 | — | 90% | -| claude-sonnet-4-5 | venice | — | $3.75 | $0.375 | — | 90% | -| claude-sonnet-4-6 | venice | — | $3.6 | $0.36 | — | 90% | -| deepseek-v3.2 | venice | — | $0.33 | $0.16 | — | 52% | -| deepseek-v4-flash | venice | — | $0.17 | $0.028 | — | 84% | -| deepseek-v4-pro | venice | — | $1.73 | $0.33 | — | 81% | -| gemini-3-1-pro-preview | venice | — | $2.5 | $0.5 | — | 80% | -| gemini-3-flash-preview | venice | — | $0.7 | $0.07 | — | 90% | -| grok-4-20 | venice | — | $1.42 | $0.23 | — | 84% | -| grok-4-20-multi-agent | venice | — | $1.42 | $0.23 | — | 84% | -| grok-4-3 | venice | — | $1.42 | $0.23 | — | 84% | -| kimi-k2-5 | venice | — | $0.56 | $0.22 | — | 61% | -| kimi-k2-6 | venice | — | $0.85 | $0.22 | — | 74% | -| mercury-2 | venice | — | $0.3125 | $0.03125 | — | 90% | -| minimax-m25 | venice | — | $0.34 | $0.04 | — | 88% | -| minimax-m27 | venice | — | $0.375 | $0.075 | — | 80% | -| openai-gpt-4o-mini-2024-07-18 | venice | — | $0.1875 | $0.09375 | — | 50% | -| openai-gpt-52 | venice | — | $2.19 | $0.219 | — | 90% | -| openai-gpt-52-codex | venice | — | $2.19 | $0.219 | — | 90% | -| openai-gpt-53-codex | venice | — | $2.19 | $0.219 | — | 90% | -| openai-gpt-54 | venice | — | $3.13 | $0.313 | — | 90% | -| openai-gpt-54-mini | venice | — | $0.9375 | $0.09375 | — | 90% | -| openai-gpt-55 | venice | — | $6.25 | $0.625 | — | 90% | -| qwen-3-6-plus | venice | — | $0.625 | $0.0625 | — | 90% | -| qwen3-5-35b-a3b | venice | — | $0.3125 | $0.15625 | — | 50% | -| qwen3-coder-480b-a35b-instruct-turbo | venice | — | $0.35 | $0.04 | — | 89% | -| z-ai-glm-5-turbo | venice | — | $1.2 | $0.24 | — | 80% | -| z-ai-glm-5v-turbo | venice | — | $1.5 | $0.3 | — | 80% | -| zai-org-glm-4.6 | venice | — | $0.85 | $0.3 | — | 65% | -| zai-org-glm-4.7 | venice | — | $0.55 | $0.11 | — | 80% | -| zai-org-glm-5 | venice | — | $1 | $0.2 | — | 80% | -| zai-org-glm-5-1 | venice | — | $1.75 | $0.325 | — | 81% | -| GLM-5.1 | wafer | — | $1.5 | $0.15 | — | 90% | -| Qwen3.5-397B-A17B | wafer | — | $0.6 | $0.06 | — | 90% | +| Model | Provider | Context | Input $/M | Cache Read $/M | Cache Write $/M | Savings | +| --------------------------------------------- | -------------- | ------- | ------------------- | -------------------- | --------------- | ------- | +| aistudio_gemini-2.0-flash | aihubmix | — | $0.05 | $0.125 | — | -150% | +| aistudio_gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| anthropic-opus-4-6 | aihubmix | — | $2.5 | $0.25 | $3.125 | 90% | +| claude-haiku-4-5 | aihubmix | — | $0.55 | $0.055 | $0.6875 | 90% | +| claude-sonnet-4-0 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5-think | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| codex-mini-latest | aihubmix | — | $0.75 | $0.1875 | — | 75% | +| deepseek-v3.2 | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| deepseek-v3.2-exp | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-exp-think | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-think | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| doubao-1.5-lite-32k | aihubmix | — | $0.025 | $0.005 | — | 80% | +| doubao-1.5-pro-32k | aihubmix | — | $0.067 | $0.0134 | — | 80% | +| doubao-lite-32k | aihubmix | — | $0.03 | $0.006 | — | 80% | +| doubao-pro-32k | aihubmix | — | $0.07 | $0.014 | — | 80% | +| doubao-seed-1-6 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-flash | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-flash-250615 | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-lite | aihubmix | — | $0.041 | $0.0082 | — | 80% | +| doubao-seed-1-6-thinking | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-thinking-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-vision-250815 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| doubao-seed-1-8 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| gemini-2.0-flash | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.0-flash-001 | aihubmix | — | $0.05 | $0.125 | — | -150% | +| gemini-2.0-flash-search | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.5-flash | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-lite | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-pro | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| glm-4.5-airx | aihubmix | — | $0.55 | $0.11 | — | 80% | +| glm-4.5-x | aihubmix | — | $1.1 | $0.22 | — | 80% | +| glm-4.6 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| glm-4.6v | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| glm-4.7 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| gpt-4.1 | aihubmix | — | $1 | $0.25 | — | 75% | +| gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| gpt-4.1-nano | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gpt-4o | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06-global | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-11-20 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-mini | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-2024-07-18 | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-global | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-search-preview | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-search-preview | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-5 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5-nano | aihubmix | — | $0.025 | $0.0025 | — | 90% | +| gpt-5.1 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-max | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5.2 | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-chat-latest | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-codex | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-high | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-low | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-pro | aihubmix | — | $10.5 | $1.05 | — | 90% | +| grok-4 | aihubmix | — | $1.65 | $0.4125 | — | 75% | +| grok-4-1-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-1-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4.20-beta-0309-non-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-beta-0309-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-beta-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-code-fast-1 | aihubmix | — | $0.1 | $0.025 | — | 75% | +| kimi-k2-thinking | aihubmix | — | $0.274 | $0.0685 | — | 75% | +| kimi-k2-turbo-preview | aihubmix | — | $0.6 | $0.15 | — | 75% | +| kimi-k2.5 | aihubmix | — | $0.3 | $0.0525 | — | 82% | +| mimo-v2-flash | aihubmix | — | $0.0959 | $0.01918 | — | 80% | +| mimo-v2-omni | aihubmix | — | $0.22 | $0.044 | — | 80% | +| mimo-v2-pro | aihubmix | — | $0.55 | $0.11 | — | 80% | +| nvidia-nemotron-3-super-120b-a12b | aihubmix | — | $0.055 | $0.01375 | — | 75% | +| o1 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-2024-12-17 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-global | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-mini | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-mini-2024-09-12 | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-preview | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-preview-2024-09-12 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o3 | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-deep-research | aihubmix | — | $5 | $1.25 | — | 75% | +| o3-global | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-mini | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o3-mini-global | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o4-mini | aihubmix | — | $0.55 | $0.1375 | — | 75% | +| qwen-plus | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-04-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-07-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-latest | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-turbo | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen-turbo-latest | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen3-coder-plus | aihubmix | — | $0.27 | $0.054 | — | 80% | +| qwen3-max | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-2026-01-23 | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-preview | aihubmix | — | $0.423 | $0.0846 | — | 80% | +| qwen3-vl-flash | aihubmix | — | $0.0103 | $0.00206 | — | 80% | +| qwen3-vl-plus | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| zai-glm-5-turbo | aihubmix | — | $0.6 | $0.12 | — | 80% | +| aion-2.0 | aion | — | $0.7999999999999999 | $0.19999999999999998 | — | 75% | +| aion-2.5 | aion | — | $1 | $0.35 | — | 65% | +| amazon-nova-2-lite | amazon-bedrock | — | $0.33 | $0.0825 | — | 75% | +| amazon-nova-lite | amazon-bedrock | — | $0.06 | $0.015 | — | 75% | +| amazon-nova-micro | amazon-bedrock | — | $0.035 | $0.00875 | — | 75% | +| amazon-nova-premier | amazon-bedrock | — | $2.5 | $0.625 | — | 75% | +| amazon-nova-pro | amazon-bedrock | — | $0.8 | $0.2 | — | 75% | +| claude-haiku-4-5-20251001 | auriko | — | $1 | $0.1 | $1.25 | 90% | +| claude-opus-4-1-20250805 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-20250514 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-5-20251101 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-6 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-7 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-sonnet-4-20250514 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5-20250929 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-6 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| deepseek-r1-0528 | auriko | — | $0.5 | $0.35 | — | 30% | +| deepseek-v3-0324 | auriko | — | $0.2 | $0.135 | — | 32% | +| deepseek-v3.1 | auriko | — | $0.21 | $0.13 | — | 38% | +| deepseek-v3.1-terminus | auriko | — | $0.27 | $0.13 | — | 52% | +| deepseek-v3.2 | auriko | — | $0.26 | $0.13 | — | 50% | +| deepseek-v4-flash | auriko | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | auriko | — | $0.435 | $0.003625 | — | 99% | +| gemini-2.5-flash | auriko | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-lite | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-pro | auriko | — | $1.25 | $0.125 | — | 90% | +| gemini-3-flash-preview | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-3.1-flash-lite | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-flash-lite-preview | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-pro-preview | auriko | — | $2 | $0.2 | — | 90% | +| gemini-3.1-pro-preview-customtools | auriko | — | $2 | $0.2 | — | 90% | +| gemini-flash-latest | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-flash-lite-latest | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-pro-latest | auriko | — | $2 | $0.2 | — | 90% | +| glm-4.5 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.5-air | auriko | — | $0.2 | $0.03 | — | 85% | +| glm-4.5-airx | auriko | — | $1.1 | $0.22 | — | 80% | +| glm-4.5-x | auriko | — | $2.2 | $0.45 | — | 80% | +| glm-4.5v | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6v | auriko | — | $0.3 | $0.05 | — | 83% | +| glm-4.6v-flashx | auriko | — | $0.04 | $0.004 | — | 90% | +| glm-4.7 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.7-flashx | auriko | — | $0.07 | $0.01 | — | 86% | +| glm-5 | auriko | — | $1 | $0.2 | — | 80% | +| glm-5-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| glm-5.1 | auriko | — | $1.4 | $0.26 | — | 81% | +| glm-5v-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| gpt-4.1-2025-04-14 | auriko | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini-2025-04-14 | auriko | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano-2025-04-14 | auriko | — | $0.1 | $0.025 | — | 75% | +| gpt-4o-2024-08-06 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-2024-11-20 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-mini-2024-07-18 | auriko | — | $0.15 | $0.075 | — | 50% | +| gpt-5-2025-08-07 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-mini-2025-08-07 | auriko | — | $0.25 | $0.025 | — | 90% | +| gpt-5-nano-2025-08-07 | auriko | — | $0.05 | $0.005 | — | 90% | +| gpt-5.1-2025-11-13 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.1-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.2-2025-12-11 | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.2-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.3-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.4-2026-03-05 | auriko | — | $2.5 | $0.25 | — | 90% | +| gpt-5.4-mini-2026-03-17 | auriko | — | $0.75 | $0.075 | — | 90% | +| gpt-5.4-nano-2026-03-17 | auriko | — | $0.2 | $0.02 | — | 90% | +| gpt-5.5-2026-04-23 | auriko | — | $5 | $0.5 | — | 90% | +| gpt-oss-120b | auriko | — | $0.15 | $0.01 | — | 93% | +| gpt-oss-20b | auriko | — | $0.07 | $0.04 | — | 43% | +| grok-4.20-0309-non-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.20-0309-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.3 | auriko | — | $1.25 | $0.2 | — | 84% | +| hy3-preview | auriko | — | $0.066 | $0.029 | — | 56% | +| kimi-k2-0711-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-0905-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking-turbo | auriko | — | $1.15 | $0.15 | — | 87% | -> 💡 Use the [interactive catalog](https://i-need-token.github.io/ai-models/) to search and filter models with more criteria. +> 📄 Showing first 200 of 1374 models. Use the [interactive catalog](https://i-need-token.github.io/ai-models/) to browse all. --- diff --git a/docs/zh/cached-pricing.md b/docs/zh/cached-pricing.md index d4556d7e..c2701e50 100644 --- a/docs/zh/cached-pricing.md +++ b/docs/zh/cached-pricing.md @@ -27,1384 +27,210 @@ ## 模型定价 -| 模型 | 提供商 | 上下文 | 输入 $/M | 缓存读取 $/M | 缓存写入 $/M | 节省 | -| ---------------------------------------------------------- | ---------------- | ------ | ------------------- | --------------------- | ------------ | ----- | -| aistudio_gemini-2.0-flash | aihubmix | — | $0.05 | $0.125 | — | -150% | -| aistudio_gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | -| anthropic-opus-4-6 | aihubmix | — | $2.5 | $0.25 | $3.125 | 90% | -| claude-haiku-4-5 | aihubmix | — | $0.55 | $0.055 | $0.6875 | 90% | -| claude-sonnet-4-0 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | -| claude-sonnet-4-5 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | -| claude-sonnet-4-5-think | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | -| codex-mini-latest | aihubmix | — | $0.75 | $0.1875 | — | 75% | -| deepseek-v3.2 | aihubmix | — | $0.151 | $0.0151 | — | 90% | -| deepseek-v3.2-exp | aihubmix | — | $0.137 | $0.0137 | — | 90% | -| deepseek-v3.2-exp-think | aihubmix | — | $0.137 | $0.0137 | — | 90% | -| deepseek-v3.2-think | aihubmix | — | $0.151 | $0.0151 | — | 90% | -| doubao-1.5-lite-32k | aihubmix | — | $0.025 | $0.005 | — | 80% | -| doubao-1.5-pro-32k | aihubmix | — | $0.067 | $0.0134 | — | 80% | -| doubao-lite-32k | aihubmix | — | $0.03 | $0.006 | — | 80% | -| doubao-pro-32k | aihubmix | — | $0.07 | $0.014 | — | 80% | -| doubao-seed-1-6 | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-flash | aihubmix | — | $0.022 | $0.0044 | — | 80% | -| doubao-seed-1-6-flash-250615 | aihubmix | — | $0.022 | $0.0044 | — | 80% | -| doubao-seed-1-6-lite | aihubmix | — | $0.041 | $0.0082 | — | 80% | -| doubao-seed-1-6-thinking | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-thinking-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | -| doubao-seed-1-6-vision-250815 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | -| doubao-seed-1-8 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | -| gemini-2.0-flash | aihubmix | — | $0.05 | $0.0125 | — | 75% | -| gemini-2.0-flash-001 | aihubmix | — | $0.05 | $0.125 | — | -150% | -| gemini-2.0-flash-search | aihubmix | — | $0.05 | $0.0125 | — | 75% | -| gemini-2.5-flash | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-lite | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | -| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-flash-search | aihubmix | — | $0.15 | $0.015 | — | 90% | -| gemini-2.5-pro | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gemini-2.5-pro-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| glm-4.5-airx | aihubmix | — | $0.55 | $0.11 | — | 80% | -| glm-4.5-x | aihubmix | — | $1.1 | $0.22 | — | 80% | -| glm-4.6 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | -| glm-4.6v | aihubmix | — | $0.0685 | $0.0137 | — | 80% | -| glm-4.7 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | -| gpt-4.1 | aihubmix | — | $1 | $0.25 | — | 75% | -| gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | -| gpt-4.1-nano | aihubmix | — | $0.05 | $0.0125 | — | 75% | -| gpt-4o | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-2024-08-06 | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-2024-08-06-global | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-2024-11-20 | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-4o-mini | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-mini-2024-07-18 | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-mini-global | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-mini-search-preview | aihubmix | — | $0.075 | $0.0375 | — | 50% | -| gpt-4o-search-preview | aihubmix | — | $1.25 | $0.625 | — | 50% | -| gpt-5 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | -| gpt-5-nano | aihubmix | — | $0.025 | $0.0025 | — | 90% | -| gpt-5.1 | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-codex-max | aihubmix | — | $0.625 | $0.0625 | — | 90% | -| gpt-5.1-codex-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | -| gpt-5.2 | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-chat-latest | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-codex | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-high | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-low | aihubmix | — | $0.875 | $0.0875 | — | 90% | -| gpt-5.2-pro | aihubmix | — | $10.5 | $1.05 | — | 90% | -| grok-4 | aihubmix | — | $1.65 | $0.4125 | — | 75% | -| grok-4-1-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4-1-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | -| grok-4.20-beta-0309-non-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-4.20-beta-0309-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-4.20-multi-agent-0309 | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-4.20-multi-agent-beta-0309 | aihubmix | — | $1 | $0.1 | — | 90% | -| grok-code-fast-1 | aihubmix | — | $0.1 | $0.025 | — | 75% | -| kimi-k2-thinking | aihubmix | — | $0.274 | $0.0685 | — | 75% | -| kimi-k2-turbo-preview | aihubmix | — | $0.6 | $0.15 | — | 75% | -| kimi-k2.5 | aihubmix | — | $0.3 | $0.0525 | — | 82% | -| mimo-v2-flash | aihubmix | — | $0.0959 | $0.01918 | — | 80% | -| mimo-v2-omni | aihubmix | — | $0.22 | $0.044 | — | 80% | -| mimo-v2-pro | aihubmix | — | $0.55 | $0.11 | — | 80% | -| nvidia-nemotron-3-super-120b-a12b | aihubmix | — | $0.055 | $0.01375 | — | 75% | -| o1 | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-2024-12-17 | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-global | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-mini | aihubmix | — | $1.5 | $0.75 | — | 50% | -| o1-mini-2024-09-12 | aihubmix | — | $1.5 | $0.75 | — | 50% | -| o1-preview | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o1-preview-2024-09-12 | aihubmix | — | $7.5 | $3.75 | — | 50% | -| o3 | aihubmix | — | $1 | $0.25 | — | 75% | -| o3-deep-research | aihubmix | — | $5 | $1.25 | — | 75% | -| o3-global | aihubmix | — | $1 | $0.25 | — | 75% | -| o3-mini | aihubmix | — | $0.55 | $0.275 | — | 50% | -| o3-mini-global | aihubmix | — | $0.55 | $0.275 | — | 50% | -| o4-mini | aihubmix | — | $0.55 | $0.1375 | — | 75% | -| qwen-plus | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-plus-2025-04-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-plus-2025-07-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-plus-latest | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | -| qwen-turbo | aihubmix | — | $0.023 | $0.0046 | — | 80% | -| qwen-turbo-latest | aihubmix | — | $0.023 | $0.0046 | — | 80% | -| qwen3-coder-plus | aihubmix | — | $0.27 | $0.054 | — | 80% | -| qwen3-max | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | -| qwen3-max-2026-01-23 | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | -| qwen3-max-preview | aihubmix | — | $0.423 | $0.0846 | — | 80% | -| qwen3-vl-flash | aihubmix | — | $0.0103 | $0.00206 | — | 80% | -| qwen3-vl-plus | aihubmix | — | $0.0685 | $0.0137 | — | 80% | -| zai-glm-5-turbo | aihubmix | — | $0.6 | $0.12 | — | 80% | -| aion-2.0 | aion | — | $0.7999999999999999 | $0.19999999999999998 | — | 75% | -| aion-2.5 | aion | — | $1 | $0.35 | — | 65% | -| amazon-nova-2-lite | amazon-bedrock | — | $0.33 | $0.0825 | — | 75% | -| amazon-nova-lite | amazon-bedrock | — | $0.06 | $0.015 | — | 75% | -| amazon-nova-micro | amazon-bedrock | — | $0.035 | $0.00875 | — | 75% | -| amazon-nova-premier | amazon-bedrock | — | $2.5 | $0.625 | — | 75% | -| amazon-nova-pro | amazon-bedrock | — | $0.8 | $0.2 | — | 75% | -| claude-haiku-4-5-20251001 | auriko | — | $1 | $0.1 | $1.25 | 90% | -| claude-opus-4-1-20250805 | auriko | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-20250514 | auriko | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-5-20251101 | auriko | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-6 | auriko | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-7 | auriko | — | $5 | $0.5 | $6.25 | 90% | -| claude-sonnet-4-20250514 | auriko | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-5-20250929 | auriko | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-6 | auriko | — | $3 | $0.3 | $3.75 | 90% | -| deepseek-r1-0528 | auriko | — | $0.5 | $0.35 | — | 30% | -| deepseek-v3-0324 | auriko | — | $0.2 | $0.135 | — | 32% | -| deepseek-v3.1 | auriko | — | $0.21 | $0.13 | — | 38% | -| deepseek-v3.1-terminus | auriko | — | $0.27 | $0.13 | — | 52% | -| deepseek-v3.2 | auriko | — | $0.26 | $0.13 | — | 50% | -| deepseek-v4-flash | auriko | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-pro | auriko | — | $0.435 | $0.003625 | — | 99% | -| gemini-2.5-flash | auriko | — | $0.3 | $0.03 | — | 90% | -| gemini-2.5-flash-lite | auriko | — | $0.1 | $0.01 | — | 90% | -| gemini-2.5-pro | auriko | — | $1.25 | $0.125 | — | 90% | -| gemini-3-flash-preview | auriko | — | $0.5 | $0.05 | — | 90% | -| gemini-3.1-flash-lite | auriko | — | $0.25 | $0.025 | — | 90% | -| gemini-3.1-flash-lite-preview | auriko | — | $0.25 | $0.025 | — | 90% | -| gemini-3.1-pro-preview | auriko | — | $2 | $0.2 | — | 90% | -| gemini-3.1-pro-preview-customtools | auriko | — | $2 | $0.2 | — | 90% | -| gemini-flash-latest | auriko | — | $0.5 | $0.05 | — | 90% | -| gemini-flash-lite-latest | auriko | — | $0.1 | $0.01 | — | 90% | -| gemini-pro-latest | auriko | — | $2 | $0.2 | — | 90% | -| glm-4.5 | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.5-air | auriko | — | $0.2 | $0.03 | — | 85% | -| glm-4.5-airx | auriko | — | $1.1 | $0.22 | — | 80% | -| glm-4.5-x | auriko | — | $2.2 | $0.45 | — | 80% | -| glm-4.5v | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.6 | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.6v | auriko | — | $0.3 | $0.05 | — | 83% | -| glm-4.6v-flashx | auriko | — | $0.04 | $0.004 | — | 90% | -| glm-4.7 | auriko | — | $0.6 | $0.11 | — | 82% | -| glm-4.7-flashx | auriko | — | $0.07 | $0.01 | — | 86% | -| glm-5 | auriko | — | $1 | $0.2 | — | 80% | -| glm-5-turbo | auriko | — | $1.2 | $0.24 | — | 80% | -| glm-5.1 | auriko | — | $1.4 | $0.26 | — | 81% | -| glm-5v-turbo | auriko | — | $1.2 | $0.24 | — | 80% | -| gpt-4.1-2025-04-14 | auriko | — | $2 | $0.5 | — | 75% | -| gpt-4.1-mini-2025-04-14 | auriko | — | $0.4 | $0.1 | — | 75% | -| gpt-4.1-nano-2025-04-14 | auriko | — | $0.1 | $0.025 | — | 75% | -| gpt-4o-2024-08-06 | auriko | — | $2.5 | $1.25 | — | 50% | -| gpt-4o-2024-11-20 | auriko | — | $2.5 | $1.25 | — | 50% | -| gpt-4o-mini-2024-07-18 | auriko | — | $0.15 | $0.075 | — | 50% | -| gpt-5-2025-08-07 | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5-mini-2025-08-07 | auriko | — | $0.25 | $0.025 | — | 90% | -| gpt-5-nano-2025-08-07 | auriko | — | $0.05 | $0.005 | — | 90% | -| gpt-5.1-2025-11-13 | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5.1-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | -| gpt-5.2-2025-12-11 | auriko | — | $1.75 | $0.175 | — | 90% | -| gpt-5.2-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | -| gpt-5.3-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | -| gpt-5.4-2026-03-05 | auriko | — | $2.5 | $0.25 | — | 90% | -| gpt-5.4-mini-2026-03-17 | auriko | — | $0.75 | $0.075 | — | 90% | -| gpt-5.4-nano-2026-03-17 | auriko | — | $0.2 | $0.02 | — | 90% | -| gpt-5.5-2026-04-23 | auriko | — | $5 | $0.5 | — | 90% | -| gpt-oss-120b | auriko | — | $0.15 | $0.01 | — | 93% | -| gpt-oss-20b | auriko | — | $0.07 | $0.04 | — | 43% | -| grok-4.20-0309-non-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | -| grok-4.20-0309-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | -| grok-4.3 | auriko | — | $1.25 | $0.2 | — | 84% | -| hy3-preview | auriko | — | $0.066 | $0.029 | — | 56% | -| kimi-k2-0711-preview | auriko | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-0905-preview | auriko | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-thinking | auriko | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-thinking-turbo | auriko | — | $1.15 | $0.15 | — | 87% | -| kimi-k2-turbo-preview | auriko | — | $1.15 | $0.15 | — | 87% | -| kimi-k2.5 | auriko | — | $0.6 | $0.1 | — | 83% | -| kimi-k2.6 | auriko | — | $0.95 | $0.16 | — | 83% | -| mimo-v2.5 | auriko | — | $0.4 | $0.08 | — | 80% | -| mimo-v2.5-pro | auriko | — | $1 | $0.2 | — | 80% | -| minimax-m2 | auriko | — | $0.3 | — | $0.375 | — | -| minimax-m2-1 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | -| minimax-m2-1-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | -| minimax-m2-5 | auriko | — | $0.3 | $0.03 | $0.375 | 90% | -| minimax-m2-5-highspeed | auriko | — | $0.6 | $0.03 | $0.375 | 95% | -| minimax-m2-7 | auriko | — | $0.3 | — | $0.375 | — | -| minimax-m2-7-highspeed | auriko | — | $0.6 | — | $0.375 | — | -| o3-2025-04-16 | auriko | — | $2 | $0.5 | — | 75% | -| o3-mini-2025-01-31 | auriko | — | $1.1 | $0.55 | — | 50% | -| o4-mini-2025-04-16 | auriko | — | $1.1 | $0.275 | — | 75% | -| qwen-3-coder-480b-a35b-instruct | auriko | — | $0.3 | $0.1 | — | 67% | -| qwen-3-max | auriko | — | $1.2 | $0.24 | — | 80% | -| qwen-3-max-thinking | auriko | — | $1.2 | $0.24 | — | 80% | -| qwen-3-vl-235b-a22b-instruct | auriko | — | $0.2 | $0.11 | — | 45% | -| qwen-3.5-35b-a3b | auriko | — | $0.14 | $0.05 | — | 64% | -| seed-1.8 | auriko | — | $0.25 | $0.05 | — | 80% | -| seed-2.0-code | auriko | — | $0.5 | $0.1 | — | 80% | -| seed-2.0-mini | auriko | — | $0.1 | $0.02 | — | 80% | -| seed-2.0-pro | auriko | — | $0.5 | $0.1 | — | 80% | -| step-3.5-flash | auriko | — | $0.1 | $0.02 | — | 80% | -| deepseek-v3.2 | baidu | — | $0.252 | $0.0252 | — | 90% | -| deepseek-v4-flash | baidu | — | $0.126 | $0.0252 | — | 80% | -| deepseek-v4-pro | baidu | — | $1.521 | $0.126 | — | 92% | -| glm-5 | baidu | — | $0.7 | $0.14 | — | 80% | -| glm-5.1 | baidu | — | $0.98 | $0.182 | — | 81% | -| minimax-m2.5 | baidu | — | $0.27 | $0.027 | — | 90% | -| deepseek-v3-1 | baseten | — | $0.5 | $0.25 | — | 50% | -| deepseek-v4 | baseten | — | $1.74 | $0.145 | — | 92% | -| glm-4-7 | baseten | — | $0.6 | $0.12 | — | 80% | -| glm-5 | baseten | — | $0.95 | $0.2 | — | 79% | -| kimi-k2-5 | baseten | — | $0.6 | $0.12 | — | 80% | -| kimi-k2-6 | baseten | — | $1 | $0.2 | — | 80% | -| minimax-m2-5 | baseten | — | $0.3 | $0.06 | — | 80% | -| nvidia-nemotron-3-super | baseten | — | $0.3 | $0.06 | — | 80% | -| MiniMaxAI--MiniMax-M2.5-TEE | chutes | — | $0.15 | $0.075 | — | 50% | -| Qwen--Qwen3-235B-A22B-Thinking-2507 | chutes | — | $0.11 | $0.055 | — | 50% | -| Qwen--Qwen3-32B-TEE | chutes | — | $0.08 | $0.04 | — | 50% | -| Qwen--Qwen3.5-397B-A17B-TEE | chutes | — | $0.39 | $0.195 | — | 50% | -| Qwen--Qwen3.6-27B-TEE | chutes | — | $0.5 | $0.25 | — | 50% | -| deepseek-ai--DeepSeek-V3.2-TEE | chutes | — | $0.28 | $0.14 | — | 50% | -| google--gemma-4-31B-turbo-TEE | chutes | — | $0.13 | $0.065 | — | 50% | -| moonshotai--Kimi-K2.5-TEE | chutes | — | $0.44 | $0.22 | — | 50% | -| moonshotai--Kimi-K2.6-TEE | chutes | — | $0.74 | $0.37 | — | 50% | -| zai-org--GLM-5-TEE | chutes | — | $0.95 | $0.475 | — | 50% | -| zai-org--GLM-5-Turbo | chutes | — | $0.4891 | $0.24455 | — | 50% | -| zai-org--GLM-5.1-TEE | chutes | — | $1.05 | $0.525 | — | 50% | -| kimi-k2-thinking | clarifai | — | $1.5 | $0.15 | — | 90% | -| kimi-k2.5 | cloudflare | — | $0.6 | $0.1 | — | 83% | -| kimi-k2.6 | cloudflare | — | $0.95 | $0.16 | — | 83% | -| claude-4-5-sonnet | cortecs | — | $2.683 | $0.293 | — | 89% | -| claude-4-6-sonnet | cortecs | — | $2.869 | $0.287 | — | 90% | -| claude-haiku-4-5 | cortecs | — | $0.894 | $0.089 | — | 90% | -| claude-opus4-5 | cortecs | — | $4.769 | $0.477 | — | 90% | -| claude-opus4-6 | cortecs | — | $4.769 | $0.477 | — | 90% | -| claude-opus4-7 | cortecs | — | $4.769 | $0.477 | — | 90% | -| claude-sonnet-4 | cortecs | — | $2.601 | $0.26 | — | 90% | -| codestral-2508 | cortecs | — | $0.3 | $0.03 | — | 90% | -| deepseek-chat-v3.1 | cortecs | — | $0.177 | $0.044 | — | 75% | -| deepseek-r1-0528 | cortecs | — | $0.585 | $0.146 | — | 75% | -| deepseek-v3-0324 | cortecs | — | $0.266 | $0.067 | — | 75% | -| deepseek-v3.2 | cortecs | — | $0.266 | $0.067 | — | 75% | -| deepseek-v4-flash | cortecs | — | $0.133 | $0.033 | — | 75% | -| deepseek-v4-pro | cortecs | — | $1.553 | $0.388 | — | 75% | -| devstral-2512 | cortecs | — | $0.4 | $0.04 | — | 90% | -| devstral-medium-2507 | cortecs | — | $0.4 | $0.04 | — | 90% | -| devstral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | -| gemini-2.5-flash | cortecs | — | $0.268 | $0.026 | — | 90% | -| gemini-2.5-pro | cortecs | — | $1.342 | $0.217 | — | 84% | -| gemini-3.1-flash-lite | cortecs | — | $0.244 | $0.022 | — | 91% | -| glm-4.6 | cortecs | — | $0.355 | $0.089 | — | 75% | -| glm-4.7 | cortecs | — | $0.532 | $0.133 | — | 75% | -| glm-5 | cortecs | — | $0.887 | $0.222 | — | 75% | -| glm-5-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | -| glm-5.1 | cortecs | — | $1.242 | $0.311 | — | 75% | -| glm-5v-turbo | cortecs | — | $1.065 | $0.266 | — | 75% | -| gpt-4.1 | cortecs | — | $1.968 | $0.49 | — | 75% | -| gpt-4.1-mini | cortecs | — | $0.39 | $0.12 | — | 69% | -| gpt-4.1-nano | cortecs | — | $0.1 | $0.05 | — | 50% | -| gpt-4o | cortecs | — | $2.387 | $1.194 | — | 50% | -| gpt-4o-mini | cortecs | — | $0.143 | $0.073 | — | 49% | -| gpt-5 | cortecs | — | $1.234 | $0.14 | — | 89% | -| gpt-5-mini | cortecs | — | $0.25 | $0.05 | — | 80% | -| gpt-5-nano | cortecs | — | $0.054 | $0.017 | — | 69% | -| gpt-5.1 | cortecs | — | $1.234 | $0.14 | — | 89% | -| gpt-5.4 | cortecs | — | $2.601 | $0.217 | — | 92% | -| gpt-oss-120b | cortecs | — | $0.035 | $0.009 | — | 74% | -| gpt-oss-20b | cortecs | — | $0.027 | $0.007 | — | 74% | -| kimi-k2.5 | cortecs | — | $0.444 | $0.111 | — | 75% | -| kimi-k2.6 | cortecs | — | $0.694 | $0.173 | — | 75% | -| llama-3.3-70b-instruct | cortecs | — | $0.089 | $0.023 | — | 74% | -| llama-4-maverick | cortecs | — | $0.124 | $0.03 | — | 76% | -| magistral-medium-2509 | cortecs | — | $2 | $0.2 | — | 90% | -| magistral-small-2509 | cortecs | — | $0.5 | $0.05 | — | 90% | -| minimax-m2 | cortecs | — | $0.222 | $0.055 | — | 75% | -| minimax-m2.5 | cortecs | — | $0.266 | $0.027 | — | 90% | -| minimax-m2.7 | cortecs | — | $0.444 | $0.111 | — | 75% | -| ministral-14b-2512 | cortecs | — | $0.2 | $0.02 | — | 90% | -| ministral-3b-2512 | cortecs | — | $0.1 | $0.01 | — | 90% | -| ministral-8b-2512 | cortecs | — | $0.15 | $0.015 | — | 90% | -| mistral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | -| mistral-large-2512 | cortecs | — | $0.5 | $0.05 | — | 90% | -| mistral-medium-2508 | cortecs | — | $0.4 | $0.04 | — | 90% | -| mistral-medium-3.5 | cortecs | — | $1.25 | $0.125 | — | 90% | -| mistral-nemo-instruct-2407 | cortecs | — | $0.13 | $0.013 | — | 90% | -| mistral-small-2506 | cortecs | — | $0.1 | $0.01 | — | 90% | -| mistral-small-2603 | cortecs | — | $0.128 | $0.013 | — | 90% | -| nemotron-3-super-120b-a12b | cortecs | — | $0.266 | $0.067 | — | 75% | -| pixtral-large-2411 | cortecs | — | $1.8 | $0.18 | — | 90% | -| qwen-2.5-72b-instruct | cortecs | — | $0.062 | $0.016 | — | 74% | -| qwen3-235b-a22b-instruct-2507 | cortecs | — | $0.062 | $0.016 | — | 74% | -| qwen3-coder-30b-a3b-instruct | cortecs | — | $0.053 | $0.013 | — | 75% | -| qwen3-vl-235b-a22b | cortecs | — | $0.186 | $0.047 | — | 75% | -| qwen3.5-122b-a10b | cortecs | — | $0.444 | $0.111 | — | 75% | -| voxtral-small-2507 | cortecs | — | $0.1 | $0.01 | — | 90% | -| databricks-claude-haiku-4-5 | databricks | — | $1 | $0.1 | $1.25 | 90% | -| databricks-claude-opus-4-1 | databricks | — | $15 | $1.5 | $18.75 | 90% | -| databricks-claude-opus-4-5 | databricks | — | $5 | $0.5 | $6.25 | 90% | -| databricks-claude-sonnet-4 | databricks | — | $3 | $0.3 | $3.75 | 90% | -| databricks-claude-sonnet-4-5 | databricks | — | $3 | $0.3 | $3.75 | 90% | -| databricks-gemini-3-1-flash-lite | databricks | — | $0.25 | $0.02 | $0.25 | 92% | -| databricks-gemini-3-1-pro | databricks | — | $2.5 | $0.25 | $2.5 | 90% | -| databricks-gemini-3-flash | databricks | — | $0.63 | $0.06 | $0.63 | 90% | -| databricks-gpt-5 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | -| databricks-gpt-5-1 | databricks | — | $1.25 | $0.13 | $1.25 | 90% | -| databricks-gpt-5-1-codex-max | databricks | — | $1.25 | $0.13 | $1.25 | 90% | -| databricks-gpt-5-1-codex-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | -| databricks-gpt-5-2 | databricks | — | $1.75 | $0.18 | $1.75 | 90% | -| databricks-gpt-5-2-codex | databricks | — | $1.75 | $0.18 | $1.75 | 90% | -| databricks-gpt-5-4 | databricks | — | $2.5 | $0.25 | $2.5 | 90% | -| databricks-gpt-5-4-mini | databricks | — | $0.75 | $0.07 | $0.75 | 91% | -| databricks-gpt-5-4-nano | databricks | — | $0.2 | $0.02 | $0.2 | 90% | -| databricks-gpt-5-5 | databricks | — | $5 | $0.5 | $5 | 90% | -| databricks-gpt-5-mini | databricks | — | $0.25 | $0.02 | $0.25 | 92% | -| databricks-gpt-5-nano | databricks | — | $0.05 | Free | $0.05 | — | -| deepseek-r1-0528 | deepinfra | — | $0.5 | $0.35 | — | 30% | -| deepseek-v3-0324 | deepinfra | — | $0.2 | $0.135 | — | 32% | -| deepseek-v3.1 | deepinfra | — | $0.21 | $0.13 | — | 38% | -| deepseek-v3.1-terminus | deepinfra | — | $0.27 | $0.13 | — | 52% | -| deepseek-v3.2 | deepinfra | — | $0.26 | $0.13 | — | 50% | -| deepseek-v4-flash | deepinfra | — | $0.14 | $0.028 | — | 80% | -| deepseek-v4-pro | deepinfra | — | $1.74 | $0.145 | — | 92% | -| glm-4.6 | deepinfra | — | $0.43 | $0.08 | — | 81% | -| glm-4.7 | deepinfra | — | $0.4 | $0.08 | — | 80% | -| glm-4.7-flash | deepinfra | — | $0.06 | $0.01 | — | 83% | -| glm-5 | deepinfra | — | $0.6 | $0.12 | — | 80% | -| glm-5.1 | deepinfra | — | $1.05 | $0.205 | — | 80% | -| kimi-k2.5 | deepinfra | — | $0.45 | $0.07 | — | 84% | -| kimi-k2.6 | deepinfra | — | $0.75 | $0.15 | — | 80% | -| mimo-v2.5 | deepinfra | — | $0.4 | $0.08 | — | 80% | -| mimo-v2.5-pro | deepinfra | — | $1 | $0.2 | — | 80% | -| minimax-m2.5 | deepinfra | — | $0.15 | $0.03 | — | 80% | -| qwen3-235b-a22b-thinking-2507 | deepinfra | — | $0.23 | $0.2 | — | 13% | -| qwen3-coder-480b-a35b-instruct-turbo | deepinfra | — | $0.3 | $0.1 | — | 67% | -| qwen3-max | deepinfra | — | $1.2 | $0.24 | — | 80% | -| qwen3-max-thinking | deepinfra | — | $1.2 | $0.24 | — | 80% | -| qwen3-vl-235b-a22b-instruct | deepinfra | — | $0.2 | $0.11 | — | 45% | -| qwen3.5-35b-a3b | deepinfra | — | $0.14 | $0.05 | — | 64% | -| qwen3.5-397b-a17b | deepinfra | — | $0.49 | $0.3 | — | 39% | -| seed-1.8 | deepinfra | — | $0.25 | $0.05 | — | 80% | -| seed-2.0-code | deepinfra | — | $0.5 | $0.1 | — | 80% | -| seed-2.0-mini | deepinfra | — | $0.1 | $0.02 | — | 80% | -| seed-2.0-pro | deepinfra | — | $0.5 | $0.1 | — | 80% | -| step-3.5-flash | deepinfra | — | $0.1 | $0.02 | — | 80% | -| deepseek-chat | deepseek | — | $0.14 | $0.0028 | — | 98% | -| deepseek-reasoner | deepseek | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-flash | deepseek | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-pro | deepseek | — | $0.435 | $0.003625 | — | 99% | -| arcee-trinity-large-thinking | digitalocean | — | $0.25 | $0.06 | — | 76% | -| anthropic--claude-3-5-haiku-20241022 | fastrouter | — | $0.8 | $0.08 | $1 | 90% | -| anthropic--claude-4.5-sonnet | fastrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-haiku-4.5 | fastrouter | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4-20250514 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.1 | fastrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.5 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.6 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.7 | fastrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-sonnet-4-20250514 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4.6 | fastrouter | — | $3 | $0.3 | $3.75 | 90% | -| deepseek--deepseek-v3.2 | fastrouter | — | $0.27 | $0.216 | — | 20% | -| deepseek--deepseek-v4-pro | fastrouter | — | $2.1 | $0.2 | — | 90% | -| google--gemini-2.0-flash-001 | fastrouter | — | $0.1 | $0.025 | $0.1833 | 75% | -| google--gemini-2.5-flash | fastrouter | — | $0.3 | $0.075 | $0.3833 | 75% | -| google--gemini-2.5-pro | fastrouter | — | $1.25 | $0.31 | $1.625 | 75% | -| google--gemini-3-flash-preview | fastrouter | — | $0.5 | $0.05 | — | 90% | -| google--gemini-3.1-flash-lite-preview | fastrouter | — | $0.25 | $0.025 | $0.083333 | 90% | -| google--gemini-3.1-pro-preview | fastrouter | — | $2 | $0.31 | $1.625 | 84% | -| minimax--minimax-m2.1 | fastrouter | — | $0.27 | $0.03 | — | 89% | -| minimax--minimax-m2.5 | fastrouter | — | $0.3 | $0.03 | — | 90% | -| minimax--minimax-m2.5-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | -| minimax--minimax-m2.7 | fastrouter | — | $0.3 | $0.06 | — | 80% | -| minimax--minimax-m2.7-highspeed | fastrouter | — | $0.6 | $0.03 | — | 95% | -| moonshotai--kimi-k2-thinking | fastrouter | — | $0.6 | $0.15 | — | 75% | -| moonshotai--kimi-k2.6 | fastrouter | — | $0.55 | $0.15 | — | 73% | -| openai--gpt-4.1 | fastrouter | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-mini | fastrouter | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-nano | fastrouter | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4o | fastrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-08-06 | fastrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-11-20 | fastrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-mini | fastrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-mini-2024-07-18 | fastrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-realtime-preview-2024-12-17 | fastrouter | — | $5 | $2.5 | — | 50% | -| openai--gpt-4o-realtime-preview-2025-06-03 | fastrouter | — | $5 | $2.5 | — | 50% | -| openai--gpt-5 | fastrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-mini | fastrouter | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-nano | fastrouter | — | $0.05 | $0.005 | — | 90% | -| openai--gpt-5.1 | fastrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.2 | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-chat | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-codex | fastrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.4 | fastrouter | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | fastrouter | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | fastrouter | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | fastrouter | — | $5 | $0.5 | — | 90% | -| openai--gpt-realtime | fastrouter | — | $4 | $0.5 | — | 88% | -| openai--gpt-realtime-1.5 | fastrouter | — | $4 | $0.4 | — | 90% | -| openai--gpt-realtime-mini | fastrouter | — | $0.6 | $0.06 | — | 90% | -| openai--o1 | fastrouter | — | $15 | $7.5 | — | 50% | -| openai--o3 | fastrouter | — | $2 | $0.5 | — | 75% | -| openai--o3-mini | fastrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o3-mini-high | fastrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o4-mini | fastrouter | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-high | fastrouter | — | $1.1 | $0.275 | — | 75% | -| x-ai--grok-3-beta | fastrouter | — | $5 | $1.25 | — | 75% | -| x-ai--grok-3-mini-beta | fastrouter | — | $0.6 | $0.15 | — | 75% | -| x-ai--grok-4 | fastrouter | — | $3 | $0.75 | — | 75% | -| x-ai--grok-4.20-beta | fastrouter | — | $2 | $0.2 | — | 90% | -| x-ai--grok-4.3 | fastrouter | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-code-fast-1 | fastrouter | — | $0.2 | $0.02 | — | 90% | -| z-ai--glm-4.5-air | fastrouter | — | $0.2 | $0.03 | — | 85% | -| LGAI-EXAONE--K-EXAONE-236B-A23B | friendli | — | $0.2 | $0.1 | — | 50% | -| MiniMaxAI--MiniMax-M2.5 | friendli | — | $0.3 | $0.06 | — | 80% | -| deepseek-ai--DeepSeek-V3.2 | friendli | — | $0.5 | $0.25 | — | 50% | -| zai-org--GLM-5 | friendli | — | $1 | $0.5 | — | 50% | -| zai-org--GLM-5.1 | friendli | — | $1.4 | $0.26 | — | 81% | -| gemini-1.5-flash | google | — | $0.075 | $0.01875 | $0.2625 | 75% | -| gemini-1.5-flash-8b | google | — | $0.075 | $0.01875 | $0.2625 | 75% | -| gemini-1.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | -| gemini-2.0-flash | google | — | $0.1 | $0.025 | $0.35 | 75% | -| gemini-2.0-flash-lite | google | — | $0.075 | $0.01875 | $0.2625 | 75% | -| gemini-2.5-flash | google | — | $0.15 | $0.0375 | $0.525 | 75% | -| gemini-2.5-flash-lite | google | — | $0.1 | $0.025 | $0.35 | 75% | -| gemini-2.5-pro | google | — | $1.25 | $0.315 | $4.375 | 75% | -| claude-haiku-4-5 | google-vertex | — | $1 | $0.1 | — | 90% | -| claude-opus-4 | google-vertex | — | $15 | $1.5 | — | 90% | -| claude-opus-4-1 | google-vertex | — | $15 | $1.5 | — | 90% | -| claude-opus-4-5 | google-vertex | — | $5 | $0.5 | — | 90% | -| claude-opus-4-6 | google-vertex | — | $5 | $0.5 | — | 90% | -| claude-opus-4-7 | google-vertex | — | $5 | $0.5 | — | 90% | -| claude-sonnet-4 | google-vertex | — | $3 | $0.3 | — | 90% | -| claude-sonnet-4-5 | google-vertex | — | $3 | $0.3 | — | 90% | -| claude-sonnet-4-6 | google-vertex | — | $3 | $0.3 | — | 90% | -| deepseek-v3-1 | google-vertex | — | $0.6 | $0.06 | — | 90% | -| deepseek-v3-2 | google-vertex | — | $0.56 | $0.056 | — | 90% | -| gemini-2-5-flash | google-vertex | — | $0.3 | $0.03 | — | 90% | -| gemini-2-5-flash-lite | google-vertex | — | $0.1 | $0.01 | — | 90% | -| gemini-2-5-pro | google-vertex | — | $1.25 | $0.13 | — | 90% | -| gemini-3-1-flash-lite | google-vertex | — | $0.25 | $0.025 | — | 90% | -| gemini-3-flash | google-vertex | — | $0.5 | $0.05 | — | 90% | -| gemini-3-pro | google-vertex | — | $2 | $0.2 | — | 90% | -| gemma-4-26b | google-vertex | — | $0.15 | $0.015 | — | 90% | -| glm-5 | google-vertex | — | $1 | $0.1 | — | 90% | -| gpt-oss-20b | google-vertex | — | $0.07 | $0.007 | — | 90% | -| grok-4-1-fast | google-vertex | — | $0.2 | $0.05 | — | 75% | -| grok-4-20 | google-vertex | — | $2 | $0.2 | — | 90% | -| kimi-k2-thinking | google-vertex | — | $0.6 | $0.06 | — | 90% | -| minimax-m2 | google-vertex | — | $0.3 | $0.03 | — | 90% | -| qwen3-coder-480b-a35b | google-vertex | — | $0.22 | $0.022 | — | 90% | -| gpt-oss-120b | groq | — | $0.15 | $0.075 | — | 50% | -| gpt-oss-20b | groq | — | $0.075 | $0.0375 | — | 50% | -| kimi-k2-instruct-0905 | groq | — | $1 | $0.5 | — | 50% | -| deepseek--deepseek-v4-flash | hpc-ai | — | $0.14 | $0.028 | — | 80% | -| deepseek--deepseek-v4-pro | hpc-ai | — | $1.74 | $0.145 | — | 92% | -| minimax--minimax-m2.5 | hpc-ai | — | $0.3 | $0.03 | — | 90% | -| moonshotai--kimi-k2.5 | hpc-ai | — | $0.6 | $0.1 | — | 83% | -| moonshotai--kimi-k2.6 | hpc-ai | — | $0.95 | $0.16 | — | 83% | -| xiaomi--mimo-v2.5 | hpc-ai | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | hpc-ai | — | $1 | $0.2 | — | 80% | -| zai--glm-5.1 | hpc-ai | — | $1.4 | $0.26 | — | 81% | -| mercury | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-coder | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-edit | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| mercury-edit-2 | inception | — | $0.25 | $0.024999999999999998 | — | 90% | -| claude-opus-4-6 | jiekou | — | $2.75 | $0.475 | — | 83% | -| claude-opus-4-6-dd | jiekou | — | $2.85 | $0.275 | — | 90% | -| claude-opus-4-7 | jiekou | — | $4.75 | $0.475 | — | 90% | -| claude-sonnet-4-6 | jiekou | — | $1.65 | $0.285 | — | 83% | -| claude-sonnet-4-6-dd | jiekou | — | $4.75 | $0.165 | — | 97% | -| doubao-seed-1-8-251228 | jiekou | — | $0.11 | $0.0221 | — | 80% | -| gemini-2.5-flash | jiekou | — | $0.095 | $0.0285 | — | 70% | -| gemini-2.5-flash-lite | jiekou | — | $1.1875 | $0.0095 | — | 99% | -| gemini-2.5-pro | jiekou | — | $0.285 | $0.1187 | — | 58% | -| gemini-3-flash-preview | jiekou | — | $0.095 | $0.0475 | — | 50% | -| gemini-3.1-flash-lite-preview | jiekou | — | $1.9 | $0.0237 | — | 99% | -| gemini-3.1-pro-preview | jiekou | — | $0.475 | $0.19 | — | 60% | -| gpt-4.1 | jiekou | — | $9.5 | $0.5 | — | 95% | -| gpt-4.1-mini | jiekou | — | $0.1 | $0.1 | — | 0% | -| gpt-4o | jiekou | — | $1.9 | $1.1875 | — | 38% | -| gpt-5-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | -| gpt-5-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | -| gpt-5-nano | jiekou | — | $0.2375 | $0.0047 | — | 98% | -| gpt-5.1-chat-latest | jiekou | — | $1.1875 | $0.1187 | — | 90% | -| gpt-5.1-codex | jiekou | — | $1.1875 | $0.1187 | — | 90% | -| gpt-5.1-codex-max | jiekou | — | $0.2375 | $0.1187 | — | 50% | -| gpt-5.1-codex-mini | jiekou | — | $1.1875 | $0.0237 | — | 98% | -| gpt-5.2-chat-latest | jiekou | — | $1.1875 | $0.1662 | — | 86% | -| gpt-5.2-codex | jiekou | — | $1.6625 | $0.175 | — | 89% | -| gpt-5.3-chat-latest | jiekou | — | $1.6625 | $0.1662 | — | 90% | -| gpt-5.3-codex | jiekou | — | $1.75 | $0.1662 | — | 91% | -| gpt-5.4 | jiekou | — | $1.6625 | $0.2375 | — | 86% | -| gpt-5.4-nano | jiekou | — | $0.7125 | $0.019 | — | 97% | -| gpt-5.5 | jiekou | — | $0.19 | $0.5 | — | -163% | -| gpt-5.5-light | jiekou | — | $5 | $0.025 | — | 100% | -| grok-code-fast-1 | jiekou | — | $0.19 | $0.019 | — | 90% | -| minimax-minimax-m2.1 | jiekou | — | $0.55 | $0.03 | — | 95% | -| moonshotai-kimi-k2.5 | jiekou | — | $0.6 | $0.1 | — | 83% | -| o3 | jiekou | — | $1.045 | $0.475 | — | 55% | -| o3-mini | jiekou | — | $1.045 | $0.5225 | — | 50% | -| zai-org-glm-4.7-flash | jiekou | — | $0.6 | $0.01 | — | 98% | -| claude-3-7-sonnet-20250219 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-3-opus | llmgateway | — | $15 | $1.5 | $18.75 | 90% | -| claude-haiku-4-5 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | -| claude-haiku-4-5-20251001 | llmgateway | — | $1 | $0.1 | $1.25 | 90% | -| claude-opus-4-1-20250805 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-20250514 | llmgateway | — | $15 | $1.5 | $18.75 | 90% | -| claude-opus-4-5-20251101 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-6 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | -| claude-opus-4-7 | llmgateway | — | $5 | $0.5 | $6.25 | 90% | -| claude-sonnet-4-20250514 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-5 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-5-20250929 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| claude-sonnet-4-6 | llmgateway | — | $3 | $0.3 | $3.75 | 90% | -| deepseek-v3.1 | llmgateway | — | $0.56 | $0.112 | — | 80% | -| deepseek-v3.2 | llmgateway | — | $0.28 | $0.028 | — | 90% | -| deepseek-v4-flash | llmgateway | — | $0.14 | $0.0028 | — | 98% | -| deepseek-v4-pro | llmgateway | — | $0.435 | $0.003625 | — | 99% | -| gemini-2.5-flash | llmgateway | — | $0.3 | $0.03 | — | 90% | -| gemini-2.5-flash-image | llmgateway | — | $0.3 | $0.03 | — | 90% | -| gemini-2.5-flash-lite | llmgateway | — | $0.1 | $0.01 | — | 90% | -| gemini-2.5-pro | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gemini-3-flash-preview | llmgateway | — | $0.5 | $0.05 | — | 90% | -| gemini-3-pro-image-preview | llmgateway | — | $2 | $0.2 | — | 90% | -| gemini-3.1-flash-lite | llmgateway | — | $0.25 | $0.025 | $0.08333 | 90% | -| gemini-3.1-pro-preview | llmgateway | — | $2 | $0.2 | — | 90% | -| gemini-3.5-flash | llmgateway | — | $1.5 | $0.15 | $0.08333 | 90% | -| gemini-pro-latest | llmgateway | — | $2 | $0.2 | — | 90% | -| glm-4.5 | llmgateway | — | $0.6 | $0.11 | — | 82% | -| glm-4.5-air | llmgateway | — | $0.2 | $0.03 | — | 85% | -| glm-4.5-airx | llmgateway | — | $1.1 | $0.22 | — | 80% | -| glm-4.5-x | llmgateway | — | $2.2 | $0.45 | — | 80% | -| glm-4.5v | llmgateway | — | $0.6 | $0.11 | — | 82% | -| glm-4.6v | llmgateway | — | $0.3 | $0.05 | — | 83% | -| glm-4.6v-flashx | llmgateway | — | $0.04 | $0.004 | — | 90% | -| glm-4.7 | llmgateway | — | $0.6 | $0.11 | — | 82% | -| glm-4.7-flashx | llmgateway | — | $0.07 | $0.01 | — | 86% | -| glm-5.1 | llmgateway | — | $1.4 | $0.26 | — | 81% | -| gpt-4.1 | llmgateway | — | $2 | $0.5 | — | 75% | -| gpt-4.1-mini | llmgateway | — | $0.4 | $0.1 | — | 75% | -| gpt-4.1-nano | llmgateway | — | $0.1 | $0.025 | — | 75% | -| gpt-4o | llmgateway | — | $2.5 | $1.25 | — | 50% | -| gpt-5 | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gpt-5-chat-latest | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gpt-5-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | -| gpt-5-nano | llmgateway | — | $0.05 | $0.005 | — | 90% | -| gpt-5.1 | llmgateway | — | $1.25 | $0.125 | — | 90% | -| gpt-5.1-codex-mini | llmgateway | — | $0.25 | $0.025 | — | 90% | -| gpt-5.2 | llmgateway | — | $1.75 | $0.175 | — | 90% | -| gpt-5.2-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | -| gpt-5.3-codex | llmgateway | — | $1.75 | $0.175 | — | 90% | -| gpt-5.4 | llmgateway | — | $2.5 | $0.25 | — | 90% | -| gpt-5.4-mini | llmgateway | — | $0.75 | $0.075 | — | 90% | -| gpt-5.4-nano | llmgateway | — | $0.2 | $0.02 | — | 90% | -| gpt-5.5 | llmgateway | — | $5 | $0.5 | — | 90% | -| grok-4 | llmgateway | — | $3 | $0.75 | — | 75% | -| grok-4-20-beta-0309-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-20-beta-0309-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-20-non-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-20-reasoning | llmgateway | — | $2 | $0.2 | — | 90% | -| grok-4-3 | llmgateway | — | $1.25 | $0.3125 | — | 75% | -| kimi-k2-thinking | llmgateway | — | $0.6 | $0.15 | — | 75% | -| kimi-k2-thinking-turbo | llmgateway | — | $1.15 | $0.15 | — | 87% | -| kimi-k2.5 | llmgateway | — | $0.6 | $0.1 | — | 83% | -| kimi-k2.6 | llmgateway | — | $0.95 | $0.16 | — | 83% | -| mimo-v2-flash | llmgateway | — | $0.1 | $0.01 | — | 90% | -| mimo-v2-omni | llmgateway | — | $0.4 | $0.08 | — | 80% | -| mimo-v2-pro | llmgateway | — | $1 | $0.2 | — | 80% | -| mimo-v2.5 | llmgateway | — | $0.4 | $0.08 | — | 80% | -| mimo-v2.5-pro | llmgateway | — | $1 | $0.2 | — | 80% | -| minimax-m2 | llmgateway | — | $0.2 | $0.03 | — | 85% | -| minimax-m2.5-highspeed | llmgateway | — | $0.6 | $0.03 | — | 95% | -| minimax-m2.7 | llmgateway | — | $0.3 | $0.06 | — | 80% | -| minimax-m2.7-highspeed | llmgateway | — | $0.6 | $0.06 | — | 90% | -| o1 | llmgateway | — | $15 | $7.5 | — | 50% | -| o3 | llmgateway | — | $2 | $0.5 | — | 75% | -| o3-mini | llmgateway | — | $1.1 | $0.55 | — | 50% | -| o4-mini | llmgateway | — | $1.1 | $0.275 | — | 75% | -| qwen-flash | llmgateway | — | $0.05 | $0.01 | $0.0625 | 80% | -| qwen-plus | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | -| qwen-plus-latest | llmgateway | — | $0.4 | $0.08 | $0.5 | 80% | -| qwen3-coder-flash | llmgateway | — | $0.3 | $0.06 | $0.375 | 80% | -| qwen3-coder-next | llmgateway | — | $0.108 | $0.06 | — | 44% | -| qwen3-coder-plus | llmgateway | — | $6 | $1.2 | $7.5 | 80% | -| qwen3-max | llmgateway | — | $3 | $0.6 | $3.75 | 80% | -| qwen3-max-2026-01-23 | llmgateway | — | $1.2 | $0.24 | $1.5 | 80% | -| qwen3-vl-flash | llmgateway | — | $0.05 | $0.01 | — | 80% | -| qwen3-vl-plus | llmgateway | — | $0.2 | $0.04 | $0.25 | 80% | -| qwen3.6-max-preview | llmgateway | — | $1.3 | $0.13 | — | 90% | -| qwen3.6-plus | llmgateway | — | $0.5 | $0.05 | — | 90% | -| seed-1-6-250615 | llmgateway | — | $0.25 | $0.05 | — | 80% | -| seed-1-6-250915 | llmgateway | — | $0.25 | $0.05 | — | 80% | -| seed-1-6-flash-250715 | llmgateway | — | $0.07 | $0.015 | — | 79% | -| seed-1-8-251228 | llmgateway | — | $0.25 | $0.05 | — | 80% | -| aion-labs--aion-2.0 | martian | — | $0.8 | $0.2 | — | 75% | -| alibaba--tongyi-deepresearch-30b-a3b | martian | — | $0.09 | $0.09 | — | 0% | -| amazon--nova-premier-v1 | martian | — | $2.5 | $0.625 | — | 75% | -| anthropic--claude-haiku-4-5 | martian | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-haiku-4-5-20251001 | martian | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4-0 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-1 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-1-20250805 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-20250514 | martian | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-5 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-5-20251101 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-6 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-7 | martian | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-sonnet-4-0 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-20250514 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-5 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-5-20250929 | martian | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-6 | martian | — | $3 | $0.3 | $3.75 | 90% | -| arcee-ai--trinity-large-thinking | martian | — | $0.22 | $0.06 | — | 73% | -| bytedance--ui-tars-1.5-7b | martian | — | $0.1 | $0.1 | — | 0% | -| deepinfra--deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | -| deepinfra--deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | -| deepinfra--deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | -| deepinfra--deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | -| deepinfra--deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | -| deepinfra--google--gemini-2.5-flash | martian | — | $0.3 | $0.03 | $0.083333 | 90% | -| deepinfra--google--gemini-2.5-pro | martian | — | $1.25 | $0.125 | $0.375 | 90% | -| deepinfra--z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | -| deepinfra--z-ai--glm-4.7-flash | martian | — | $0.06 | $0.01 | — | 83% | -| deepinfra--z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | -| deepinfra--z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | -| deepseek--deepseek-chat-v3-0324 | martian | — | $0.2 | $0.135 | — | 32% | -| deepseek--deepseek-chat-v3.1 | martian | — | $0.21 | $0.13 | — | 38% | -| deepseek--deepseek-r1-0528 | martian | — | $0.5 | $0.35 | — | 30% | -| deepseek--deepseek-v3.1-terminus | martian | — | $0.27 | $0.13 | — | 52% | -| deepseek--deepseek-v3.2 | martian | — | $0.252 | $0.0252 | — | 90% | -| deepseek--deepseek-v3.2-speciale | martian | — | $0.287 | $0.058 | — | 80% | -| deepseek--deepseek-v4-flash | martian | — | $0.14 | $0.0028 | — | 98% | -| deepseek--deepseek-v4-pro | martian | — | $0.435 | $0.003625 | — | 99% | -| ibm-granite--granite-4.1-8b | martian | — | $0.05 | $0.05 | — | 0% | -| inception--mercury-2 | martian | — | $0.25 | $0.025 | — | 90% | -| inclusionai--ling-2.6-1t | martian | — | $0.3 | $0.06 | — | 80% | -| inclusionai--ling-2.6-flash | martian | — | $0.08 | $0.016 | — | 80% | -| kwaipilot--kat-coder-pro-v2 | martian | — | $0.3 | $0.06 | — | 80% | -| microsoft--phi-4-mini-instruct | martian | — | $0.08 | $0.08 | — | 0% | -| minimax--minimax-m2 | martian | — | $0.255 | $0.03 | — | 88% | -| minimax--minimax-m2-her | martian | — | $0.3 | $0.03 | — | 90% | -| minimax--minimax-m2.1 | martian | — | $0.29 | $0.03 | — | 90% | -| mistralai--codestral-2508 | martian | — | $0.3 | $0.03 | — | 90% | -| mistralai--devstral-medium | martian | — | $0.4 | $0.04 | — | 90% | -| mistralai--devstral-small | martian | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-14b-2512 | martian | — | $0.2 | $0.02 | — | 90% | -| mistralai--ministral-3b-2512 | martian | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-8b-2512 | martian | — | $0.15 | $0.015 | — | 90% | -| mistralai--mistral-large | martian | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2407 | martian | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2411 | martian | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2512 | martian | — | $0.5 | $0.05 | — | 90% | -| mistralai--mistral-medium-3 | martian | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-medium-3.1 | martian | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-saba | martian | — | $0.2 | $0.02 | — | 90% | -| mistralai--mistral-small-2603 | martian | — | $0.15 | $0.015 | — | 90% | -| mistralai--mixtral-8x22b-instruct | martian | — | $2 | $0.2 | — | 90% | -| mistralai--pixtral-large-2411 | martian | — | $2 | $0.2 | — | 90% | -| mistralai--voxtral-small-24b-2507 | martian | — | $0.1 | $0.01 | — | 90% | -| moonshotai--kimi-k2-thinking | martian | — | $0.6 | $0.15 | — | 75% | -| moonshotai--kimi-k2.5 | martian | — | $0.4 | $0.05 | — | 88% | -| moonshotai--kimi-k2.6 | martian | — | $0.74 | $0.25 | — | 66% | -| openai--gpt-4.1 | martian | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-2025-04-14 | martian | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-mini | martian | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-mini-2025-04-14 | martian | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-nano | martian | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4.1-nano-2025-04-14 | martian | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4o-2024-08-06 | martian | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-11-20 | martian | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-mini | martian | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-mini-2024-07-18 | martian | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-5 | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-2025-08-07 | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-codex | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-mini | martian | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-mini-2025-08-07 | martian | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-nano | martian | — | $0.05 | $0.01 | — | 80% | -| openai--gpt-5-nano-2025-08-07 | martian | — | $0.05 | $0.01 | — | 80% | -| openai--gpt-5.1 | martian | — | $1.25 | $0.13 | — | 90% | -| openai--gpt-5.1-2025-11-13 | martian | — | $1.25 | $0.13 | — | 90% | -| openai--gpt-5.1-codex | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-max | martian | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-mini | martian | — | $0.25 | $0.03 | — | 88% | -| openai--gpt-5.2 | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-2025-12-11 | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-codex | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-codex | martian | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.4 | martian | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-2026-03-05 | martian | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | martian | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-mini-2026-03-17 | martian | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | martian | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.4-nano-2026-03-17 | martian | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | martian | — | $5 | $0.5 | — | 90% | -| openai--gpt-5.5-2026-04-23 | martian | — | $5 | $0.5 | — | 90% | -| openai--o1 | martian | — | $15 | $7.5 | — | 50% | -| openai--o1-2024-12-17 | martian | — | $15 | $7.5 | — | 50% | -| openai--o3 | martian | — | $2 | $0.5 | — | 75% | -| openai--o3-2025-04-16 | martian | — | $2 | $0.5 | — | 75% | -| openai--o3-mini | martian | — | $1.1 | $0.55 | — | 50% | -| openai--o3-mini-2025-01-31 | martian | — | $1.1 | $0.55 | — | 50% | -| openai--o4-mini | martian | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-2025-04-16 | martian | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-deep-research | martian | — | $2 | $0.5 | — | 75% | -| openai--o4-mini-deep-research-2025-06-26 | martian | — | $2 | $0.5 | — | 75% | -| qwen--qwen-plus | martian | — | $0.26 | $0.052 | $0.325 | 80% | -| qwen--qwen-plus-2025-07-28 | martian | — | $0.26 | — | $0.325 | — | -| qwen--qwen-plus-2025-07-28--thinking | martian | — | $0.26 | — | $0.325 | — | -| qwen--qwen3-30b-a3b-thinking-2507 | martian | — | $0.08 | $0.08 | — | 0% | -| qwen--qwen3-8b | martian | — | $0.05 | $0.05 | — | 0% | -| qwen--qwen3-coder-flash | martian | — | $0.195 | $0.039 | $0.24375 | 80% | -| qwen--qwen3-coder-next | martian | — | $0.11 | $0.07 | — | 36% | -| qwen--qwen3-coder-plus | martian | — | $0.65 | $0.13 | $0.8125 | 80% | -| qwen--qwen3-max | martian | — | $0.78 | $0.156 | $0.975 | 80% | -| qwen--qwen3-vl-235b-a22b-instruct | martian | — | $0.2 | $0.11 | — | 45% | -| qwen--qwen3.5-35b-a3b | martian | — | $0.14 | $0.05 | — | 64% | -| qwen--qwen3.5-flash-02-23 | martian | — | $0.065 | — | $0.08125 | — | -| qwen--qwen3.5-plus-02-15 | martian | — | $0.26 | — | $0.325 | — | -| qwen--qwen3.6-35b-a3b | martian | — | $0.15 | $0.05 | — | 67% | -| qwen--qwen3.6-flash | martian | — | $0.25 | — | $0.3125 | — | -| qwen--qwen3.6-max-preview | martian | — | $1.04 | — | $1.3 | — | -| qwen--qwen3.6-plus | martian | — | $0.325 | — | $0.40625 | — | -| tencent--hy3-preview | martian | — | $0.066 | $0.029 | — | 56% | -| thedrummer--cydonia-24b-v4.1 | martian | — | $0.3 | $0.15 | — | 50% | -| thedrummer--skyfall-36b-v2 | martian | — | $0.55 | $0.25 | — | 55% | -| x-ai--grok-3 | martian | — | $3 | $0.75 | — | 75% | -| x-ai--grok-3-mini | martian | — | $0.3 | $0.075 | — | 75% | -| x-ai--grok-4 | martian | — | $3 | $0.75 | — | 75% | -| x-ai--grok-4.20-multi-agent | martian | — | $2 | $0.2 | — | 90% | -| x-ai--grok-4.3 | martian | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-code-fast-1 | martian | — | $0.2 | $0.02 | — | 90% | -| xiaomi--mimo-v2-flash | martian | — | $0.1 | $0.01 | — | 90% | -| xiaomi--mimo-v2-omni | martian | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2-pro | martian | — | $1 | $0.2 | — | 80% | -| xiaomi--mimo-v2.5 | martian | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | martian | — | $1 | $0.2 | — | 80% | -| z-ai--glm-4.5 | martian | — | $0.6 | $0.11 | — | 82% | -| z-ai--glm-4.5-air | martian | — | $0.13 | $0.025 | — | 81% | -| z-ai--glm-4.7 | martian | — | $0.4 | $0.08 | — | 80% | -| z-ai--glm-5 | martian | — | $0.6 | $0.12 | — | 80% | -| z-ai--glm-5.1 | martian | — | $0.98 | $0.182 | — | 81% | -| MiniMax-M2 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | -| MiniMax-M2.1 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | -| MiniMax-M2.1-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | -| MiniMax-M2.5 | minimax | — | $2.1 | $0.21 | $2.625 | 90% | -| MiniMax-M2.5-highspeed | minimax | — | $4.2 | $0.21 | $2.625 | 95% | -| MiniMax-M2.7 | minimax | — | $2.1 | $0.42 | $2.625 | 80% | -| MiniMax-M2.7-highspeed | minimax | — | $4.2 | $0.42 | $2.625 | 90% | -| kimi-k2-0711-preview | moonshotai | — | $4 | $1 | — | 75% | -| kimi-k2-0905-preview | moonshotai | — | $4 | $1 | — | 75% | -| kimi-k2-thinking | moonshotai | — | $4 | $1 | — | 75% | -| kimi-k2-thinking-turbo | moonshotai | — | $8 | $1 | — | 88% | -| kimi-k2-turbo-preview | moonshotai | — | $8 | $1 | — | 88% | -| kimi-k2.5 | moonshotai | — | $4 | $0.7 | — | 82% | -| kimi-k2.6 | moonshotai | — | $6.5 | $1.1 | — | 83% | -| kimi-k2.6-long | moonshotai | — | $6.5 | $1.1 | — | 83% | -| kimi-vl-a3b | moonshotai | — | $4 | $0.7 | — | 82% | -| kimi-vl-a3b-thinking | moonshotai | — | $4 | $0.7 | — | 82% | -| aion-labs--aion-2.5 | nanogpt | — | $1 | $0.35 | — | 65% | -| alibaba--qwen3.6-flash | nanogpt | — | $0.19 | $0.02 | $0.24 | 89% | -| deepseek--deepseek-latest | nanogpt | — | $1.1 | $0.11 | — | 90% | -| deepseek--deepseek-v4-flash | nanogpt | — | $0.14 | $0.028 | — | 80% | -| deepseek--deepseek-v4-flash--thinking | nanogpt | — | $0.14 | $0.028 | — | 80% | -| deepseek--deepseek-v4-pro | nanogpt | — | $1.1 | $0.11 | — | 90% | -| deepseek--deepseek-v4-pro--thinking | nanogpt | — | $1.1 | $0.11 | — | 90% | -| deepseek--deepseek-v4-pro-cheaper | nanogpt | — | $0.435 | $0.0435 | — | 90% | -| deepseek--deepseek-v4-pro-cheaper--thinking | nanogpt | — | $0.435 | $0.0435 | — | 90% | -| ernie-5.1 | nanogpt | — | $0.75 | $0.75 | — | 0% | -| ernie-5.1--thinking | nanogpt | — | $0.75 | $0.75 | — | 0% | -| google--gemini-3-flash-preview | nanogpt | — | $0.5 | — | $0.08333 | — | -| google--gemini-3.1-flash-lite | nanogpt | — | $0.25 | — | $0.08333 | — | -| google--gemini-3.1-pro-preview | nanogpt | — | $2 | — | $0.375 | — | -| google--gemini-3.1-pro-preview-customtools | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-pro-preview-high | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-pro-preview-low | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.5-flash | nanogpt | — | $1.5 | — | $0.08333 | — | -| google--gemini-flash-lite-latest | nanogpt | — | $0.25 | $0.025 | $0.08333 | 90% | -| google--gemini-pro-latest | nanogpt | — | $2 | $0.2 | $0.375 | 90% | -| ibm-granite--granite-4.1-8b | nanogpt | — | $0.05 | $0.05 | — | 0% | -| inclusionai--ling-2.6-1t | nanogpt | — | $0.3 | $0.06 | — | 80% | -| mercury-2 | nanogpt | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5.4 | nanogpt | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | nanogpt | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | nanogpt | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | nanogpt | — | $5 | $0.5 | — | 90% | -| openai--gpt-chat-latest | nanogpt | — | $5 | $0.5 | — | 90% | -| openai--gpt-latest | nanogpt | — | $5 | $0.5 | — | 90% | -| qwen-3.6-plus | nanogpt | — | $0.325 | $0.0325 | $0.40625 | 90% | -| sarvam-105b | nanogpt | — | $0.045 | $0.028 | — | 38% | -| sarvam-30b | nanogpt | — | $0.028 | $0.017 | — | 39% | -| tencent--hy3-preview | nanogpt | — | $0.066 | $0.029 | — | 56% | -| upstage--solar-pro-3 | nanogpt | — | $0.15 | $0.015 | — | 90% | -| x-ai--grok-4.3 | nanogpt | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-latest | nanogpt | — | $1.25 | $0.2 | — | 84% | -| xiaomi--mimo-v2-omni | nanogpt | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2-pro | nanogpt | — | $1 | $0.2 | — | 80% | -| xiaomi--mimo-v2.5 | nanogpt | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | nanogpt | — | $0.6 | $0.12 | — | 80% | -| z-ai--glm-5-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | -| z-ai--glm-5v-turbo | nanogpt | — | $1.2 | $0.24 | — | 80% | -| z-ai--glm-5v-turbo--thinking | nanogpt | — | $1.2 | $0.24 | — | 80% | -| zai-org--glm-4.7-original | nanogpt | — | $0.6 | $0.11 | — | 82% | -| zai-org--glm-4.7-original--thinking | nanogpt | — | $0.6 | $0.11 | — | 82% | -| gpt-4.1 | openai | — | $2 | $0.5 | — | 75% | -| gpt-4.1-mini | openai | — | $0.4 | $0.1 | — | 75% | -| gpt-4.1-nano | openai | — | $0.1 | $0.025 | — | 75% | -| gpt-4o | openai | — | $2.5 | $1.25 | — | 50% | -| gpt-4o-mini | openai | — | $0.15 | $0.075 | — | 50% | -| o1 | openai | — | $15 | $7.5 | — | 50% | -| o1-mini | openai | — | $1.5 | $0.75 | — | 50% | -| o3 | openai | — | $10 | $2.5 | — | 75% | -| o3-mini | openai | — | $1.1 | $0.55 | — | 50% | -| o4-mini | openai | — | $1.1 | $0.275 | — | 75% | -| aion-labs--aion-2.0 | openrouter | — | $0.8 | $0.2 | — | 75% | -| alibaba--tongyi-deepresearch-30b-a3b | openrouter | — | $0.09 | $0.09 | — | 0% | -| amazon--nova-premier-v1 | openrouter | — | $2.5 | $0.625 | — | 75% | -| anthropic--claude-3-haiku | openrouter | — | $0.25 | $0.03 | $0.3 | 88% | -| anthropic--claude-3.5-haiku | openrouter | — | $0.8 | $0.08 | $1 | 90% | -| anthropic--claude-haiku-4.5 | openrouter | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4 | openrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.1 | openrouter | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4.5 | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.6 | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.6-fast | openrouter | — | $30 | $3 | $37.5 | 90% | -| anthropic--claude-opus-4.7 | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4.7-fast | openrouter | — | $30 | $3 | $37.5 | 90% | -| anthropic--claude-sonnet-4 | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4.5 | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4.6 | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| arcee-ai--trinity-large-thinking | openrouter | — | $0.22 | $0.06 | — | 73% | -| bytedance--ui-tars-1.5-7b | openrouter | — | $0.1 | $0.1 | — | 0% | -| deepseek--deepseek-chat-v3-0324 | openrouter | — | $0.2 | $0.135 | — | 32% | -| deepseek--deepseek-chat-v3.1 | openrouter | — | $0.21 | $0.13 | — | 38% | -| deepseek--deepseek-r1-0528 | openrouter | — | $0.5 | $0.35 | — | 30% | -| deepseek--deepseek-v3.1-terminus | openrouter | — | $0.27 | $0.13 | — | 52% | -| deepseek--deepseek-v3.2 | openrouter | — | $0.252 | $0.0252 | — | 90% | -| deepseek--deepseek-v3.2-speciale | openrouter | — | $0.287 | $0.058 | — | 80% | -| deepseek--deepseek-v4-flash | openrouter | — | $0.112 | $0.022 | — | 80% | -| deepseek--deepseek-v4-pro | openrouter | — | $0.435 | $0.003625 | — | 99% | -| google--gemini-2.0-flash-001 | openrouter | — | $0.1 | $0.025 | $0.083333 | 75% | -| google--gemini-2.5-flash | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | -| google--gemini-2.5-flash-image | openrouter | — | $0.3 | $0.03 | $0.083333 | 90% | -| google--gemini-2.5-flash-lite | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | -| google--gemini-2.5-flash-lite-preview-09-2025 | openrouter | — | $0.1 | $0.01 | $0.083333 | 90% | -| google--gemini-2.5-pro | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | -| google--gemini-2.5-pro-preview | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | -| google--gemini-2.5-pro-preview-05-06 | openrouter | — | $1.25 | $0.125 | $0.375 | 90% | -| google--gemini-3-flash-preview | openrouter | — | $0.5 | $0.05 | $0.083333 | 90% | -| google--gemini-3-pro-image-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-flash-lite | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | -| google--gemini-3.1-flash-lite-preview | openrouter | — | $0.25 | $0.025 | $0.083333 | 90% | -| google--gemini-3.1-pro-preview | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.1-pro-preview-customtools | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| google--gemini-3.5-flash | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | -| ibm-granite--granite-4.1-8b | openrouter | — | $0.05 | $0.05 | — | 0% | -| inception--mercury-2 | openrouter | — | $0.25 | $0.025 | — | 90% | -| inclusionai--ling-2.6-1t | openrouter | — | $0.3 | $0.06 | — | 80% | -| inclusionai--ling-2.6-flash | openrouter | — | $0.01 | $0.002 | — | 80% | -| inclusionai--ring-2.6-1t | openrouter | — | $0.075 | $0.015 | — | 80% | -| kwaipilot--kat-coder-pro-v2 | openrouter | — | $0.3 | $0.06 | — | 80% | -| microsoft--phi-4-mini-instruct | openrouter | — | $0.08 | $0.08 | — | 0% | -| minimax--minimax-m2 | openrouter | — | $0.255 | $0.03 | — | 88% | -| minimax--minimax-m2-her | openrouter | — | $0.3 | $0.03 | — | 90% | -| minimax--minimax-m2.1 | openrouter | — | $0.29 | $0.03 | — | 90% | -| mistralai--codestral-2508 | openrouter | — | $0.3 | $0.03 | — | 90% | -| mistralai--devstral-2512 | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--devstral-medium | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--devstral-small | openrouter | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-14b-2512 | openrouter | — | $0.2 | $0.02 | — | 90% | -| mistralai--ministral-3b-2512 | openrouter | — | $0.1 | $0.01 | — | 90% | -| mistralai--ministral-8b-2512 | openrouter | — | $0.15 | $0.015 | — | 90% | -| mistralai--mistral-large | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2407 | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--mistral-large-2512 | openrouter | — | $0.5 | $0.05 | — | 90% | -| mistralai--mistral-medium-3 | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-medium-3.1 | openrouter | — | $0.4 | $0.04 | — | 90% | -| mistralai--mistral-saba | openrouter | — | $0.2 | $0.02 | — | 90% | -| mistralai--mistral-small-2603 | openrouter | — | $0.15 | $0.015 | — | 90% | -| mistralai--mixtral-8x22b-instruct | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--pixtral-large-2411 | openrouter | — | $2 | $0.2 | — | 90% | -| mistralai--voxtral-small-24b-2507 | openrouter | — | $0.1 | $0.01 | — | 90% | -| moonshotai--kimi-k2.5 | openrouter | — | $0.4 | $0.09 | — | 78% | -| moonshotai--kimi-k2.6 | openrouter | — | $0.73 | $0.25 | — | 66% | -| openai--gpt-4.1 | openrouter | — | $2 | $0.5 | — | 75% | -| openai--gpt-4.1-mini | openrouter | — | $0.4 | $0.1 | — | 75% | -| openai--gpt-4.1-nano | openrouter | — | $0.1 | $0.025 | — | 75% | -| openai--gpt-4o-2024-08-06 | openrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-2024-11-20 | openrouter | — | $2.5 | $1.25 | — | 50% | -| openai--gpt-4o-mini | openrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-4o-mini-2024-07-18 | openrouter | — | $0.15 | $0.075 | — | 50% | -| openai--gpt-5 | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-chat | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-codex | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5-image | openrouter | — | $10 | $1.25 | — | 88% | -| openai--gpt-5-image-mini | openrouter | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5-mini | openrouter | — | $0.25 | $0.025 | — | 90% | -| openai--gpt-5-nano | openrouter | — | $0.05 | $0.01 | — | 80% | -| openai--gpt-5.1 | openrouter | — | $1.25 | $0.13 | — | 90% | -| openai--gpt-5.1-chat | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-max | openrouter | — | $1.25 | $0.125 | — | 90% | -| openai--gpt-5.1-codex-mini | openrouter | — | $0.25 | $0.03 | — | 88% | -| openai--gpt-5.2 | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-chat | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.2-codex | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-chat | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.3-codex | openrouter | — | $1.75 | $0.175 | — | 90% | -| openai--gpt-5.4 | openrouter | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-image-2 | openrouter | — | $8 | $2 | — | 75% | -| openai--gpt-5.4-mini | openrouter | — | $0.75 | $0.075 | — | 90% | -| openai--gpt-5.4-nano | openrouter | — | $0.2 | $0.02 | — | 90% | -| openai--gpt-5.5 | openrouter | — | $5 | $0.5 | — | 90% | -| openai--gpt-chat-latest | openrouter | — | $5 | $0.5 | — | 90% | -| openai--gpt-oss-safeguard-20b | openrouter | — | $0.075 | $0.037 | — | 51% | -| openai--o1 | openrouter | — | $15 | $7.5 | — | 50% | -| openai--o3 | openrouter | — | $2 | $0.5 | — | 75% | -| openai--o3-deep-research | openrouter | — | $10 | $2.5 | — | 75% | -| openai--o3-mini | openrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o3-mini-high | openrouter | — | $1.1 | $0.55 | — | 50% | -| openai--o4-mini | openrouter | — | $1.1 | $0.275 | — | 75% | -| openai--o4-mini-deep-research | openrouter | — | $2 | $0.5 | — | 75% | -| openai--o4-mini-high | openrouter | — | $1.1 | $0.275 | — | 75% | -| qwen--qwen-plus | openrouter | — | $0.26 | $0.052 | $0.325 | 80% | -| qwen--qwen-plus-2025-07-28 | openrouter | — | $0.26 | — | $0.325 | — | -| qwen--qwen-plus-2025-07-28--thinking | openrouter | — | $0.26 | — | $0.325 | — | -| qwen--qwen3-30b-a3b-thinking-2507 | openrouter | — | $0.08 | $0.08 | — | 0% | -| qwen--qwen3-8b | openrouter | — | $0.05 | $0.05 | — | 0% | -| qwen--qwen3-coder-flash | openrouter | — | $0.195 | $0.039 | $0.24375 | 80% | -| qwen--qwen3-coder-next | openrouter | — | $0.11 | $0.07 | — | 36% | -| qwen--qwen3-coder-plus | openrouter | — | $0.65 | $0.13 | $0.8125 | 80% | -| qwen--qwen3-max | openrouter | — | $0.78 | $0.156 | $0.975 | 80% | -| qwen--qwen3-vl-235b-a22b-instruct | openrouter | — | $0.2 | $0.11 | — | 45% | -| qwen--qwen3.5-397b-a17b | openrouter | — | $0.39 | $0.195 | — | 50% | -| qwen--qwen3.5-flash-02-23 | openrouter | — | $0.065 | — | $0.08125 | — | -| qwen--qwen3.5-plus-02-15 | openrouter | — | $0.26 | — | $0.325 | — | -| qwen--qwen3.6-35b-a3b | openrouter | — | $0.15 | $0.05 | — | 67% | -| qwen--qwen3.6-flash | openrouter | — | $0.1875 | — | $0.234375 | — | -| qwen--qwen3.6-max-preview | openrouter | — | $1.04 | — | $1.3 | — | -| qwen--qwen3.6-plus | openrouter | — | $0.325 | — | $0.40625 | — | -| tencent--hy3-preview | openrouter | — | $0.066 | $0.029 | — | 56% | -| thedrummer--cydonia-24b-v4.1 | openrouter | — | $0.3 | $0.15 | — | 50% | -| thedrummer--skyfall-36b-v2 | openrouter | — | $0.55 | $0.25 | — | 55% | -| upstage--solar-pro-3 | openrouter | — | $0.15 | $0.015 | — | 90% | -| x-ai--grok-4.20 | openrouter | — | $1.25 | $0.2 | — | 84% | -| x-ai--grok-4.20-multi-agent | openrouter | — | $2 | $0.2 | — | 90% | -| x-ai--grok-4.3 | openrouter | — | $1.25 | $0.2 | — | 84% | -| xiaomi--mimo-v2-flash | openrouter | — | $0.1 | $0.01 | — | 90% | -| xiaomi--mimo-v2-omni | openrouter | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2-pro | openrouter | — | $1 | $0.2 | — | 80% | -| xiaomi--mimo-v2.5 | openrouter | — | $0.4 | $0.08 | — | 80% | -| xiaomi--mimo-v2.5-pro | openrouter | — | $1 | $0.2 | — | 80% | -| z-ai--glm-4.5 | openrouter | — | $0.6 | $0.11 | — | 82% | -| z-ai--glm-4.5-air | openrouter | — | $0.13 | $0.025 | — | 81% | -| z-ai--glm-4.5v | openrouter | — | $0.6 | $0.11 | — | 82% | -| z-ai--glm-4.6 | openrouter | — | $0.43 | $0.08 | — | 81% | -| z-ai--glm-4.6v | openrouter | — | $0.3 | $0.05 | — | 83% | -| z-ai--glm-4.7 | openrouter | — | $0.4 | $0.08 | — | 80% | -| z-ai--glm-4.7-flash | openrouter | — | $0.06 | $0.01 | — | 83% | -| z-ai--glm-5 | openrouter | — | $0.6 | $0.12 | — | 80% | -| z-ai--glm-5-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | -| z-ai--glm-5v-turbo | openrouter | — | $1.2 | $0.24 | — | 80% | -| ~anthropic--claude-haiku-latest | openrouter | — | $1 | $0.1 | $1.25 | 90% | -| ~anthropic--claude-opus-latest | openrouter | — | $5 | $0.5 | $6.25 | 90% | -| ~anthropic--claude-sonnet-latest | openrouter | — | $3 | $0.3 | $3.75 | 90% | -| ~google--gemini-flash-latest | openrouter | — | $1.5 | $0.15 | $0.083333 | 90% | -| ~google--gemini-pro-latest | openrouter | — | $2 | $0.2 | $0.375 | 90% | -| ~moonshotai--kimi-latest | openrouter | — | $0.73 | $0.25 | — | 66% | -| ~openai--gpt-latest | openrouter | — | $5 | $0.5 | — | 90% | -| ~openai--gpt-mini-latest | openrouter | — | $0.75 | $0.075 | — | 90% | -| deepseek--deepseek-v3-0324 | ppio | — | $2 | $0.6 | — | 70% | -| deepseek--deepseek-v3.1 | ppio | — | $4 | $2 | — | 50% | -| deepseek--deepseek-v3.1-terminus | ppio | — | $4 | $2 | — | 50% | -| deepseek--deepseek-v3.2 | ppio | — | $2 | $0.2 | — | 90% | -| deepseek--deepseek-v4-flash | ppio | — | $1 | $0.2 | — | 80% | -| deepseek--deepseek-v4-pro | ppio | — | $12 | $1 | — | 92% | -| minimax--minimax-m2 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | -| minimax--minimax-m2.1 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | -| minimax--minimax-m2.5 | ppio | — | $2.1 | $0.21 | $2.625 | 90% | -| minimax--minimax-m2.5-highspeed | ppio | — | $4.2 | $0.21 | $2.625 | 95% | -| minimax--minimax-m2.7 | ppio | — | $2.1 | $0.42 | $2.625 | 80% | -| minimax--minimax-m2.7-highspeed | ppio | — | $4.2 | $0.42 | $2.625 | 90% | -| moonshotai--kimi-k2.5 | ppio | — | $4 | $0.7 | — | 82% | -| moonshotai--kimi-k2.6 | ppio | — | $6.5 | $1.1 | — | 83% | -| xiaomimimo--mimo-v2-flash | ppio | — | $0.7 | $0.07 | — | 90% | -| zai-org--glm-4.5 | ppio | — | $4 | $0.8 | — | 80% | -| zai-org--glm-4.5-air | ppio | — | $1.2 | $0.24 | — | 80% | -| zai-org--glm-4.5v | ppio | — | $4 | $0.8 | — | 80% | -| zai-org--glm-4.6 | ppio | — | $4 | $0.8 | $0.8 | 80% | -| zai-org--glm-4.7-flash | ppio | — | $0.5 | $0.1 | — | 80% | -| gemma-3-27b | privatemode | — | $0.77 | $0.08 | — | 90% | -| gemma-4-31b | privatemode | — | $0.77 | $0.08 | — | 90% | -| gpt-oss-120b | privatemode | — | $0.43 | $0.04 | — | 91% | -| kimi-k2-5 | privatemode | — | $1.55 | $0.15 | — | 90% | -| qwen3-coder-30b-a3b | privatemode | — | $0.43 | $0.04 | — | 91% | -| alibaba--qwen-max | requesty | — | $1.6 | $1.6 | $1.6 | 0% | -| alibaba--qwen-plus | requesty | — | $0.4 | $0.4 | $0.4 | 0% | -| alibaba--qwen-turbo | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| alibaba--qwen3-30b-a3b-instruct-2507 | requesty | — | $0.2 | $0.2 | $0.8 | 0% | -| alibaba--qwen3-coder-flash | requesty | — | $0.3 | $0.08 | $0.3 | 73% | -| alibaba--qwen3.5 | requesty | — | $0.6 | $0.6 | $3.6 | 0% | -| anthropic--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | -| anthropic--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| anthropic--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| anthropic--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| anthropic--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| azure--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| azure--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| azure--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| azure--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| azure--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | -| azure--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | -| azure--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| azure--gpt-5.2 | requesty | — | $1.75 | $0.175 | $14 | 90% | -| azure--openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| azure--openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| azure--openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| bedrock--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | -| bedrock--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | -| bedrock--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| bedrock--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| bedrock--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| bedrock--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| bedrock--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| bedrock--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| deepinfra--deepseek-ai--deepseek-r1 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | -| deepinfra--deepseek-ai--deepseek-r1-distill-llama-70b | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--deepseek-ai--deepseek-v3 | requesty | — | $0.85 | $0.85 | $0.85 | 0% | -| deepinfra--deepseek-ai--deepseek-v3.1 | requesty | — | $0.3 | $0.3 | $1 | 0% | -| deepinfra--kimi k2.5 | requesty | — | $0.45 | $0.07 | — | 84% | -| deepinfra--meta-llama--llama-3.2-90b-vision-instruct | requesty | — | $0.35 | $0.35 | $0.35 | 0% | -| deepinfra--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.12 | $0.12 | $0.12 | 0% | -| deepinfra--meta-llama--meta-llama-3.1-405b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | -| deepinfra--meta-llama--meta-llama-3.1-70b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.02 | $0.02 | $0.02 | 0% | -| deepinfra--microsoft--phi-4 | requesty | — | $0.07 | $0.07 | $0.07 | 0% | -| deepinfra--qwen--qwen2.5-72b-instruct | requesty | — | $0.23 | $0.23 | $0.23 | 0% | -| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | — | $0.07 | $0.07 | $0.07 | 0% | -| deepinfra--qwen--qwen3-235b-a22b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| deepinfra--qwen--qwen3-32b | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| deepinfra--qwen--qwen3-coder-480b-a35b-instruct | requesty | — | $0.4 | $0.4 | $1.6 | 0% | -| deepinfra--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | -| deepinfra--zai-org--glm-4.5-air | requesty | — | $0.2 | $0.2 | $1.1 | 0% | -| deepseek--deepseek-chat | requesty | — | $0.14 | $0.028 | $0.14 | 80% | -| deepseek--deepseek-reasoner | requesty | — | $0.14 | $0.028 | $0.14 | 80% | -| deepseek--deepseek-v4-flash | requesty | — | $0.14 | $0.028 | $0.14 | 80% | -| deepseek--deepseek-v4-pro | requesty | — | $1.74 | $0.145 | $1.74 | 92% | -| fireworks--deepseek-v3.2 | requesty | — | $0.56 | $0.28 | — | 50% | -| fireworks--deepseek-v4-pro | requesty | — | $1.74 | $0.15 | — | 91% | -| fireworks--glm-5 | requesty | — | $1 | $0.2 | — | 80% | -| fireworks--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | -| fireworks--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $3 | 83% | -| fireworks--kimi-k2.6 | requesty | — | $0.95 | $0.16 | — | 83% | -| fireworks--minimax-m2.5 | requesty | — | $0.3 | $0.03 | — | 90% | -| fireworks--minimax-m2.7 | requesty | — | $0.3 | $0.06 | — | 80% | -| fireworks--qwen3.6-plus | requesty | — | $0.5 | $0.1 | — | 80% | -| google--gemini-2.0-flash-001 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| google--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | -| google--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | -| google--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | -| google--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | -| google--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| google--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | -| google--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| groq--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.75 | 0% | -| groq--openai--gpt-oss-20b | requesty | — | $0.1 | $0.1 | $0.5 | 0% | -| inceptron--glm-5.1 | requesty | — | $1.4 | $0.26 | — | 81% | -| inceptron--kimi-k2.6 | requesty | — | $0.8 | $0.2 | — | 75% | -| inceptron--minimax-m2.5 | requesty | — | $0.28 | $0.03 | — | 89% | -| minimaxi--minimax-m2 | requesty | — | $0.3 | $0.3 | $1.2 | 0% | -| minimaxi--minimax-m2.5 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | -| minimaxi--minimax-m2.5-highspeed | requesty | — | $0.6 | $0.06 | $2.4 | 90% | -| minimaxi--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | -| minimaxi--minimax-m2.7-highspeed | requesty | — | $0.6 | $0.06 | $1.2 | 90% | -| mistral--codestral-latest | requesty | — | $0.3 | $0.3 | $0.9 | 0% | -| mistral--devstral-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | -| mistral--devstral-medium-2507 | requesty | — | $0.4 | $0.4 | $2 | 0% | -| mistral--devstral-small-2507 | requesty | — | $0.1 | $0.1 | $0.3 | 0% | -| mistral--devstral-small-latest | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| mistral--mistral-large-latest | requesty | — | $0.5 | $0.5 | $0.5 | 0% | -| mistral--mistral-medium-latest | requesty | — | $0.4 | $0.4 | $2 | 0% | -| mistral--mistral-small-2503 | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| mistral--mistral-small-2603 | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| mistral--mistral-small-latest | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| mistral--open-mistral-7b | requesty | — | $0.25 | $0.25 | $0.25 | 0% | -| mistral--pixtral-large-latest | requesty | — | $2 | $2 | $5 | 0% | -| moonshot--kimi-k2-0711-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-0905-preview | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-thinking | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-thinking-turbo | requesty | — | $0.6 | $0.15 | $0.6 | 75% | -| moonshot--kimi-k2-turbo-preview | requesty | — | $1.2 | $0.3 | $5 | 75% | -| moonshot--kimi-k2.5 | requesty | — | $0.6 | $0.1 | $0.6 | 83% | -| moonshot--kimi-k2.6 | requesty | — | $0.95 | $0.16 | $0.96 | 83% | -| nebius--deepseek-ai--deepseek-v3.2 | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| nebius--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.13 | $0.13 | $0.13 | 0% | -| nebius--moonshotai--kimi-k2.5 | requesty | — | $0.5 | $0.5 | $2.5 | 0% | -| nebius--openai--gpt-oss-120b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| novita--deepseek--deepseek-prover-v2-671b | requesty | — | $0.7 | $0.7 | $0.7 | 0% | -| novita--deepseek--deepseek-r1 | requesty | — | $4 | $4 | $4 | 0% | -| novita--deepseek--deepseek-r1-distill-llama-70b | requesty | — | $0.8 | $0.8 | $0.8 | 0% | -| novita--deepseek--deepseek-r1-distill-qwen-14b | requesty | — | $0.15 | $0.15 | $0.15 | 0% | -| novita--deepseek--deepseek-r1-distill-qwen-32b | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| novita--deepseek--deepseek-r1-turbo | requesty | — | $0.7 | $0.7 | $2.5 | 0% | -| novita--deepseek--deepseek-v3-0324 | requesty | — | $0.4 | $0.4 | $0.4 | 0% | -| novita--deepseek--deepseek-v3-turbo | requesty | — | $0.4 | $0.4 | $0.4 | 0% | -| novita--deepseek--deepseek-v3.2 | requesty | — | $0.269 | $0.1345 | $0.269 | 50% | -| novita--deepseek--deepseek_v3 | requesty | — | $0.89 | $0.89 | $0.89 | 0% | -| novita--glm-5 | requesty | — | $1 | $0.2 | — | 80% | -| novita--gryphe--mythomax-l2-13b | requesty | — | $0.09 | $0.09 | $0.09 | 0% | -| novita--meta-llama--llama-3-70b-instruct | requesty | — | $0.51 | $0.51 | $0.51 | 0% | -| novita--meta-llama--llama-3-8b-instruct | requesty | — | $0.04 | $0.04 | $0.04 | 0% | -| novita--meta-llama--llama-3.1-8b-instruct | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| novita--meta-llama--llama-3.2-1b-instruct | requesty | — | $0.02 | $0.02 | $0.02 | 0% | -| novita--meta-llama--llama-3.2-3b-instruct | requesty | — | $0.03 | $0.03 | $0.03 | 0% | -| novita--meta-llama--llama-3.3-70b-instruct | requesty | — | $0.39 | $0.39 | $0.39 | 0% | -| novita--meta-llama--llama-4-maverick-17b-128e-instruct-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| novita--microsoft--wizardlm-2-8x22b | requesty | — | $0.62 | $0.62 | $0.62 | 0% | -| novita--minimax--minimax-m2.7 | requesty | — | $0.3 | $0.06 | $1.2 | 80% | -| novita--mistralai--mistral-nemo | requesty | — | $0.17 | $0.17 | $0.17 | 0% | -| novita--moonshotai--kimi-k2-instruct | requesty | — | $0.57 | $0.57 | $0.57 | 0% | -| novita--nousresearch--hermes-2-pro-llama-3-8b | requesty | — | $0.14 | $0.14 | $0.14 | 0% | -| novita--qwen--qwen-2.5-72b-instruct | requesty | — | $0.38 | $0.38 | $0.38 | 0% | -| novita--qwen--qwen2.5-vl-72b-instruct | requesty | — | $0.8 | $0.8 | $0.8 | 0% | -| novita--qwen--qwen3-235b-a22b-fp8 | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| novita--qwen--qwen3.5-397b-a17b | requesty | — | $0.6 | $0.6 | $3.6 | 0% | -| novita--sao10k--l3-70b-euryale-v2.1 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | -| novita--sao10k--l3-8b-lunaris | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| novita--sao10k--l3-8b-stheno-v3.2 | requesty | — | $0.05 | $0.05 | $0.05 | 0% | -| novita--sao10k--l31-70b-euryale-v2.2 | requesty | — | $1.48 | $1.48 | $1.48 | 0% | -| novita--zai-org--glm-4.5 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | -| novita--zai-org--glm-4.6 | requesty | — | $0.6 | $0.6 | $2.2 | 0% | -| openai--chatgpt-4o | requesty | — | $5 | $5 | $5 | 0% | -| openai--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| openai--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| openai--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| openai--gpt-4o | requesty | — | $2.5 | $1.25 | $2.5 | 50% | -| openai--gpt-4o-2024-05-13 | requesty | — | $2.5 | $2.5 | $2.5 | 0% | -| openai--gpt-4o-2024-08-06 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | -| openai--gpt-4o-2024-11-20 | requesty | — | $2.5 | $1.25 | $2.5 | 50% | -| openai--gpt-4o-mini | requesty | — | $0.15 | $0.075 | $0.15 | 50% | -| openai--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | -| openai--gpt-5-mini:flex | requesty | — | $0.125 | $0.0125 | $0.125 | 90% | -| openai--gpt-5-mini:priority | requesty | — | $0.45 | $0.045 | $3.6 | 90% | -| openai--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | -| openai--gpt-5-nano:flex | requesty | — | $0.025 | $0.0025 | $0.025 | 90% | -| openai--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5.1-chat | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | -| openai--gpt-5.2-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai--gpt-5.4 | requesty | — | $2.5 | $0.25 | — | 90% | -| openai--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | -| openai--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | -| openai--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | -| openai--gpt-5:flex | requesty | — | $0.625 | $0.0625 | $0.625 | 90% | -| openai--gpt-5:priority | requesty | — | $2.5 | $0.25 | $20 | 90% | -| openai--o1 | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o1-pro | requesty | — | $150 | $150 | $150 | 0% | -| openai--o1:high | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o1:low | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o1:medium | requesty | — | $15 | $7.5 | $15 | 50% | -| openai--o3 | requesty | — | $2 | $0.5 | $2 | 75% | -| openai--o3-deep-research | requesty | — | $10 | $2.5 | $10 | 75% | -| openai--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3-mini:high | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3-mini:low | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3-mini:medium | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai--o3:flex | requesty | — | $1 | $0.25 | $1 | 75% | -| openai--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai--o4-mini-deep-research | requesty | — | $2 | $0.5 | $2 | 75% | -| openai--o4-mini:flex | requesty | — | $0.55 | $0.138 | $0.55 | 75% | -| openai--o4-mini:high | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai--o4-mini:low | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai--o4-mini:medium | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| openai-responses--gpt-4.1 | requesty | — | $2 | $0.5 | $2 | 75% | -| openai-responses--gpt-4.1-mini | requesty | — | $0.4 | $0.1 | $0.4 | 75% | -| openai-responses--gpt-4.1-nano | requesty | — | $0.1 | $0.025 | $0.1 | 75% | -| openai-responses--gpt-5 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai-responses--gpt-5-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | -| openai-responses--gpt-5-mini | requesty | — | $0.25 | $0.025 | $0.25 | 90% | -| openai-responses--gpt-5-nano | requesty | — | $0.05 | $0.005 | $0.05 | 90% | -| openai-responses--gpt-5-pro | requesty | — | $15 | $15 | $120 | 0% | -| openai-responses--gpt-5.1 | requesty | — | $1.25 | $0.125 | $10 | 90% | -| openai-responses--gpt-5.1-codex | requesty | — | $1.25 | $0.125 | $1.25 | 90% | -| openai-responses--gpt-5.2 | requesty | — | $1.75 | $0.175 | $0.175 | 90% | -| openai-responses--gpt-5.2-codex | requesty | — | $1.75 | $0.175 | $1.75 | 90% | -| openai-responses--gpt-5.3-chat | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai-responses--gpt-5.3-codex | requesty | — | $1.75 | $0.175 | $14 | 90% | -| openai-responses--gpt-5.4 | requesty | — | $2.5 | $0.25 | $2.5 | 90% | -| openai-responses--gpt-5.4-mini | requesty | — | $0.75 | $0.075 | $4.5 | 90% | -| openai-responses--gpt-5.4-nano | requesty | — | $0.2 | $0.02 | $1.25 | 90% | -| openai-responses--gpt-5.4-pro | requesty | — | $30 | $30 | $180 | 0% | -| openai-responses--gpt-5.5 | requesty | — | $5 | $0.5 | — | 90% | -| openai-responses--o3-mini | requesty | — | $1.1 | $0.55 | $1.1 | 50% | -| openai-responses--o3-pro | requesty | — | $20 | $20 | $20 | 0% | -| openai-responses--o4-mini | requesty | — | $1.1 | $0.275 | $1.1 | 75% | -| parasail--parasail-gemma3-27b-it | requesty | — | $0.3 | $0.3 | $0.5 | 0% | -| parasail--parasail-kimi-k2-instruct | requesty | — | $0.99 | $0.99 | $2.99 | 0% | -| parasail--parasail-qwen25-vl-72b-instruct | requesty | — | $0.7 | $0.7 | $0.7 | 0% | -| parasail--parasail-qwen3-235b-a22b-instruct-2507 | requesty | — | $0.15 | $0.15 | $0.85 | 0% | -| perplexity--sonar | requesty | — | $1 | $1 | $1 | 0% | -| perplexity--sonar-pro | requesty | — | $3 | $3 | $3 | 0% | -| perplexity--sonar-reasoning-pro | requesty | — | $2 | $2 | $2 | 0% | -| together--deepseek-ai--deepseek-r1 | requesty | — | $3 | $3 | $7 | 0% | -| together--deepseek-ai--deepseek-v3 | requesty | — | $1.25 | $1.25 | $1.25 | 0% | -| together--meta-llama--llama-3.2-3b-instruct-turbo | requesty | — | $0.06 | $0.06 | $0.06 | 0% | -| together--meta-llama--llama-3.3-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | -| together--meta-llama--llamaguard-2-8b | requesty | — | $0.2 | $0.2 | $0.2 | 0% | -| together--meta-llama--meta-llama-3-8b-instruct-lite | requesty | — | $0.1 | $0.1 | $0.1 | 0% | -| together--meta-llama--meta-llama-3.1-70b-instruct-turbo | requesty | — | $0.88 | $0.88 | $0.88 | 0% | -| together--meta-llama--meta-llama-3.1-8b-instruct-turbo | requesty | — | $0.18 | $0.18 | $0.18 | 0% | -| together--qwen--qwen2.5-72b-instruct-turbo | requesty | — | $1.2 | $1.2 | $1.2 | 0% | -| together--qwen--qwen2.5-7b-instruct-turbo | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| vertex--claude-3-7-sonnet | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--claude-haiku-4-5 | requesty | — | $1 | $0.1 | $1.25 | 90% | -| vertex--claude-opus-4 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| vertex--claude-opus-4-1 | requesty | — | $15 | $1.5 | $18.75 | 90% | -| vertex--claude-opus-4-5 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| vertex--claude-opus-4-6 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| vertex--claude-opus-4-7 | requesty | — | $5 | $0.5 | $6.25 | 90% | -| vertex--claude-sonnet-4 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--claude-sonnet-4-5 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--claude-sonnet-4-6 | requesty | — | $3 | $0.3 | $3.75 | 90% | -| vertex--deepseek-v3.2 | requesty | — | $0.56 | $0.056 | $1.68 | 90% | -| vertex--gemini-2.5-flash | requesty | — | $0.3 | $0.075 | $0.55 | 75% | -| vertex--gemini-2.5-flash-image | requesty | — | $0.3 | $0.3 | $2.5 | 0% | -| vertex--gemini-2.5-flash-lite | requesty | — | $0.1 | $0.01 | $0.18333 | 90% | -| vertex--gemini-2.5-pro | requesty | — | $1.25 | $0.31 | $2.375 | 75% | -| vertex--gemini-3-flash-preview | requesty | — | $0.5 | $0.05 | $1 | 90% | -| vertex--gemini-3-flash-preview:flex | requesty | — | $0.25 | $0.025 | $0.5 | 90% | -| vertex--gemini-3-pro-image-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| vertex--gemini-3-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| vertex--gemini-3.1-flash-lite | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | -| vertex--gemini-3.1-flash-lite-preview | requesty | — | $0.25 | $0.025 | $0.08333 | 90% | -| vertex--gemini-3.1-pro-preview | requesty | — | $2 | $0.2 | $4.5 | 90% | -| vertex--gemini-3.1-pro-preview:flex | requesty | — | $1 | $0.1 | $2.75 | 90% | -| vertex--gemini-3.5-flash | requesty | — | $1.5 | $0.15 | $1.583 | 90% | -| vertex--kimi-k2 | requesty | — | $0.6 | $0.06 | $2.5 | 90% | -| xai--grok-2-1212 | requesty | — | $2 | $2 | $2 | 0% | -| xai--grok-3 | requesty | — | $5 | $5 | $5 | 0% | -| xai--grok-3-mini | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| xai--grok-3-mini:high | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| xai--grok-3-mini:low | requesty | — | $0.3 | $0.3 | $0.3 | 0% | -| xai--grok-4 | requesty | — | $3 | $0.75 | $3 | 75% | -| xai--grok-4-1-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4-1-fast-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4-fast | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4-fast-non-reasoning | requesty | — | $0.2 | $0.05 | $0.2 | 75% | -| xai--grok-4.2-beta | requesty | — | $2 | $0.2 | $2 | 90% | -| xai--grok-4.3 | requesty | — | $1.25 | $0.2 | $1.25 | 84% | -| xai--grok-code-fast-1 | requesty | — | $0.2 | $0.02 | $0.2 | 90% | -| zai--glm-4.5 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | -| zai--glm-4.6 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | -| zai--glm-4.7 | requesty | — | $0.6 | $0.11 | $0.6 | 82% | -| zai--glm-5 | requesty | — | $1 | $0.2 | — | 80% | -| zai--glm-5.1 | requesty | — | $1.4 | $0.26 | $4.4 | 81% | -| deepseek-v3.2 | siliconflow | — | $0.27 | $0.135 | — | 50% | -| deepseek-v4-flash | siliconflow | — | $0.14 | $0.028 | — | 80% | -| deepseek-v4-pro | siliconflow | — | $1.74 | $0.145 | — | 92% | -| glm-4.7 | siliconflow | — | $0.42 | $0.11 | — | 74% | -| glm-5 | siliconflow | — | $0.95 | $0.2 | — | 79% | -| glm-5.1 | siliconflow | — | $1.4 | $0.26 | — | 81% | -| hy3-preview | siliconflow | — | $0.066 | $0.029 | — | 56% | -| kimi-k2.5 | siliconflow | — | $0.45 | $0.07 | — | 84% | -| kimi-k2.6 | siliconflow | — | $0.9 | $0.2 | — | 78% | -| minimax-m2.5 | siliconflow | — | $0.3 | $0.03 | — | 90% | -| step-1-32k | stepfun | — | $15 | $3 | — | 80% | -| step-1-8k | stepfun | — | $5 | $1 | — | 80% | -| step-1o-audio | stepfun | — | $25 | $5 | — | 80% | -| step-1o-turbo-vision | stepfun | — | $2.5 | $0.5 | — | 80% | -| step-1o-vision-32k | stepfun | — | $15 | $3 | — | 80% | -| step-1v-32k | stepfun | — | $15 | $3 | — | 80% | -| step-1v-8k | stepfun | — | $5 | $1 | — | 80% | -| step-2-16k | stepfun | — | $38 | $7.6 | — | 80% | -| step-2-16k-exp | stepfun | — | $38 | $7.6 | — | 80% | -| step-2-mini | stepfun | — | $1 | $0.2 | — | 80% | -| step-3 | stepfun | — | $1.5 | $0.3 | — | 80% | -| step-3.5-flash | stepfun | — | $0.7 | $0.14 | — | 80% | -| step-3.5-flash-2603 | stepfun | — | $0.7 | $0.14 | — | 80% | -| step-audio-2 | stepfun | — | $10 | $2 | — | 80% | -| step-r1-v-mini | stepfun | — | $2.5 | $0.5 | — | 80% | -| stepaudio-2.5-chat | stepfun | — | $10 | $2 | — | 80% | -| stepaudio-2.5-realtime | stepfun | — | $10 | $2 | — | 80% | -| deepseek-v4-flash | tencent-tokenhub | — | $1 | $0.2 | — | 80% | -| deepseek-v4-pro | tencent-tokenhub | — | $12 | $1 | — | 92% | -| glm-5 | tencent-tokenhub | — | $4 | $1 | — | 75% | -| glm-5-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | -| glm-5.1 | tencent-tokenhub | — | $6 | $1.3 | — | 78% | -| glm-5v-turbo | tencent-tokenhub | — | $5 | $1.2 | — | 76% | -| hy3-preview | tencent-tokenhub | — | $1.2 | $0.4 | — | 67% | -| kimi-k2.5 | tencent-tokenhub | — | $4 | $0.7 | — | 82% | -| kimi-k2.6 | tencent-tokenhub | — | $6.5 | $1.1 | — | 83% | -| minimax-m2.5 | tencent-tokenhub | — | $2.1 | $0.21 | — | 90% | -| minimax-m2.7 | tencent-tokenhub | — | $2.1 | $0.42 | — | 80% | -| MiniMaxAI--MiniMax-M2.5 | togetherai | — | $0.3 | $0.06 | — | 80% | -| MiniMaxAI--MiniMax-M2.7 | togetherai | — | $0.3 | $0.06 | — | 80% | -| deepseek-ai--DeepSeek-V4-Pro | togetherai | — | $2.1 | $0.2 | — | 90% | -| moonshotai--Kimi-K2.6 | togetherai | — | $1.2 | $0.2 | — | 83% | -| solar-pro2 | upstage | — | $0.15 | $0.015 | — | 90% | -| solar-pro3 | upstage | — | $0.15 | $0.015 | — | 90% | -| aion-labs-aion-2-0 | venice | — | $1 | $0.25 | — | 75% | -| arcee-trinity-large-thinking | venice | — | $0.3125 | $0.075 | — | 76% | -| claude-opus-4-5 | venice | — | $6 | $0.6 | — | 90% | -| claude-opus-4-6 | venice | — | $6 | $0.6 | — | 90% | -| claude-opus-4-6-fast | venice | — | $36 | $3.6 | — | 90% | -| claude-opus-4-7 | venice | — | $6 | $0.6 | — | 90% | -| claude-opus-4-7-fast | venice | — | $36 | $3.6 | — | 90% | -| claude-sonnet-4-5 | venice | — | $3.75 | $0.375 | — | 90% | -| claude-sonnet-4-6 | venice | — | $3.6 | $0.36 | — | 90% | -| deepseek-v3.2 | venice | — | $0.33 | $0.16 | — | 52% | -| deepseek-v4-flash | venice | — | $0.17 | $0.028 | — | 84% | -| deepseek-v4-pro | venice | — | $1.73 | $0.33 | — | 81% | -| gemini-3-1-pro-preview | venice | — | $2.5 | $0.5 | — | 80% | -| gemini-3-flash-preview | venice | — | $0.7 | $0.07 | — | 90% | -| grok-4-20 | venice | — | $1.42 | $0.23 | — | 84% | -| grok-4-20-multi-agent | venice | — | $1.42 | $0.23 | — | 84% | -| grok-4-3 | venice | — | $1.42 | $0.23 | — | 84% | -| kimi-k2-5 | venice | — | $0.56 | $0.22 | — | 61% | -| kimi-k2-6 | venice | — | $0.85 | $0.22 | — | 74% | -| mercury-2 | venice | — | $0.3125 | $0.03125 | — | 90% | -| minimax-m25 | venice | — | $0.34 | $0.04 | — | 88% | -| minimax-m27 | venice | — | $0.375 | $0.075 | — | 80% | -| openai-gpt-4o-mini-2024-07-18 | venice | — | $0.1875 | $0.09375 | — | 50% | -| openai-gpt-52 | venice | — | $2.19 | $0.219 | — | 90% | -| openai-gpt-52-codex | venice | — | $2.19 | $0.219 | — | 90% | -| openai-gpt-53-codex | venice | — | $2.19 | $0.219 | — | 90% | -| openai-gpt-54 | venice | — | $3.13 | $0.313 | — | 90% | -| openai-gpt-54-mini | venice | — | $0.9375 | $0.09375 | — | 90% | -| openai-gpt-55 | venice | — | $6.25 | $0.625 | — | 90% | -| qwen-3-6-plus | venice | — | $0.625 | $0.0625 | — | 90% | -| qwen3-5-35b-a3b | venice | — | $0.3125 | $0.15625 | — | 50% | -| qwen3-coder-480b-a35b-instruct-turbo | venice | — | $0.35 | $0.04 | — | 89% | -| z-ai-glm-5-turbo | venice | — | $1.2 | $0.24 | — | 80% | -| z-ai-glm-5v-turbo | venice | — | $1.5 | $0.3 | — | 80% | -| zai-org-glm-4.6 | venice | — | $0.85 | $0.3 | — | 65% | -| zai-org-glm-4.7 | venice | — | $0.55 | $0.11 | — | 80% | -| zai-org-glm-5 | venice | — | $1 | $0.2 | — | 80% | -| zai-org-glm-5-1 | venice | — | $1.75 | $0.325 | — | 81% | -| GLM-5.1 | wafer | — | $1.5 | $0.15 | — | 90% | -| Qwen3.5-397B-A17B | wafer | — | $0.6 | $0.06 | — | 90% | +| 模型 | 提供商 | 上下文 | 输入 $/M | 缓存读取 $/M | 缓存写入 $/M | 节省 | +| --------------------------------------------- | -------------- | ------ | ------------------- | -------------------- | ------------ | ----- | +| aistudio_gemini-2.0-flash | aihubmix | — | $0.05 | $0.125 | — | -150% | +| aistudio_gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| anthropic-opus-4-6 | aihubmix | — | $2.5 | $0.25 | $3.125 | 90% | +| claude-haiku-4-5 | aihubmix | — | $0.55 | $0.055 | $0.6875 | 90% | +| claude-sonnet-4-0 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5 | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| claude-sonnet-4-5-think | aihubmix | — | $1.65 | $0.165 | $2.0625 | 90% | +| codex-mini-latest | aihubmix | — | $0.75 | $0.1875 | — | 75% | +| deepseek-v3.2 | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| deepseek-v3.2-exp | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-exp-think | aihubmix | — | $0.137 | $0.0137 | — | 90% | +| deepseek-v3.2-think | aihubmix | — | $0.151 | $0.0151 | — | 90% | +| doubao-1.5-lite-32k | aihubmix | — | $0.025 | $0.005 | — | 80% | +| doubao-1.5-pro-32k | aihubmix | — | $0.067 | $0.0134 | — | 80% | +| doubao-lite-32k | aihubmix | — | $0.03 | $0.006 | — | 80% | +| doubao-pro-32k | aihubmix | — | $0.07 | $0.014 | — | 80% | +| doubao-seed-1-6 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-flash | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-flash-250615 | aihubmix | — | $0.022 | $0.0044 | — | 80% | +| doubao-seed-1-6-lite | aihubmix | — | $0.041 | $0.0082 | — | 80% | +| doubao-seed-1-6-thinking | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-thinking-250615 | aihubmix | — | $0.09 | $0.018 | — | 80% | +| doubao-seed-1-6-vision-250815 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| doubao-seed-1-8 | aihubmix | — | $0.054795 | $0.010959 | — | 80% | +| gemini-2.0-flash | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.0-flash-001 | aihubmix | — | $0.05 | $0.125 | — | -150% | +| gemini-2.0-flash-search | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gemini-2.5-flash | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-lite | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | $0.005 | — | 90% | +| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-flash-search | aihubmix | — | $0.15 | $0.015 | — | 90% | +| gemini-2.5-pro | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gemini-2.5-pro-search | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| glm-4.5-airx | aihubmix | — | $0.55 | $0.11 | — | 80% | +| glm-4.5-x | aihubmix | — | $1.1 | $0.22 | — | 80% | +| glm-4.6 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| glm-4.6v | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| glm-4.7 | aihubmix | — | $0.136987 | $0.027397 | — | 80% | +| gpt-4.1 | aihubmix | — | $1 | $0.25 | — | 75% | +| gpt-4.1-mini | aihubmix | — | $0.2 | $0.05 | — | 75% | +| gpt-4.1-nano | aihubmix | — | $0.05 | $0.0125 | — | 75% | +| gpt-4o | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-08-06-global | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-2024-11-20 | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-4o-mini | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-2024-07-18 | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-global | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-mini-search-preview | aihubmix | — | $0.075 | $0.0375 | — | 50% | +| gpt-4o-search-preview | aihubmix | — | $1.25 | $0.625 | — | 50% | +| gpt-5 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5-nano | aihubmix | — | $0.025 | $0.0025 | — | 90% | +| gpt-5.1 | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-chat-latest | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-max | aihubmix | — | $0.625 | $0.0625 | — | 90% | +| gpt-5.1-codex-mini | aihubmix | — | $0.125 | $0.0125 | — | 90% | +| gpt-5.2 | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-chat-latest | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-codex | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-high | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-low | aihubmix | — | $0.875 | $0.0875 | — | 90% | +| gpt-5.2-pro | aihubmix | — | $10.5 | $1.05 | — | 90% | +| grok-4 | aihubmix | — | $1.65 | $0.4125 | — | 75% | +| grok-4-1-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-1-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-non-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4-fast-reasoning | aihubmix | — | $0.1 | $0.025 | — | 75% | +| grok-4.20-beta-0309-non-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-beta-0309-reasoning | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-4.20-multi-agent-beta-0309 | aihubmix | — | $1 | $0.1 | — | 90% | +| grok-code-fast-1 | aihubmix | — | $0.1 | $0.025 | — | 75% | +| kimi-k2-thinking | aihubmix | — | $0.274 | $0.0685 | — | 75% | +| kimi-k2-turbo-preview | aihubmix | — | $0.6 | $0.15 | — | 75% | +| kimi-k2.5 | aihubmix | — | $0.3 | $0.0525 | — | 82% | +| mimo-v2-flash | aihubmix | — | $0.0959 | $0.01918 | — | 80% | +| mimo-v2-omni | aihubmix | — | $0.22 | $0.044 | — | 80% | +| mimo-v2-pro | aihubmix | — | $0.55 | $0.11 | — | 80% | +| nvidia-nemotron-3-super-120b-a12b | aihubmix | — | $0.055 | $0.01375 | — | 75% | +| o1 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-2024-12-17 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-global | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-mini | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-mini-2024-09-12 | aihubmix | — | $1.5 | $0.75 | — | 50% | +| o1-preview | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o1-preview-2024-09-12 | aihubmix | — | $7.5 | $3.75 | — | 50% | +| o3 | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-deep-research | aihubmix | — | $5 | $1.25 | — | 75% | +| o3-global | aihubmix | — | $1 | $0.25 | — | 75% | +| o3-mini | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o3-mini-global | aihubmix | — | $0.55 | $0.275 | — | 50% | +| o4-mini | aihubmix | — | $0.55 | $0.1375 | — | 75% | +| qwen-plus | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-04-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-2025-07-28 | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-plus-latest | aihubmix | — | $0.0563 | $0.01126 | $0.070375 | 80% | +| qwen-turbo | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen-turbo-latest | aihubmix | — | $0.023 | $0.0046 | — | 80% | +| qwen3-coder-plus | aihubmix | — | $0.27 | $0.054 | — | 80% | +| qwen3-max | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-2026-01-23 | aihubmix | — | $0.2254 | $0.04508 | $0.28175 | 80% | +| qwen3-max-preview | aihubmix | — | $0.423 | $0.0846 | — | 80% | +| qwen3-vl-flash | aihubmix | — | $0.0103 | $0.00206 | — | 80% | +| qwen3-vl-plus | aihubmix | — | $0.0685 | $0.0137 | — | 80% | +| zai-glm-5-turbo | aihubmix | — | $0.6 | $0.12 | — | 80% | +| aion-2.0 | aion | — | $0.7999999999999999 | $0.19999999999999998 | — | 75% | +| aion-2.5 | aion | — | $1 | $0.35 | — | 65% | +| amazon-nova-2-lite | amazon-bedrock | — | $0.33 | $0.0825 | — | 75% | +| amazon-nova-lite | amazon-bedrock | — | $0.06 | $0.015 | — | 75% | +| amazon-nova-micro | amazon-bedrock | — | $0.035 | $0.00875 | — | 75% | +| amazon-nova-premier | amazon-bedrock | — | $2.5 | $0.625 | — | 75% | +| amazon-nova-pro | amazon-bedrock | — | $0.8 | $0.2 | — | 75% | +| claude-haiku-4-5-20251001 | auriko | — | $1 | $0.1 | $1.25 | 90% | +| claude-opus-4-1-20250805 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-20250514 | auriko | — | $15 | $1.5 | $18.75 | 90% | +| claude-opus-4-5-20251101 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-6 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-opus-4-7 | auriko | — | $5 | $0.5 | $6.25 | 90% | +| claude-sonnet-4-20250514 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-5-20250929 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| claude-sonnet-4-6 | auriko | — | $3 | $0.3 | $3.75 | 90% | +| deepseek-r1-0528 | auriko | — | $0.5 | $0.35 | — | 30% | +| deepseek-v3-0324 | auriko | — | $0.2 | $0.135 | — | 32% | +| deepseek-v3.1 | auriko | — | $0.21 | $0.13 | — | 38% | +| deepseek-v3.1-terminus | auriko | — | $0.27 | $0.13 | — | 52% | +| deepseek-v3.2 | auriko | — | $0.26 | $0.13 | — | 50% | +| deepseek-v4-flash | auriko | — | $0.14 | $0.0028 | — | 98% | +| deepseek-v4-pro | auriko | — | $0.435 | $0.003625 | — | 99% | +| gemini-2.5-flash | auriko | — | $0.3 | $0.03 | — | 90% | +| gemini-2.5-flash-lite | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-2.5-pro | auriko | — | $1.25 | $0.125 | — | 90% | +| gemini-3-flash-preview | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-3.1-flash-lite | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-flash-lite-preview | auriko | — | $0.25 | $0.025 | — | 90% | +| gemini-3.1-pro-preview | auriko | — | $2 | $0.2 | — | 90% | +| gemini-3.1-pro-preview-customtools | auriko | — | $2 | $0.2 | — | 90% | +| gemini-flash-latest | auriko | — | $0.5 | $0.05 | — | 90% | +| gemini-flash-lite-latest | auriko | — | $0.1 | $0.01 | — | 90% | +| gemini-pro-latest | auriko | — | $2 | $0.2 | — | 90% | +| glm-4.5 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.5-air | auriko | — | $0.2 | $0.03 | — | 85% | +| glm-4.5-airx | auriko | — | $1.1 | $0.22 | — | 80% | +| glm-4.5-x | auriko | — | $2.2 | $0.45 | — | 80% | +| glm-4.5v | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.6v | auriko | — | $0.3 | $0.05 | — | 83% | +| glm-4.6v-flashx | auriko | — | $0.04 | $0.004 | — | 90% | +| glm-4.7 | auriko | — | $0.6 | $0.11 | — | 82% | +| glm-4.7-flashx | auriko | — | $0.07 | $0.01 | — | 86% | +| glm-5 | auriko | — | $1 | $0.2 | — | 80% | +| glm-5-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| glm-5.1 | auriko | — | $1.4 | $0.26 | — | 81% | +| glm-5v-turbo | auriko | — | $1.2 | $0.24 | — | 80% | +| gpt-4.1-2025-04-14 | auriko | — | $2 | $0.5 | — | 75% | +| gpt-4.1-mini-2025-04-14 | auriko | — | $0.4 | $0.1 | — | 75% | +| gpt-4.1-nano-2025-04-14 | auriko | — | $0.1 | $0.025 | — | 75% | +| gpt-4o-2024-08-06 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-2024-11-20 | auriko | — | $2.5 | $1.25 | — | 50% | +| gpt-4o-mini-2024-07-18 | auriko | — | $0.15 | $0.075 | — | 50% | +| gpt-5-2025-08-07 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5-mini-2025-08-07 | auriko | — | $0.25 | $0.025 | — | 90% | +| gpt-5-nano-2025-08-07 | auriko | — | $0.05 | $0.005 | — | 90% | +| gpt-5.1-2025-11-13 | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.1-chat-latest | auriko | — | $1.25 | $0.125 | — | 90% | +| gpt-5.2-2025-12-11 | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.2-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.3-chat-latest | auriko | — | $1.75 | $0.175 | — | 90% | +| gpt-5.4-2026-03-05 | auriko | — | $2.5 | $0.25 | — | 90% | +| gpt-5.4-mini-2026-03-17 | auriko | — | $0.75 | $0.075 | — | 90% | +| gpt-5.4-nano-2026-03-17 | auriko | — | $0.2 | $0.02 | — | 90% | +| gpt-5.5-2026-04-23 | auriko | — | $5 | $0.5 | — | 90% | +| gpt-oss-120b | auriko | — | $0.15 | $0.01 | — | 93% | +| gpt-oss-20b | auriko | — | $0.07 | $0.04 | — | 43% | +| grok-4.20-0309-non-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.20-0309-reasoning | auriko | — | $1.25 | $0.2 | — | 84% | +| grok-4.3 | auriko | — | $1.25 | $0.2 | — | 84% | +| hy3-preview | auriko | — | $0.066 | $0.029 | — | 56% | +| kimi-k2-0711-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-0905-preview | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking | auriko | — | $0.6 | $0.15 | — | 75% | +| kimi-k2-thinking-turbo | auriko | — | $1.15 | $0.15 | — | 87% | -> 💡 使用[交互式目录](https://i-need-token.github.io/ai-models/)搜索和筛选更多条件的模型。 +> 📄 显示前 200 个，共 1374 个模型。使用[交互式目录](https://i-need-token.github.io/ai-models/)浏览全部。 --- diff --git a/llms-full.txt b/llms-full.txt index 70bf4d12..e240f31b 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -320,7 +320,7 @@ by_context = sorted( - [Glossary](docs/glossary.md) — key terms and definitions for AI model terminology - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers -- [Cached Pricing](docs/cached-pricing.md) — 75 models with prompt caching, 50-90% input cost savings +- [Cached Pricing](docs/cached-pricing.md) — 1,374 models with prompt caching, 50-90% input cost savings - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models diff --git a/llms.txt b/llms.txt index 75a3cc2c..0e5f3b4c 100644 --- a/llms.txt +++ b/llms.txt @@ -75,7 +75,7 @@ modalities: - [Glossary](docs/glossary.md) — key terms and definitions for AI model terminology - [Model Comparison](docs/model-comparison.md) — flagship, cost-effective, free, open-weight - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers -- [Cached Pricing](docs/cached-pricing.md) — 75 models with prompt caching, 50-90% input cost savings +- [Cached Pricing](docs/cached-pricing.md) — 1,374 models with prompt caching, 50-90% input cost savings - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models From 77f213d13ee9b838f466f0c697b9b8e556e9dff1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:01:36 +0800 Subject: [PATCH 142/311] Add video-models.md link to llms.txt and llms-full.txt --- llms-full.txt | 1 + llms.txt | 1 + 2 files changed, 2 insertions(+) diff --git a/llms-full.txt b/llms-full.txt index e240f31b..c61d86b2 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -329,6 +329,7 @@ by_context = sorted( - [Vision Models](docs/vision-models.md) — 1,487 vision models - [Image Generation](docs/image-generation.md) — 28 image generation models - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models +- [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video - [Provider Overview](docs/providers.md) — all 95 providers by type and market diff --git a/llms.txt b/llms.txt index 0e5f3b4c..5acee9b5 100644 --- a/llms.txt +++ b/llms.txt @@ -84,6 +84,7 @@ modalities: - [Vision Models](docs/vision-models.md) — 1,487 vision models - [Image Generation](docs/image-generation.md) — 28 image generation models - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models +- [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video - [Provider Overview](docs/providers.md) — all 95 providers by type and market From 08329ba0029ece5777766cd53ea306fc30ac3ef6 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:02:51 +0800 Subject: [PATCH 143/311] Add 'Try it now' one-liner to README hero section --- README.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/README.md b/README.md index 08503a5f..ad652d1d 100644 --- a/README.md +++ b/README.md @@ -21,6 +21,12 @@ Machine-readable YAML catalog of every major AI model provider and their models **[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[🔍 Search models →](https://i-need-token.github.io/ai-models/)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN access →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** +> 💡 **Try it now** — fetch model data in one command: +> +> ```bash +> curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d)} models across {len(set(m["provider"] for m in d))} providers')" +> ``` + ## Why This Catalog? | | | From 4d578da8f52aa6c80c43b0cc3eaac0a7fd7f8a1a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:08:39 +0800 Subject: [PATCH 144/311] =?UTF-8?q?Add=20model-selection.md=20guide=20(EN?= =?UTF-8?q?=20+=20ZH)=20=E2=80=94=20decision=20framework=20for=20choosing?= =?UTF-8?q?=20models?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Free, best value, large context model recommendations - Cross-linked to all related docs - Added to README, AGENTS.md, llms.txt, llms-full.txt --- README.md | 1 + docs/model-selection.md | 127 ++++++++++++++++ docs/zh/model-selection.md | 127 ++++++++++++++++ llms-full.txt | 1 + llms.txt | 290 ++++++++++++++++++++++++++++++++++++- 5 files changed, 541 insertions(+), 5 deletions(-) create mode 100644 docs/model-selection.md create mode 100644 docs/zh/model-selection.md diff --git a/README.md b/README.md index ad652d1d..06a8eeb0 100644 --- a/README.md +++ b/README.md @@ -421,6 +421,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | +| [Model Selection Guide](docs/model-selection.md) | Decision framework: free, best value, large context models | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | [Code Examples](docs/code-examples.md) | Practical examples in TypeScript, Python, Go, Rust, jq | | [FAQ](docs/faq.md) | Common questions about the catalog, data, and contributing | diff --git a/docs/model-selection.md b/docs/model-selection.md new file mode 100644 index 00000000..dc3c1720 --- /dev/null +++ b/docs/model-selection.md @@ -0,0 +1,127 @@ +# Model Selection Guide + +[中文](zh/model-selection.md) + +How to choose the right AI model for your use case — practical recommendations based on cost, capabilities, and context windows. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Decision Framework + +``` +What do you need? +├── Cheapest possible → Free models (81 available) +│ ├── With tool calling → See "Free + Tool Calling" below +│ ├── With reasoning → See "Free + Reasoning" below +│ └── Best overall free → See "Best Free Models" below +├── Best value (cheap + capable) → See "Best Value Models" below +├── Largest context → See "Large Context Models" below +├── Specific capability +│ ├── Tool calling → [Tool Calling Models](tool-calling.md) +│ ├── Reasoning → [Reasoning Models](reasoning-models.md) +│ ├── Vision → [Vision Models](vision-models.md) +│ ├── Structured output → [Structured Output](structured-output.md) +│ └── Prompt caching → [Cached Pricing](cached-pricing.md) +└── Full comparison → [Model Comparison](model-comparison.md) · [Pricing Comparison](pricing-comparison.md) +``` + +## Best Free Models + +Models with $0 input and $0 output pricing — perfect for prototyping and development. + +| Model | Provider | Context | Capabilities | +| ----- | -------- | ------- | ------------ | + +> See [Free AI Models](free-models.md) for the complete list of 81 free models. + +## Free + Tool Calling + +Free models that support function/tool calling — ideal for building agents at zero cost. + +| Model | Provider | Context | Capabilities | +| ----- | -------- | ------- | ------------ | + +## Free + Reasoning + +Free models with chain-of-thought reasoning — complex problem solving at zero cost. + +| Model | Provider | Context | Capabilities | +| ----- | -------- | ------- | ------------ | + +## Best Value Models + +Cheapest models with tool calling — best bang for the buck for production agents. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| --------------------------- | ------------ | ------- | --------- | ---------- | ------------ | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🔧 👁 📋 | +| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔧 🔓 | +| inclusionai--ling-2.6-flash | openrouter | 262K | $0.01 | $0.03 | 🔧 📋 | +| ling-2.6-flash | inclusionai | 262K | $0.01 | $0.03 | 🔧 | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 🔧 👁 | + +Cheapest models with vision: + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| -------------------- | --------- | ------- | --------- | ---------- | ------------ | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🔧 👁 📋 | +| deepseek-ocr | aihubmix | 0 | $0.01 | $0.01 | 👁 | +| gemini-2.0-flash-exp | aihubmix | 0 | $0.01 | $0.04 | 👁 | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 🔧 👁 | +| qwen3.5-0.8b | deepinfra | 262K | $0.01 | $0.05 | 🧠 👁 | + +Cheapest models with reasoning: + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| --------------------- | --------- | ------- | --------- | ---------- | ------------ | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 🔧 👁 | +| qwen3.5-0.8b | deepinfra | 262K | $0.01 | $0.05 | 🧠 👁 | +| gemma-2-2b-it | cortecs | 0 | $0.018 | $0.054 | 🧠 | +| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | 🧠 🔧 | +| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | 🧠 🔧 👁 | + +## Large Context Models + +Models with the largest context windows — for long documents, multi-turn conversations, and codebases. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ---------------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 🔧 👁 📋 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | 🔧 👁 | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 🔧 👁 📋 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | 🔧 👁 | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | 🔧 👁 | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | 🔧 👁 | + +> See [Context Window Comparison](context-windows.md) for the full analysis. + +## Cost Optimization Tips + +1. **Use free models for development** — prototype with free models, switch to paid for production +2. **Enable prompt caching** — [1,374 models](cached-pricing.md) support caching with 50-90% input cost savings +3. **Choose the smallest capable model** — e.g., GPT-4.1 Mini instead of GPT-4.1 for simple tasks +4. **Use open-weight models** — [527 models](open-weights.md) can run on your own infrastructure +5. **Compare across providers** — the same model is often cheaper through alternative providers (e.g., Groq, Together AI, DeepInfra) +6. **Batch requests** — some providers offer 50% discount for batch API calls +7. **Monitor usage** — track input/output token ratios to optimize model selection + +## Related Documentation + +- [Model Comparison](model-comparison.md) — flagship, cost-effective, free, and open-weight models +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Cached Pricing](cached-pricing.md) — models with prompt caching support +- [Free AI Models](free-models.md) — 81 free models by capability +- [Open-Weight Models](open-weights.md) — 527 models you can run yourself +- [Context Window Comparison](context-windows.md) — largest context windows +- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Quick Start Guide](quick-start.md) — get started in 30 seconds + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/model-selection.md b/docs/zh/model-selection.md new file mode 100644 index 00000000..f1386fe5 --- /dev/null +++ b/docs/zh/model-selection.md @@ -0,0 +1,127 @@ +# 模型选择指南 + +[English](../model-selection.md) + +如何根据使用场景选择合适的 AI 模型 — 基于成本、能力和上下文窗口的实用建议。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 决策框架 + +``` +你需要什么？ +├── 尽可能便宜 → 免费模型（81 个可用） +│ ├── 带工具调用 → 见下方"免费 + 工具调用" +│ ├── 带推理 → 见下方"免费 + 推理" +│ └── 最佳免费模型 → 见下方"最佳免费模型" +├── 最佳性价比（便宜 + 能力强）→ 见下方"最佳性价比模型" +├── 最大上下文 → 见下方"大上下文模型" +├── 特定能力 +│ ├── 工具调用 → [工具调用模型](tool-calling.md) +│ ├── 推理 → [推理模型](reasoning-models.md) +│ ├── 视觉 → [视觉模型](vision-models.md) +│ ├── 结构化输出 → [结构化输出](structured-output.md) +│ └── 提示缓存 → [缓存定价](cached-pricing.md) +└── 完整对比 → [模型对比](model-comparison.md) · [定价对比](pricing-comparison.md) +``` + +## 最佳免费模型 + +输入和输出定价均为 $0 的模型 — 非常适合原型开发和开发阶段。 + +| 模型 | 提供商 | 上下文 | 能力 | +| ---- | ------ | ------ | ---- | + +> 查看[免费 AI 模型](free-models.md)获取 81 个免费模型的完整列表。 + +## 免费 + 工具调用 + +支持函数/工具调用的免费模型 — 零成本构建 Agent 的理想选择。 + +| 模型 | 提供商 | 上下文 | 能力 | +| ---- | ------ | ------ | ---- | + +## 免费 + 推理 + +具有链式思维推理的免费模型 — 零成本解决复杂问题。 + +| 模型 | 提供商 | 上下文 | 能力 | +| ---- | ------ | ------ | ---- | + +## 最佳性价比模型 + +最便宜的工具调用模型 — 生产 Agent 的最佳性价比。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| --------------------------- | ------------ | ------ | -------- | -------- | -------- | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🔧 👁 📋 | +| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔧 🔓 | +| inclusionai--ling-2.6-flash | openrouter | 262K | $0.01 | $0.03 | 🔧 📋 | +| ling-2.6-flash | inclusionai | 262K | $0.01 | $0.03 | 🔧 | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 🔧 👁 | + +最便宜的视觉模型： + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| -------------------- | --------- | ------ | -------- | -------- | -------- | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🔧 👁 📋 | +| deepseek-ocr | aihubmix | 0 | $0.01 | $0.01 | 👁 | +| gemini-2.0-flash-exp | aihubmix | 0 | $0.01 | $0.04 | 👁 | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 🔧 👁 | +| qwen3.5-0.8b | deepinfra | 262K | $0.01 | $0.05 | 🧠 👁 | + +最便宜的推理模型： + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| --------------------- | --------- | ------ | -------- | -------- | -------- | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 🔧 👁 | +| qwen3.5-0.8b | deepinfra | 262K | $0.01 | $0.05 | 🧠 👁 | +| gemma-2-2b-it | cortecs | 0 | $0.018 | $0.054 | 🧠 | +| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | 🧠 🔧 | +| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | 🧠 🔧 👁 | + +## 大上下文模型 + +上下文窗口最大的模型 — 适用于长文档、多轮对话和代码库。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---------------------------- | ---------- | ------ | -------- | -------- | -------- | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 🔧 👁 📋 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | 🔧 👁 | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 🔧 👁 📋 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | 🔧 👁 | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | 🔧 👁 | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 👁 | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | 🔧 👁 | + +> 查看[上下文窗口对比](context-windows.md)获取完整分析。 + +## 成本优化技巧 + +1. **开发时使用免费模型** — 用免费模型原型开发，生产环境切换到付费模型 +2. **启用提示缓存** — [1,374 个模型](cached-pricing.md)支持缓存，输入成本节省 50-90% +3. **选择满足需求的最小模型** — 例如简单任务用 GPT-4.1 Mini 而非 GPT-4.1 +4. **使用开源权重模型** — [527 个模型](open-weights.md)可在自己的基础设施上运行 +5. **跨提供商对比** — 同一模型通过替代提供商（如 Groq、Together AI、DeepInfra）通常更便宜 +6. **批量请求** — 部分提供商对批量 API 调用提供 50% 折扣 +7. **监控用量** — 跟踪输入/输出 token 比率以优化模型选择 + +## 相关文档 + +- [模型对比](model-comparison.md) — 旗舰、高性价比、免费和开源模型 +- [定价对比](pricing-comparison.md) — 各提供商定价并排对比 +- [缓存定价](cached-pricing.md) — 支持提示缓存的模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 +- [上下文窗口对比](context-windows.md) — 最大上下文窗口 +- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [视觉模型](vision-models.md) — 1,487 个支持图像理解的模型 +- [快速入门](quick-start.md) — 30 秒上手 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/llms-full.txt b/llms-full.txt index c61d86b2..9d02af58 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -314,6 +314,7 @@ by_context = sorted( ## Documentation - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds +- [Model Selection Guide](docs/model-selection.md) — decision framework: free, best value, large context models - [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq - [FAQ](docs/faq.md) — frequently asked questions about the catalog, data, and contributing diff --git a/llms.txt b/llms.txt index 5acee9b5..9d02af58 100644 --- a/llms.txt +++ b/llms.txt @@ -1,8 +1,6 @@ # AI Models Catalog -> Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data only. - -See [llms-full.txt](./llms-full.txt) for the complete data schema, pricing types, provider list, and usage examples. +> Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data only. npm package available. ## What is this? @@ -19,6 +17,10 @@ A machine-readable YAML catalog of every major AI model provider and their model - 527 open-weight models - 81 free models - 1,487 vision models +- 829 structured output models +- 28 image generation models +- 118 audio input models +- 167 video input models ## Install @@ -64,11 +66,255 @@ limit: modalities: input: [text, image] output: [text] +release_date: "2026-05-18" +last_updated: "2026-05-18" +``` + +## Pricing Types + +| Type | When | Example | +| -------------- | ------------------------- | ------------------------------- | +| `TokenPricing` | Per-million-token pricing | `input: 2.5, output: 10` | +| `VideoPricing` | Per-second pricing | `unit: per_second, price: 0.03` | +| `UnitPricing` | Per-image or per-request | `unit: per_image, price: 0.04` | +| `FreePricing` | No cost | `unit: free` | + +## Covered Providers + +OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, Amazon Bedrock, Azure OpenAI, Google Vertex AI, OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, SiliconFlow, Novita AI, SambaNova, Cloudflare Workers AI, Chutes, Kluster AI, NanoGPT, and 75+ more. + +Full list: 01.AI, 302.AI, AI21 Labs, AIHubMix, AI/ML API, Aion Labs, Alibaba Cloud, Amazon Bedrock, Amazon Nova, Anthropic, Arcee AI, Auriko, Azure OpenAI, Baichuan AI, Baidu, Baseten, Berget, ByteDance, Cerebras, Chutes, Clarifai, CloudFerro Sherlock, Cloudflare Workers AI, Cohere, Cortecs, DInference, Databricks, DeepInfra, DeepSeek, DigitalOcean, evroc, FastRouter, Fireworks AI, FriendliAI, GMI Cloud, Google, Google Vertex AI, Groq, HPC-AI Cloud, Hyperbolic, IBM Granite, iFlytek SparkDesk, Inception Labs, InclusionAI, Inference.net, Kluster AI, LLM Gateway, Martian, MegaNova, Meta Llama, Microsoft Phi, MiniMax, Mistral AI, Mixlayer, MoArk AI, Moonshot AI, Morph, NanoGPT, Nebius, NeuralWatt, Nous Research, Novita AI, NVIDIA, OpenAI, OpenRouter, OrcaRouter, OVHcloud, PPIO, Perplexity, Privatemode AI, Qiniu AI, Regolo, Reka AI, Requesty, SambaNova, Sarvam AI, Scaleway, SiliconFlow, SiliconFlow CN, StepFun, SubModel, Tencent Cloud TokenHub, Tencent Hunyuan, TextSynth, Together AI, Upstage, Venice AI, Voyage AI, Vultr, Wafer, Writer, xAI Grok, Xiaomi, Zhipu AI, 接口 AI + +## Data Schema + +### Model Schema (Required Fields) + +| Field | Type | Description | Example | +| -------------- | ------- | ---------------------------------------- | ------------------------------------------ | +| `id` | string | Stable model ID (no date suffix) | `gpt-4o`, `claude-sonnet-4-5` | +| `name` | string | Display name | `GPT-4o`, `Claude Sonnet 4.5` | +| `family` | string | Model family (broad lineage) | `gpt-4o`, `claude-sonnet` | +| `pricing` | Pricing | Model pricing (see below) | — | +| `modalities` | object | Input/output modalities | `{ input: [text, image], output: [text] }` | +| `last_updated` | string | Last data update (YYYY-MM-DD or YYYY-MM) | `2024-08-06` | + +### Model Schema (Optional Fields) + +| Field | Type | Default | Description | Example | +| ------------------- | ------- | ------- | -------------------------------- | ------------------------------------ | +| `reasoning` | boolean | `false` | Supports reasoning/thinking mode | `true` | +| `temperature` | boolean | `true` | Supports temperature parameter | `false` | +| `tool_call` | boolean | `false` | Supports tool/function calling | `true` | +| `attachment` | boolean | `false` | Supports file attachments | `true` | +| `structured_output` | boolean | `false` | Supports structured/JSON output | `true` | +| `open_weights` | boolean | `false` | Open-weight model | `true` | +| `deprecated` | boolean | `false` | Deprecated but still accessible | `true` | +| `limit` | object | — | Token limits | `{ context: 128000, output: 16384 }` | +| `knowledge` | string | — | Training data cutoff | `2023-10` | +| `release_date` | string | — | Model release date | `2024-05-13` | +| `snapshots` | array | — | Dated model versions | See below | + +### Modality Types + +| Modality | Description | +| -------- | --------------------- | +| `text` | Text input or output | +| `image` | Image input or output | +| `video` | Video input | +| `audio` | Audio input or output | +| `pdf` | PDF document input | + +### Pricing Schema + +Pricing is a union of four types: + +**TokenPricing** (most common — per-million-token pricing): + +```yaml +pricing: + input: 2.5 # $/M input tokens + output: 10 # $/M output tokens + cache_write: 1.25 # optional + cache_read: 0.625 # optional +``` + +**VideoPricing** (per-second, optionally tiered by resolution): + +```yaml +pricing: + unit: per_second + price: 0.03 +``` + +**UnitPricing** (per-image or per-request): + +```yaml +pricing: + unit: per_image + price: 0.04 +``` + +**FreePricing** (no cost): + +```yaml +pricing: + unit: free +``` + +### Snapshot Schema + +Snapshots represent dated versions of a model. They inherit all parent fields and only override what differs: + +```yaml +id: gpt-4o +name: GPT-4o +snapshots: + - id: gpt-4o-2024-08-06 + last_updated: "2024-08-06" + - id: gpt-4o-2024-05-13 + deprecated: true + last_updated: "2024-05-13" +``` + +### Provider Schema + +Each provider has a `provider.yaml` file: + +| Field | Type | Required | Description | Example | +| ---------------- | ------ | -------- | ------------------------------------ | ---------------------------------- | +| `id` | string | ✅ | Provider ID (matches directory name) | `openai` | +| `name` | string | ✅ | Display name | `OpenAI` | +| `url` | string | ✅ | Official website URL | `https://openai.com` | +| `api_docs` | string | ❌ | API documentation URL | `https://platform.openai.com/docs` | +| `apis` | object | ✅ | API endpoints keyed by format | See below | +| `apis.openai` | string | ❌ | OpenAI-compatible API endpoint | `https://api.openai.com/v1` | +| `apis.anthropic` | string | ❌ | Anthropic API endpoint | — | +| `apis.google` | string | ❌ | Google AI API endpoint | — | +| `currency` | string | ❌ | Default currency (USD/CNY/EUR) | `USD` | + +## Quick Start + +### Find the cheapest model + +→ See [Pricing Comparison](docs/pricing-comparison.md) + +**Cheapest models with tool calling:** + +| Model | Provider | Input (per 1M tokens) | Output (per 1M tokens) | +| ---------------- | ------------- | --------------------: | ---------------------: | +| DeepSeek-V3 | DeepSeek | $0.27 | $1.10 | +| Qwen3-235B-A22B | Alibaba Cloud | $0.14 | $0.42 | +| Llama 4 Maverick | Together AI | $0.20 | $0.80 | + +### Find the most capable model + +→ See [Model Comparison](docs/model-comparison.md) + +**Top-tier flagships:** + +| Model | Context | Tool Call | Vision | Input $/1M | Output $/1M | +| -------------- | ------- | --------- | ------ | ---------: | ----------: | +| GPT-4.1 | 1M | ✅ | ✅ | $2.00 | $8.00 | +| Claude Opus 4 | 200K | ✅ | ✅ | $15.00 | $75.00 | +| Gemini 2.5 Pro | 1M | ✅ | ✅ | $1.25 | $10.00 | + +### Find a free model + +→ See [Free AI Models](docs/free-models.md) + +- Google Gemini 2.0 Flash (1M context, tool calling, vision, reasoning) +- 70+ models on Chutes (Llama 4, Qwen3, DeepSeek-R1, etc.) + +### Find the largest context window + +→ See [Context Window Comparison](docs/context-windows.md) + +| Model | Context Window | +| --------------- | -------------: | +| Llama 4 Scout | 10M tokens | +| Gemini 2.5 Pro | 1M tokens | +| GPT-4.1 | ~1M tokens | + +## Programmatic Access + +### npm package + +```bash +npm install ai-models +``` + +```typescript +import catalog from "ai-models"; // 4,587 models as JSON +import type { Model } from "ai-models"; // TypeScript types + +// Find models with tool calling under $1/1M input +const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < 1); +``` + +### Download data files + +```bash +# JSON — full metadata (2.3 MB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# CSV — flat table for Excel/Google Sheets (560 KB) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` + +### Python usage + +```python +import json + +with open("models.json") as f: + catalog = json.load(f) + +# Find all reasoning models +reasoning = [m for m in catalog["models"] if m.get("reasoning")] + +# Find models with largest context windows +by_context = sorted( + catalog["models"], + key=lambda m: (m.get("limit", {}) or {}).get("context", 0), + reverse=True, +)[:10] +``` + +### JSON structure + +```json +{ + "generated_at": "2026-05-21T02:13:04.076Z", + "stats": { + "providers": 95, + "models": 4587, + "unique_model_ids": 2712, + "families": 441 + }, + "providers": { + "openai": { "name": "OpenAI", "model_count": 28 }, + "anthropic": { "name": "Anthropic", "model_count": 11 } + }, + "models": [ + { + "id": "gpt-4.1", + "name": "GPT-4.1", + "family": "gpt-4.1", + "provider": "openai", + "tool_call": true, + "structured_output": true, + "pricing": { "input": 2, "output": 8, "cache_read": 0.5 }, + "limit": { "context": 1047576, "output": 32768 }, + "modalities": { "input": ["text", "image"], "output": ["text"] } + } + ] +} ``` ## Documentation - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds +- [Model Selection Guide](docs/model-selection.md) — decision framework: free, best value, large context models - [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq - [FAQ](docs/faq.md) — frequently asked questions about the catalog, data, and contributing @@ -92,6 +338,40 @@ modalities: - [Data Acquisition](docs/data-acquisition.md) — how we acquire and update data - [Design Principles](docs/lessons-learned.md) — lessons learned -## Covered Providers +## Design Principles + +- **First-party data only** — all model data comes from the provider's own API or website +- **Dynamic discovery** — scrape functions discover models from the source, not from hardcoded lists +- **Include deprecated, exclude retired** — deprecated models are included with `deprecated: true`; retired (inaccessible) models are excluded +- **Never fabricate data** — if required data is missing, skip the model with a warning rather than filling in guessed values +- **YAML source format** — human-readable, supports comments, machine-parseable +- **Snapshot inheritance** — dated model versions are nested within the parent model, inheriting all fields + +## Adding a New Provider + +1. Create `providers//scrape.ts` with a `scrape()` function that returns `{ provider, models }` +2. Data must come from a first-party source (provider's API or website) +3. Include a discovery step — no hardcoded model ID lists +4. Run `npx tsx scripts/sync.ts ` to generate initial data +5. Validate with `npx tsx scripts/validate.ts` + +## CLI Tools + +```bash +# Validate all YAML data +npx tsx scripts/validate.ts + +# Compute catalog statistics +npx tsx scripts/stats.ts # table format +npx tsx scripts/stats.ts json # JSON format + +# Compile to models.json +npx tsx scripts/compile.ts + +# Sync data from providers +npx tsx scripts/sync.ts openai # single provider +npx tsx scripts/sync.ts # all providers -OpenAI, Anthropic, Google, Meta, DeepSeek, Alibaba Cloud, Mistral, xAI, Cohere, NVIDIA, IBM, Microsoft, Amazon Bedrock, Azure OpenAI, Google Vertex AI, OpenRouter, Together AI, Fireworks AI, Groq, Cerebras, DeepInfra, and 75+ more. \ No newline at end of file +# Export to CSV +npx tsx scripts/export-csv.ts +``` \ No newline at end of file From 1e0079287ca63dd10eaf75718a6469c2ca71f940 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:10:16 +0800 Subject: [PATCH 145/311] Add cross-links to 10 key docs (EN + ZH) for internal SEO - free-models, tool-calling, reasoning, vision, structured-output - open-weights, context-windows, pricing-comparison, cached-pricing, quick-start - Each doc now has 'Related Documentation' section linking to 5 related docs --- docs/cached-pricing.md | 8 ++++++++ docs/context-windows.md | 8 ++++++++ docs/free-models.md | 8 ++++++++ docs/open-weights.md | 8 ++++++++ docs/pricing-comparison.md | 8 ++++++++ docs/quick-start.md | 8 ++++++++ docs/reasoning-models.md | 8 ++++++++ docs/structured-output.md | 8 ++++++++ docs/tool-calling.md | 8 ++++++++ docs/vision-models.md | 8 ++++++++ docs/zh/cached-pricing.md | 8 ++++++++ docs/zh/context-windows.md | 8 ++++++++ docs/zh/free-models.md | 8 ++++++++ docs/zh/open-weights.md | 8 ++++++++ docs/zh/pricing-comparison.md | 8 ++++++++ docs/zh/quick-start.md | 8 ++++++++ docs/zh/reasoning-models.md | 8 ++++++++ docs/zh/structured-output.md | 8 ++++++++ docs/zh/tool-calling.md | 8 ++++++++ docs/zh/vision-models.md | 8 ++++++++ 20 files changed, 160 insertions(+) diff --git a/docs/cached-pricing.md b/docs/cached-pricing.md index 0714acab..edfe5629 100644 --- a/docs/cached-pricing.md +++ b/docs/cached-pricing.md @@ -235,3 +235,11 @@ Prompt caching lets you store repeated prompt prefixes (system prompts, few-shot --- Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — cost optimization tips +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Free AI Models](free-models.md) — 81 free models +- [Context Window Comparison](context-windows.md) — largest context windows +- [Open-Weight Models](open-weights.md) — 527 models you can run yourself diff --git a/docs/context-windows.md b/docs/context-windows.md index 525dcff7..4bf6d850 100644 --- a/docs/context-windows.md +++ b/docs/context-windows.md @@ -63,3 +63,11 @@ Which models have the largest context windows? This page lists models by context - **128K context** is the most common tier (1,310 models) — sufficient for most use cases - **Cost scales with context**: 1M+ context models cost 2–10x more per token than 128K models - **Cache read pricing** can reduce costs significantly for repeated queries (up to 90% discount) + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — large context model recommendations +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Free AI Models](free-models.md) — 81 free models by context window +- [Vision Models](vision-models.md) — 1,487 vision models with context info +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching diff --git a/docs/free-models.md b/docs/free-models.md index 0caf6759..370b7421 100644 --- a/docs/free-models.md +++ b/docs/free-models.md @@ -109,3 +109,11 @@ Groq offers free tier for some models with rate limits: - **Llama 4 Scout** on Chutes offers the largest free context window at 10M tokens - Free tiers typically have rate limits (requests per minute) — check provider docs for specifics - For production use, consider upgrading to paid tiers for reliability and higher rate limits + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Open-Weight Models](open-weights.md) — 527 models you can run yourself +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning diff --git a/docs/open-weights.md b/docs/open-weights.md index 58785975..20d3f822 100644 --- a/docs/open-weights.md +++ b/docs/open-weights.md @@ -104,3 +104,11 @@ Lowest per-token pricing for open-weight inference: - **DeepSeek-R1** is the most popular open-weight reasoning model, available on 3 providers - **MiMo V2.5** is the only open-weight model combining 1M context, reasoning, and vision - Pricing varies widely — the cheapest open-weight models cost under $0.01/1M tokens + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Free AI Models](free-models.md) — 81 free models +- [Provider Overview](providers.md) — all 95 providers organized by type +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching diff --git a/docs/pricing-comparison.md b/docs/pricing-comparison.md index 91217e00..6004203a 100644 --- a/docs/pricing-comparison.md +++ b/docs/pricing-comparison.md @@ -119,3 +119,11 @@ The absolute cheapest per-token models across all providers. --- **Note**: All pricing from first-party sources as of data collection date. Inference platform prices may differ. Check `providers//models/` for current data. CNY and EUR pricing available in provider YAML files. + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [Free AI Models](free-models.md) — 81 free models +- [Context Window Comparison](context-windows.md) — largest context windows +- [Provider Overview](providers.md) — all 95 providers diff --git a/docs/quick-start.md b/docs/quick-start.md index 68daa734..0ede2c54 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -128,3 +128,11 @@ console.log(model.limit); // { context: 1047576, output: 32768 } ## I want to understand the data format → See [Data Schema Reference](data-schema.md) for the complete YAML schema. + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [FAQ](faq.md) — common questions about the catalog +- [API & Programmatic Access](api.md) — download and use the data +- [Code Examples](code-examples.md) — practical examples in TypeScript, Python, Go, Rust +- [Glossary](glossary.md) — key terms and definitions diff --git a/docs/reasoning-models.md b/docs/reasoning-models.md index e49521d3..f3d42836 100644 --- a/docs/reasoning-models.md +++ b/docs/reasoning-models.md @@ -87,3 +87,11 @@ Models that can reason about images — ideal for visual analysis: - **MiMo V2.5** is the only open-weight model combining 1M context, reasoning, and vision - **697 reasoning models** also support vision — the most common combined capability - Small reasoning models (Qwen 3.5 0.8B–4B) cost as little as $0.01–$0.03/1M tokens + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling +- [Structured Output](structured-output.md) — 829 JSON-mode models +- [Free AI Models](free-models.md) — 81 free models, some with reasoning +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching diff --git a/docs/structured-output.md b/docs/structured-output.md index cdaef929..737416b9 100644 --- a/docs/structured-output.md +++ b/docs/structured-output.md @@ -69,3 +69,11 @@ For AI agents that need to return structured data, call tools, and reason: - **Gemini 2.5 Flash** is the best value: 1M context, structured output, tool calling, and reasoning for $0.15/1M - Small models (Ernie 4.5 0.3B, Ling 2.6 Flash) cost as little as $0.01/1M with structured output - 91% of structured output models also support tool calling — these capabilities go hand-in-hand + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Free AI Models](free-models.md) — 81 free models, some with structured output +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching diff --git a/docs/tool-calling.md b/docs/tool-calling.md index 08679aaa..c2a15887 100644 --- a/docs/tool-calling.md +++ b/docs/tool-calling.md @@ -86,3 +86,11 @@ The "holy trinity" for advanced AI agents — tool calling, reasoning, and visio - **45 free models** support tool calling — start building agents at zero cost - **829 models** also support structured output — perfect for reliable JSON responses - Small models (Qwen 3.5 0.8B–4B) cost as little as $0.01–$0.03/1M tokens with tool calling + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Free AI Models](free-models.md) — 81 free models, many with tool calling +- [Structured Output](structured-output.md) — 829 JSON-mode models +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching diff --git a/docs/vision-models.md b/docs/vision-models.md index 1fdca6fa..d94af2ab 100644 --- a/docs/vision-models.md +++ b/docs/vision-models.md @@ -85,3 +85,11 @@ The most capable vision models — can see, reason, and act: - **Grok 4 Fast Reasoning** is the only model combining 2M context, vision, tool calling, and reasoning - **104 open-weight vision models** available — run vision AI on your own infrastructure - Small vision models (Qwen 3.5 0.8B–4B) cost as little as $0.01–$0.03/1M tokens + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Image Generation](image-generation.md) — 28 image generation models +- [Video Models](video-models.md) — 167 video input/output models +- [Modality Matrix](modality-matrix.md) — all modalities at a glance +- [Free AI Models](free-models.md) — 81 free models, some with vision diff --git a/docs/zh/cached-pricing.md b/docs/zh/cached-pricing.md index c2701e50..0b87ce23 100644 --- a/docs/zh/cached-pricing.md +++ b/docs/zh/cached-pricing.md @@ -235,3 +235,11 @@ --- 数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 成本优化技巧 +- [定价对比](pricing-comparison.md) — 各提供商定价并排对比 +- [免费 AI 模型](free-models.md) — 81 个免费模型 +- [上下文窗口对比](context-windows.md) — 最大上下文窗口 +- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 diff --git a/docs/zh/context-windows.md b/docs/zh/context-windows.md index 5a194d82..e49cdf89 100644 --- a/docs/zh/context-windows.md +++ b/docs/zh/context-windows.md @@ -63,3 +63,11 @@ - **128K 上下文**是最常见的层级（1,310 个模型）— 足以满足大多数用例 - **成本随上下文增长**：1M+ 上下文模型的每 token 成本是 128K 模型的 2–10 倍 - **缓存读取定价**可显著降低重复查询的成本（最高 90% 折扣） + +## 相关文档 + +- [模型选择指南](model-selection.md) — 大上下文模型推荐 +- [定价对比](pricing-comparison.md) — 各提供商定价并排对比 +- [免费 AI 模型](free-models.md) — 81 个免费模型按上下文窗口分类 +- [视觉模型](vision-models.md) — 1,487 个视觉模型含上下文信息 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 diff --git a/docs/zh/free-models.md b/docs/zh/free-models.md index 8c0dc0ef..7ae9a455 100644 --- a/docs/zh/free-models.md +++ b/docs/zh/free-models.md @@ -109,3 +109,11 @@ Groq 为部分模型提供免费层： - **Llama 4 Scout** 在 Chutes 上提供最大的免费上下文窗口（10M tokens） - 免费层通常有速率限制（每分钟请求数）— 请查看提供商文档了解具体限制 - 生产环境建议升级到付费层以获得可靠性和更高的速率限制 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 diff --git a/docs/zh/open-weights.md b/docs/zh/open-weights.md index fc1d0872..70cce2e3 100644 --- a/docs/zh/open-weights.md +++ b/docs/zh/open-weights.md @@ -104,3 +104,11 @@ - **DeepSeek-R1** 是最受欢迎的开源权重推理模型，在 3 个提供商上可用 - **MiMo V2.5** 是唯一结合 1M 上下文、推理和视觉的开源权重模型 - 定价差异很大 — 最便宜的开源权重模型每 1M tokens 不到 $0.01 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [免费 AI 模型](free-models.md) — 81 个免费模型 +- [提供商概览](providers.md) — 95 个提供商按类型分类 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 diff --git a/docs/zh/pricing-comparison.md b/docs/zh/pricing-comparison.md index 8a690853..a86991d3 100644 --- a/docs/zh/pricing-comparison.md +++ b/docs/zh/pricing-comparison.md @@ -119,3 +119,11 @@ --- **注意**：所有定价来自第一方来源，以数据采集日期为准。推理平台价格可能不同。查看 `providers//models/` 获取最新数据。人民币和欧元定价见提供商 YAML 文件。 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型 +- [上下文窗口对比](context-windows.md) — 最大上下文窗口 +- [提供商概览](providers.md) — 95 个提供商 diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md index 4a9821b6..e8c700e1 100644 --- a/docs/zh/quick-start.md +++ b/docs/zh/quick-start.md @@ -128,3 +128,11 @@ console.log(model.limit); // { context: 1047576, output: 32768 } ## 我想了解数据格式 → 查看[数据 Schema 参考](data-schema.md)，了解完整的 YAML Schema。 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [常见问题](faq.md) — 关于目录的常见问题 +- [API 与编程访问](api.md) — 下载和使用数据 +- [代码示例](code-examples.md) — TypeScript、Python、Go、Rust 实用示例 +- [术语表](glossary.md) — 关键术语和定义 diff --git a/docs/zh/reasoning-models.md b/docs/zh/reasoning-models.md index 276bc175..180b787b 100644 --- a/docs/zh/reasoning-models.md +++ b/docs/zh/reasoning-models.md @@ -87,3 +87,11 @@ - **MiMo V2.5** 是唯一结合 1M 上下文、推理和视觉的开源权重模型 - **697 个推理模型**同时支持视觉 — 最常见的组合能力 - 小型推理模型（Qwen 3.5 0.8B–4B）每 1M tokens 仅需 $0.01–$0.03 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 +- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持推理 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 diff --git a/docs/zh/structured-output.md b/docs/zh/structured-output.md index 10bc2bf6..2e996435 100644 --- a/docs/zh/structured-output.md +++ b/docs/zh/structured-output.md @@ -69,3 +69,11 @@ - **Gemini 2.5 Flash** 是最佳性价比：1M 上下文、结构化输出、工具调用和推理，仅 $0.15/1M - 小型模型（Ernie 4.5 0.3B、Ling 2.6 Flash）带结构化输出仅需 $0.01/1M - 91% 的结构化输出模型同时支持工具调用 — 这两种能力相辅相成 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持结构化输出 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 diff --git a/docs/zh/tool-calling.md b/docs/zh/tool-calling.md index e37cb9ad..360f38bf 100644 --- a/docs/zh/tool-calling.md +++ b/docs/zh/tool-calling.md @@ -86,3 +86,11 @@ - **45 个免费模型**支持工具调用 — 零成本开始构建代理 - **829 个模型**同时支持结构化输出 — 完美适合可靠的 JSON 响应 - 小型模型（Qwen 3.5 0.8B–4B）带工具调用仅需 $0.01–$0.03/1M tokens + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [免费 AI 模型](free-models.md) — 81 个免费模型，多数支持工具调用 +- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 diff --git a/docs/zh/vision-models.md b/docs/zh/vision-models.md index 06e6fc5c..ed394e7e 100644 --- a/docs/zh/vision-models.md +++ b/docs/zh/vision-models.md @@ -85,3 +85,11 @@ - **Grok 4 Fast Reasoning** 是唯一结合 2M 上下文、视觉、工具调用和推理的模型 - **104 个开源权重视觉模型** — 在自己的基础设施上运行视觉 AI - 小型视觉模型（Qwen 3.5 0.8B–4B）每 1M tokens 仅需 $0.01–$0.03 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [图像生成](image-generation.md) — 28 个图像生成模型 +- [视频模型](video-models.md) — 167 个视频输入/输出模型 +- [模态矩阵](modality-matrix.md) — 所有模态一览 +- [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持视觉 From ebd9b63d6aef367ad32073ca8277ee4fc4ea173c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:17:57 +0800 Subject: [PATCH 146/311] Add Star CTA to README hero section --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index 06a8eeb0..939442b2 100644 --- a/README.md +++ b/README.md @@ -15,6 +15,8 @@
+> ⭐ **If this catalog helps you choose the right model, please star this repo — it helps others discover it!** + --- Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. From 8a3317acc8d7e7c0dd743ecab4344795590b7ed9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:20:00 +0800 Subject: [PATCH 147/311] Add cross-links to remaining 13 docs (EN + ZH) for internal SEO - audio-models, video-models, image-generation, modality-matrix - model-comparison, providers, api, code-examples - data-schema, data-acquisition, faq, glossary, lessons-learned - All 24 docs now have 'Related Documentation' sections --- docs/api.md | 8 ++++++++ docs/audio-models.md | 8 ++++++++ docs/code-examples.md | 8 ++++++++ docs/data-acquisition.md | 8 ++++++++ docs/data-schema.md | 8 ++++++++ docs/faq.md | 8 ++++++++ docs/glossary.md | 8 ++++++++ docs/image-generation.md | 8 ++++++++ docs/lessons-learned.md | 8 ++++++++ docs/modality-matrix.md | 8 ++++++++ docs/model-comparison.md | 8 ++++++++ docs/providers.md | 8 ++++++++ docs/video-models.md | 8 ++++++++ docs/zh/api.md | 8 ++++++++ docs/zh/audio-models.md | 8 ++++++++ docs/zh/code-examples.md | 8 ++++++++ docs/zh/data-acquisition.md | 8 ++++++++ docs/zh/data-schema.md | 8 ++++++++ docs/zh/faq.md | 8 ++++++++ docs/zh/glossary.md | 8 ++++++++ docs/zh/image-generation.md | 8 ++++++++ docs/zh/lessons-learned.md | 8 ++++++++ docs/zh/modality-matrix.md | 8 ++++++++ docs/zh/model-comparison.md | 8 ++++++++ docs/zh/providers.md | 8 ++++++++ docs/zh/video-models.md | 8 ++++++++ 26 files changed, 208 insertions(+) diff --git a/docs/api.md b/docs/api.md index b26c356a..b3fbc68a 100644 --- a/docs/api.md +++ b/docs/api.md @@ -208,3 +208,11 @@ npx tsx scripts/compile.ts npx tsx scripts/sync.ts openai # single provider npx tsx scripts/sync.ts # all providers ``` + +## Related Documentation + +- [Quick Start](quick-start.md) — find the right model in 30 seconds +- [Code Examples](code-examples.md) — TypeScript, Python, Go, Rust, jq +- [Data Schema](data-schema.md) — complete YAML schema reference +- [FAQ](faq.md) — common questions +- [Model Selection Guide](model-selection.md) — decision framework diff --git a/docs/audio-models.md b/docs/audio-models.md index 502f2d6e..d0ef50bd 100644 --- a/docs/audio-models.md +++ b/docs/audio-models.md @@ -165,3 +165,11 @@ AI models that support audio input or output, sourced from the [AI Models Catalo --- Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. + +## Related Documentation + +- [Video Models](video-models.md) — 167 video input/output models +- [Vision Models](vision-models.md) — 1,487 vision models +- [Modality Matrix](modality-matrix.md) — all modalities at a glance +- [Model Selection Guide](model-selection.md) — decision framework +- [Free AI Models](free-models.md) — 81 free models diff --git a/docs/code-examples.md b/docs/code-examples.md index 68522147..7bcdb7bc 100644 --- a/docs/code-examples.md +++ b/docs/code-examples.md @@ -357,3 +357,11 @@ const results = findModels({ maxOutputPrice: 5, }); ``` + +## Related Documentation + +- [API & Programmatic Access](api.md) — npm, CDN, CSV, GitHub Action +- [Quick Start](quick-start.md) — find the right model in 30 seconds +- [Data Schema](data-schema.md) — complete YAML schema reference +- [FAQ](faq.md) — common questions +- [Glossary](glossary.md) — key terms and definitions diff --git a/docs/data-acquisition.md b/docs/data-acquisition.md index 1cb80259..7129e2bf 100644 --- a/docs/data-acquisition.md +++ b/docs/data-acquisition.md @@ -256,3 +256,11 @@ Providers that host and serve models produced by others. They are added **after 2. Updates the YAML file directly 3. Sets `last_updated` to current date 4. Validates with `npm run validate` + +## Related Documentation + +- [Data Schema](data-schema.md) — complete YAML schema reference +- [Design Principles](lessons-learned.md) — lessons learned +- [Provider Overview](providers.md) — all 95 providers +- [FAQ](faq.md) — common questions +- [Contributing](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — how to contribute diff --git a/docs/data-schema.md b/docs/data-schema.md index b9f9bd6a..7d791fa6 100644 --- a/docs/data-schema.md +++ b/docs/data-schema.md @@ -207,3 +207,11 @@ npx tsx scripts/validate.ts openai ``` The validation uses `ModelSchema` from [`types/schemas.ts`](../types/schemas.ts), which mirrors the TypeScript types exactly. Any YAML file that doesn't conform to the schema will produce a validation error with the specific field path and issue. + +## Related Documentation + +- [Data Acquisition](data-acquisition.md) — how we acquire and update data +- [API & Programmatic Access](api.md) — npm, CDN, CSV access +- [Code Examples](code-examples.md) — practical code examples +- [Design Principles](lessons-learned.md) — lessons learned +- [FAQ](faq.md) — common questions diff --git a/docs/faq.md b/docs/faq.md index 26b9279b..07056477 100644 --- a/docs/faq.md +++ b/docs/faq.md @@ -104,3 +104,11 @@ Absolutely! You can: --- More questions? [Open an issue](https://github.com/i-need-token/ai-models/issues/new) or start a [discussion](https://github.com/i-need-token/ai-models/discussions). + +## Related Documentation + +- [Quick Start](quick-start.md) — find the right model in 30 seconds +- [API & Programmatic Access](api.md) — npm, CDN, CSV access +- [Glossary](glossary.md) — key terms and definitions +- [Data Schema](data-schema.md) — complete YAML schema +- [Model Selection Guide](model-selection.md) — decision framework diff --git a/docs/glossary.md b/docs/glossary.md index 81d6f3a4..a27ae746 100644 --- a/docs/glossary.md +++ b/docs/glossary.md @@ -69,3 +69,11 @@ A quick reference for terms used throughout the AI Models Catalog. --- See [Data Schema Reference](data-schema.md) for the complete YAML field specification. + +## Related Documentation + +- [FAQ](faq.md) — common questions +- [Data Schema](data-schema.md) — complete YAML schema reference +- [Quick Start](quick-start.md) — find the right model in 30 seconds +- [Model Comparison](model-comparison.md) — compare models +- [Modality Matrix](modality-matrix.md) — all modalities at a glance diff --git a/docs/image-generation.md b/docs/image-generation.md index bf049bb0..8d244a35 100644 --- a/docs/image-generation.md +++ b/docs/image-generation.md @@ -56,3 +56,11 @@ - **GPT-5 Image Mini** offers the best combination of reasoning + generation + large context - **Amazon Nova 2.0 Omni** is the only model that generates images from audio and video input - Most image generation models accept both text and image input (for editing/reference) + +## Related Documentation + +- [Vision Models](vision-models.md) — 1,487 vision models +- [Video Models](video-models.md) — 167 video input/output models +- [Modality Matrix](modality-matrix.md) — all modalities at a glance +- [Model Selection Guide](model-selection.md) — decision framework +- [Free AI Models](free-models.md) — 81 free models diff --git a/docs/lessons-learned.md b/docs/lessons-learned.md index 1ebea94a..60e42c68 100644 --- a/docs/lessons-learned.md +++ b/docs/lessons-learned.md @@ -298,3 +298,11 @@ Platforms like OpenRouter and nano-gpt are router/aggregators that don't produce - **nano-gpt**: Public API for model list + JS bundle for per-token USD pricing (555 models). These are treated as inference platforms, not rejected as "just routers", because they provide verifiable first-party per-token pricing data. + +## Related Documentation + +- [Data Acquisition](data-acquisition.md) — how we acquire and update data +- [Data Schema](data-schema.md) — complete YAML schema reference +- [Provider Overview](providers.md) — all 95 providers +- [FAQ](faq.md) — common questions +- [Contributing](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — how to contribute diff --git a/docs/modality-matrix.md b/docs/modality-matrix.md index dc2fff3c..d40d40e0 100644 --- a/docs/modality-matrix.md +++ b/docs/modality-matrix.md @@ -92,3 +92,11 @@ Models that accept text + at least 2 additional input modalities: | Gemini 2.5 Pro | Google | text, image, audio, video | | Claude Opus 4 | Anthropic | text, image, audio | | Qwen2-Audio | Alibaba Cloud | text, image, audio | + +## Related Documentation + +- [Vision Models](vision-models.md) — 1,487 vision models +- [Audio Models](audio-models.md) — 118 audio input + 34 audio output models +- [Video Models](video-models.md) — 167 video input/output models +- [Image Generation](image-generation.md) — 28 image generation models +- [Model Selection Guide](model-selection.md) — decision framework diff --git a/docs/model-comparison.md b/docs/model-comparison.md index 91caa6fa..9813c168 100644 --- a/docs/model-comparison.md +++ b/docs/model-comparison.md @@ -95,3 +95,11 @@ Models with publicly available weights for self-hosting. --- **Note**: All pricing and capability data is from first-party sources. Prices may vary on inference platforms. Check `providers//models/` for the most current data. + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Free AI Models](free-models.md) — 81 free models +- [Open-Weight Models](open-weights.md) — 527 models you can run yourself +- [Context Window Comparison](context-windows.md) — largest context windows diff --git a/docs/providers.md b/docs/providers.md index 3e544e25..eafba4ce 100644 --- a/docs/providers.md +++ b/docs/providers.md @@ -158,3 +158,11 @@ Providers with EUR pricing, serving the European market. | Privatemode AI | `privatemode` | 5 | | Regolo | `regolo` | — | | Scaleway | `scaleway` | 13 | + +## Related Documentation + +- [Model Comparison](model-comparison.md) — flagship, cost-effective, free models +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing +- [Open-Weight Models](open-weights.md) — 527 models you can run yourself +- [Free AI Models](free-models.md) — 81 free models +- [Data Schema](data-schema.md) — complete YAML schema diff --git a/docs/video-models.md b/docs/video-models.md index 31e456d0..2c34019f 100644 --- a/docs/video-models.md +++ b/docs/video-models.md @@ -200,3 +200,11 @@ AI models that support video input or output, sourced from the [AI Models Catalo --- Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. + +## Related Documentation + +- [Audio Models](audio-models.md) — 118 audio input + 34 audio output models +- [Vision Models](vision-models.md) — 1,487 vision models +- [Image Generation](image-generation.md) — 28 image generation models +- [Modality Matrix](modality-matrix.md) — all modalities at a glance +- [Model Selection Guide](model-selection.md) — decision framework diff --git a/docs/zh/api.md b/docs/zh/api.md index c63bc61c..e48cd8cf 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -208,3 +208,11 @@ npx tsx scripts/compile.ts npx tsx scripts/sync.ts openai # 单个提供商 npx tsx scripts/sync.ts # 所有提供商 ``` + +## 相关文档 + +- [快速入门](quick-start.md) — 30 秒内找到适合的模型 +- [代码示例](code-examples.md) — TypeScript、Python、Go、Rust、jq +- [数据模式](data-schema.md) — 完整 YAML 模式参考 +- [常见问题](faq.md) — 常见问题 +- [模型选择指南](model-selection.md) — 决策框架 diff --git a/docs/zh/audio-models.md b/docs/zh/audio-models.md index 90d344dc..f01dd1f3 100644 --- a/docs/zh/audio-models.md +++ b/docs/zh/audio-models.md @@ -165,3 +165,11 @@ --- 数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 + +## 相关文档 + +- [视频模型](video-models.md) — 167 个视频输入/输出模型 +- [视觉模型](vision-models.md) — 1,487 个视觉模型 +- [模态矩阵](modality-matrix.md) — 所有模态一览 +- [模型选择指南](model-selection.md) — 决策框架 +- [免费 AI 模型](free-models.md) — 81 个免费模型 diff --git a/docs/zh/code-examples.md b/docs/zh/code-examples.md index f84b7a24..61057166 100644 --- a/docs/zh/code-examples.md +++ b/docs/zh/code-examples.md @@ -357,3 +357,11 @@ const results = findModels({ maxOutputPrice: 5, }); ``` + +## 相关文档 + +- [API 与编程访问](api.md) — npm、CDN、CSV、GitHub Action +- [快速入门](quick-start.md) — 30 秒内找到适合的模型 +- [数据模式](data-schema.md) — 完整 YAML 模式参考 +- [常见问题](faq.md) — 常见问题 +- [术语表](glossary.md) — 关键术语和定义 diff --git a/docs/zh/data-acquisition.md b/docs/zh/data-acquisition.md index 83578930..1dc15c56 100644 --- a/docs/zh/data-acquisition.md +++ b/docs/zh/data-acquisition.md @@ -256,3 +256,11 @@ parseModality | toLowerCase | toUpperCase | trim | removeCommas | identity 2. 直接更新 YAML 文件 3. 将 `last_updated` 设为当前日期 4. 用 `npm run validate` 验证 + +## 相关文档 + +- [数据模式](data-schema.md) — 完整 YAML 模式参考 +- [设计原则](lessons-learned.md) — 经验教训 +- [提供商概览](providers.md) — 95 个提供商 +- [常见问题](faq.md) — 常见问题 +- [贡献指南](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — 如何贡献 diff --git a/docs/zh/data-schema.md b/docs/zh/data-schema.md index 9fdd8cdf..8e5ed47f 100644 --- a/docs/zh/data-schema.md +++ b/docs/zh/data-schema.md @@ -207,3 +207,11 @@ npx tsx scripts/validate.ts openai ``` 校验使用 [`types/schemas.ts`](../../types/schemas.ts) 中的 `ModelSchema`，与 TypeScript 类型完全对应。任何不符合 schema 的 YAML 文件将产生校验错误，包含具体的字段路径和问题。 + +## 相关文档 + +- [数据获取](data-acquisition.md) — 如何获取和更新数据 +- [API 与编程访问](api.md) — npm、CDN、CSV 访问 +- [代码示例](code-examples.md) — 实用代码示例 +- [设计原则](lessons-learned.md) — 经验教训 +- [常见问题](faq.md) — 常见问题 diff --git a/docs/zh/faq.md b/docs/zh/faq.md index 4460d5e2..a26dd490 100644 --- a/docs/zh/faq.md +++ b/docs/zh/faq.md @@ -104,3 +104,11 @@ npx ajv validate -s schema.json -d providers/openai/models/gpt-4o.yaml --- 更多问题？[提交 issue](https://github.com/i-need-token/ai-models/issues/new) 或发起[讨论](https://github.com/i-need-token/ai-models/discussions)。 + +## 相关文档 + +- [快速入门](quick-start.md) — 30 秒内找到适合的模型 +- [API 与编程访问](api.md) — npm、CDN、CSV 访问 +- [术语表](glossary.md) — 关键术语和定义 +- [数据模式](data-schema.md) — 完整 YAML 模式 +- [模型选择指南](model-selection.md) — 决策框架 diff --git a/docs/zh/glossary.md b/docs/zh/glossary.md index 3be4f222..c040f22d 100644 --- a/docs/zh/glossary.md +++ b/docs/zh/glossary.md @@ -69,3 +69,11 @@ AI Models Catalog 中使用的术语快速参考。 --- 详见[数据 Schema 参考](data-schema.md)获取完整的 YAML 字段规范。 + +## 相关文档 + +- [常见问题](faq.md) — 常见问题 +- [数据模式](data-schema.md) — 完整 YAML 模式参考 +- [快速入门](quick-start.md) — 30 秒内找到适合的模型 +- [模型对比](model-comparison.md) — 模型对比 +- [模态矩阵](modality-matrix.md) — 所有模态一览 diff --git a/docs/zh/image-generation.md b/docs/zh/image-generation.md index 4fd17c1d..cd66d0f7 100644 --- a/docs/zh/image-generation.md +++ b/docs/zh/image-generation.md @@ -56,3 +56,11 @@ - **GPT-5 Image Mini** 提供推理 + 生成 + 大上下文的最佳组合 - **Amazon Nova 2.0 Omni** 是唯一可以从音频和视频输入生成图像的模型 - 大多数图像生成模型同时接受文本和图像输入（用于编辑/参考） + +## 相关文档 + +- [视觉模型](vision-models.md) — 1,487 个视觉模型 +- [视频模型](video-models.md) — 167 个视频输入/输出模型 +- [模态矩阵](modality-matrix.md) — 所有模态一览 +- [模型选择指南](model-selection.md) — 决策框架 +- [免费 AI 模型](free-models.md) — 81 个免费模型 diff --git a/docs/zh/lessons-learned.md b/docs/zh/lessons-learned.md index 16c05433..93af3c8c 100644 --- a/docs/zh/lessons-learned.md +++ b/docs/zh/lessons-learned.md @@ -290,3 +290,11 @@ - **nano-gpt**：公开 API 获取模型列表 + JS 包获取按 token USD 定价（555 个模型）。 这些平台被视为推理平台，而非被拒绝为“只是路由器”，因为它们提供可验证的第一方按 token 定价数据。 + +## 相关文档 + +- [数据获取](data-acquisition.md) — 如何获取和更新数据 +- [数据模式](data-schema.md) — 完整 YAML 模式参考 +- [提供商概览](providers.md) — 95 个提供商 +- [常见问题](faq.md) — 常见问题 +- [贡献指南](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — 如何贡献 diff --git a/docs/zh/modality-matrix.md b/docs/zh/modality-matrix.md index 91d77e07..11864322 100644 --- a/docs/zh/modality-matrix.md +++ b/docs/zh/modality-matrix.md @@ -92,3 +92,11 @@ | Gemini 2.5 Pro | Google | 文本、图像、音频、视频 | | Claude Opus 4 | Anthropic | 文本、图像、音频 | | Qwen2-Audio | 阿里云 | 文本、图像、音频 | + +## 相关文档 + +- [视觉模型](vision-models.md) — 1,487 个视觉模型 +- [音频模型](audio-models.md) — 118 个音频输入 + 34 个音频输出模型 +- [视频模型](video-models.md) — 167 个视频输入/输出模型 +- [图像生成](image-generation.md) — 28 个图像生成模型 +- [模型选择指南](model-selection.md) — 决策框架 diff --git a/docs/zh/model-comparison.md b/docs/zh/model-comparison.md index 3fc8a937..6c31659e 100644 --- a/docs/zh/model-comparison.md +++ b/docs/zh/model-comparison.md @@ -95,3 +95,11 @@ --- **注意**：所有定价和能力数据来自第一方来源。推理平台价格可能不同。查看 `providers//models/` 获取最新数据。 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [定价对比](pricing-comparison.md) — 各提供商定价并排对比 +- [免费 AI 模型](free-models.md) — 81 个免费模型 +- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 +- [上下文窗口对比](context-windows.md) — 最大上下文窗口 diff --git a/docs/zh/providers.md b/docs/zh/providers.md index 2911a3c8..eee37806 100644 --- a/docs/zh/providers.md +++ b/docs/zh/providers.md @@ -158,3 +158,11 @@ | Privatemode AI | `privatemode` | 5 | | Regolo | `regolo` | — | | Scaleway | `scaleway` | 13 | + +## 相关文档 + +- [模型对比](model-comparison.md) — 旗舰、高性价比、免费模型 +- [定价对比](pricing-comparison.md) — 定价并排对比 +- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型 +- [数据模式](data-schema.md) — 完整 YAML 模式 diff --git a/docs/zh/video-models.md b/docs/zh/video-models.md index 749b83f9..a66fb3ed 100644 --- a/docs/zh/video-models.md +++ b/docs/zh/video-models.md @@ -200,3 +200,11 @@ --- 数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 + +## 相关文档 + +- [音频模型](audio-models.md) — 118 个音频输入 + 34 个音频输出模型 +- [视觉模型](vision-models.md) — 1,487 个视觉模型 +- [图像生成](image-generation.md) — 28 个图像生成模型 +- [模态矩阵](modality-matrix.md) — 所有模态一览 +- [模型选择指南](model-selection.md) — 决策框架 From c8ef85a73023fadd808d4dff2d306094884281aa Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:20:40 +0800 Subject: [PATCH 148/311] Add 8 search-discoverability keywords to package.json (46 total) - model-selection, prompt-caching, cached-pricing - llm-catalog, model-directory, ai-directory - ai-model-pricing, model-metadata --- package.json | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/package.json b/package.json index 450981f1..b9537fd5 100644 --- a/package.json +++ b/package.json @@ -4,10 +4,13 @@ "description": "Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation.", "keywords": [ "ai-catalog", + "ai-directory", + "ai-model-pricing", "ai-models", "ai-pricing", "ai-provider", "anthropic", + "cached-pricing", "cdn", "cerebras", "chatgpt", @@ -27,14 +30,19 @@ "llama", "llm", "llm-api", + "llm-catalog", "llm-pricing", "machine-readable", "mistral", "model-catalog", "model-comparison", + "model-directory", + "model-metadata", "model-pricing", + "model-selection", "open-weights", "openai", + "prompt-caching", "qwen", "reasoning-models", "structured-output", From eae2c018ff746e0463faf41075efa541fa808e74 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:21:12 +0800 Subject: [PATCH 149/311] Add 6 keywords to CITATION.cff: model-selection, prompt-caching, cached-pricing, audio-models, video-models, modality-matrix --- CITATION.cff | 6 ++++++ site/index.html | 22 ++++++++++++++++++++++ 2 files changed, 28 insertions(+) diff --git a/CITATION.cff b/CITATION.cff index a71136ee..1c094492 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -29,6 +29,12 @@ keywords: - github-action - cdn - huggingface + - model-selection + - prompt-caching + - cached-pricing + - audio-models + - video-models + - modality-matrix license: MIT version: 0.1.0 date-released: "2026-05-21" diff --git a/site/index.html b/site/index.html index b340eae4..c1394941 100644 --- a/site/index.html +++ b/site/index.html @@ -394,6 +394,28 @@

🤖 AI Models CatalogGitHub Action · npm · Hugging Face +
+ Model Selection + · + Free Models + · + Tool Calling + · + Pricing + · + Quick Start + · API · + FAQ

From 6c5f6cfce5ac00830f3262d281c7f4e71f43186b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:24:22 +0800 Subject: [PATCH 150/311] Add model-selection.md link to AGENTS.md (was missing) --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index ac5cd37a..e9912393 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -27,6 +27,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/cached-pricing.md`](docs/cached-pricing.md) — Models with prompt caching, 50-90% input cost savings ([中文](docs/zh/cached-pricing.md)) - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) - [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) +- [`docs/model-selection.md`](docs/model-selection.md) — Model selection guide: free, best value, large context ([中文](docs/zh/model-selection.md)) - [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) - [`docs/code-examples.md`](docs/code-examples.md) — code examples in multiple languages ([中文](docs/zh/code-examples.md)) - [`docs/faq.md`](docs/faq.md) — frequently asked questions ([中文](docs/zh/faq.md)) From 1e6dd796f894891bd802b5861698c83b41dcbc8e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:25:39 +0800 Subject: [PATCH 151/311] =?UTF-8?q?Add=20Alternatives=20section=20to=20REA?= =?UTF-8?q?DME=20=E2=80=94=20comparison=20table=20with=20key=20differentia?= =?UTF-8?q?tors?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Compares scope, data source, format, auto-update, and pricing - Highlights first-party data, machine-readable YAML, multiple access formats - Helps users understand why this catalog is unique --- README.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/README.md b/README.md index 939442b2..f4aac02f 100644 --- a/README.md +++ b/README.md @@ -489,6 +489,26 @@ Contributions are welcome! Whether it's adding a new provider, fixing data, or i Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. +## Alternatives + +| Project | Scope | Data Source | Format | Auto-Update | Free | +| -------------------------------------------------------- | --------------------------- | ---------------- | ----------------- | ----------- | ------- | +| **This catalog** | 95 providers, 4,587+ models | First-party APIs | YAML + JSON + CSV | Weekly CI | ✅ | +| [OpenRouter models](https://openrouter.ai/models) | OpenRouter only | OpenRouter API | Web UI | ✅ | ✅ | +| [Artificial Analysis](https://artificialanalysis.ai/) | ~30 providers | Mixed | Web UI | ✅ | Partial | +| [Helicone models](https://helicone.ai/models) | ~20 providers | Mixed | Web UI | ✅ | Partial | +| [BerriAI/litellm](https://github.com/BerriAI/litellm) | 100+ providers | Community | Python config | ✅ | ✅ | +| [openai/models](https://github.com/openai/openai-python) | OpenAI only | OpenAI API | Python SDK | ✅ | ✅ | + +**Key differentiators of this catalog:** + +- **First-party data only** — scraped directly from provider APIs, not aggregated from third parties +- **Machine-readable YAML** — structured data with Zod validation, not just a web UI +- **Multiple access formats** — npm, CDN, CSV, GitHub Action, Hugging Face dataset +- **Comprehensive metadata** — pricing, context windows, modalities, capabilities, snapshots +- **Bilingual docs** — 24 English + 24 Chinese documentation pages +- **Open data** — all model data is open and programmatically accessible + ## Roadmap - 🔜 **More providers** — continuously adding new AI model providers From 7d77a2aebf5b8f1ac49b9b1f423408608fbf43f3 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:26:39 +0800 Subject: [PATCH 152/311] Add 3 more use cases to README: model selection, cost optimization, free prototyping --- README.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/README.md b/README.md index f4aac02f..1b2dd886 100644 --- a/README.md +++ b/README.md @@ -101,6 +101,9 @@ Machine-readable YAML catalog of every major AI model provider and their models | 📊 **Track the AI landscape** | 2,712 models with release dates, deprecation status | | 🤖 **Power an AI tool** | TypeScript types + Zod validation = type-safe access | | 🌍 **Find local/EU providers** | [Provider overview](docs/providers.md) with market segmentation | +| 🎯 **Choose the right model** | [Model selection guide](docs/model-selection.md) — decision framework | +| 💸 **Optimize API costs** | [Cached pricing](docs/cached-pricing.md) — 1,374 models with 50-90% savings | +| 🧪 **Prototype for free** | [Free models](docs/free-models.md) — 81 models at zero cost | ## Quick Numbers From 087adeca0170e986108ea8e2340d448346f9f12e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:27:17 +0800 Subject: [PATCH 153/311] Add 'Choose a model' link to README hero quick links, streamline labels --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 1b2dd886..4baea67a 100644 --- a/README.md +++ b/README.md @@ -21,7 +21,7 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Quick start →](docs/quick-start.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[Browse providers →](docs/providers.md)** · **[Data schema →](docs/data-schema.md)** · **[🔍 Search models →](https://i-need-token.github.io/ai-models/)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN access →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** +**[Quick start →](docs/quick-start.md)** · **[Choose a model →](docs/model-selection.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[🔍 Search →](https://i-need-token.github.io/ai-models/)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** > 💡 **Try it now** — fetch model data in one command: > From c087a28b52f95dc219672b1bf9873dc406643e30 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:28:55 +0800 Subject: [PATCH 154/311] Add GitHub stars badge to README header --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 4baea67a..9e439e4f 100644 --- a/README.md +++ b/README.md @@ -11,7 +11,7 @@ [![Hugging Face](https://img.shields.io/badge/%F0%9F%A4%97-Dataset-blue)](https://huggingface.co/datasets/i-need-token/ai-models) [![Models](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/models.json)](providers/) [![Providers](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/providers.json)](providers/) -[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) [![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) +[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) [![GitHub stars](https://img.shields.io/github/stars/i-need-token/ai-models?style=social)](https://github.com/i-need-token/ai-models/stargazers) [![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From ce6b7f6e8cef1d7372bf5ac8042ec8633662b9f9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:29:19 +0800 Subject: [PATCH 155/311] =?UTF-8?q?Add=20'Last=20Updated'=20badge=20to=20R?= =?UTF-8?q?EADME=20=E2=80=94=20shows=20data=20freshness?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 9e439e4f..c4083443 100644 --- a/README.md +++ b/README.md @@ -11,7 +11,7 @@ [![Hugging Face](https://img.shields.io/badge/%F0%9F%A4%97-Dataset-blue)](https://huggingface.co/datasets/i-need-token/ai-models) [![Models](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/models.json)](providers/) [![Providers](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/i-need-token/ai-models/main/.github/badges/providers.json)](providers/) -[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) [![GitHub stars](https://img.shields.io/github/stars/i-need-token/ai-models?style=social)](https://github.com/i-need-token/ai-models/stargazers) [![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) +[![CI](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml/badge.svg)](https://github.com/i-need-token/ai-models/actions/workflows/validate.yml) [![GitHub stars](https://img.shields.io/github/stars/i-need-token/ai-models?style=social)](https://github.com/i-need-token/ai-models/stargazers) [![Last Updated](https://img.shields.io/github/last-commit/i-need-token/ai-models?label=updated)](https://github.com/i-need-token/ai-models/commits/main) [![Star History](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From 462e379c2ccbd91f48faa0fc645b8f3dfd3efb6a Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:29:49 +0800 Subject: [PATCH 156/311] Add .all-contributorsrc for contributor recognition --- .all-contributorsrc | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) create mode 100644 .all-contributorsrc diff --git a/.all-contributorsrc b/.all-contributorsrc new file mode 100644 index 00000000..1a505349 --- /dev/null +++ b/.all-contributorsrc @@ -0,0 +1,20 @@ +{ + "files": ["README.md"], + "imageSize": 100, + "contributorsPerLine": 7, + "contributorsSortAlphabetically": true, + "badgeTemplate": "[![All Contributors](https://img.shields.io/badge/all_contributors-<%= contributors.length %>-orange.svg?style=flat-square)](#contributors-)", + "contributorTemplate": "\">\" width=\"<%= options.imageSize %>px;\" alt=\"<%= contributor.name %>\"/>
<%= contributor.name %>", + "types": { + "code": { "symbol": "💻", "description": "Code" }, + "data": { "symbol": "🔣", "description": "Data" }, + "doc": { "symbol": "📖", "description": "Documentation" }, + "ideas": { "symbol": "🤔", "description": "Ideas, Planning, & Feedback" }, + "maintenance": { "symbol": "🚧", "description": "Maintenance" }, + "review": { "symbol": "👀", "description": "Reviewed Pull Requests" }, + "test": { "symbol": "⚠️", "description": "Tests" }, + "infra": { "symbol": "🔧", "description": "Infrastructure" } + }, + "skipCi": true, + "contributors": [] +} From f340099f7db51d3b720237c9b16c2e92b315d217 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:36:04 +0800 Subject: [PATCH 157/311] =?UTF-8?q?Add=20code-models.md=20doc=20(EN=20+=20?= =?UTF-8?q?ZH)=20=E2=80=94=20189=20code-focused=20models=20across=2041=20p?= =?UTF-8?q?roviders?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Free, cheapest, and largest context code model tables - Added to README, AGENTS.md, llms.txt, llms-full.txt --- AGENTS.md | 1 + README.md | 1 + docs/code-models.md | 90 ++++++++++++++++++++++++++++++++++++++++++ docs/zh/code-models.md | 90 ++++++++++++++++++++++++++++++++++++++++++ llms-full.txt | 1 + llms.txt | 1 + 6 files changed, 184 insertions(+) create mode 100644 docs/code-models.md create mode 100644 docs/zh/code-models.md diff --git a/AGENTS.md b/AGENTS.md index e9912393..e7965624 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -19,6 +19,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/video-models.md`](docs/video-models.md) — 167 video models ([中文](docs/zh/video-models.md)) - [`docs/image-generation.md`](docs/image-generation.md) — 28 image generation models ([中文](docs/zh/image-generation.md)) - [`docs/audio-models.md`](docs/audio-models.md) — 118 audio input + 34 audio output models ([中文](docs/zh/audio-models.md)) +- [`docs/code-models.md`](docs/code-models.md) — 189 code-focused models across 41 providers ([中文](docs/zh/code-models.md)) - [`docs/structured-output.md`](docs/structured-output.md) — 829 structured output models ([中文](docs/zh/structured-output.md)) - [`docs/modality-matrix.md`](docs/modality-matrix.md) — Model capabilities matrix ([中文](docs/zh/modality-matrix.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) diff --git a/README.md b/README.md index c4083443..e98105cb 100644 --- a/README.md +++ b/README.md @@ -422,6 +422,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Vision Models](docs/vision-models.md) | 1,487 vision models — cheapest, largest context, open-weight | | [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | | [Audio Models](docs/audio-models.md) | 118 audio input + 34 audio output models | +| [Code Models](docs/code-models.md) | 189 code-focused models across 41 providers | | [Video Models](docs/video-models.md) | 167 video input + 4 video output models | | [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | diff --git a/docs/code-models.md b/docs/code-models.md new file mode 100644 index 00000000..3ae25403 --- /dev/null +++ b/docs/code-models.md @@ -0,0 +1,90 @@ +# Code Models + +[中文](zh/code-models.md) + +AI models specifically designed or optimized for code generation, code understanding, and software development tasks. Includes models from 41 providers with code-focused capabilities. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Code Models Matter + +Code-focused models are optimized for: + +- **Code generation** — write functions, classes, and entire programs from descriptions +- **Code completion** — autocomplete suggestions in IDEs and editors +- **Code review** — identify bugs, security issues, and style violations +- **Code explanation** — understand and document existing code +- **Code translation** — convert code between programming languages +- **Test generation** — write unit tests and integration tests automatically + +## Stats + +| Metric | Count | +| ----------------------- | ----- | +| Code-focused models | 189 | +| Providers | 41 | +| Free code models | 0 | +| Open-weight code models | 14 | +| With reasoning | 43 | +| With tool calling | 119 | +| With structured output | 43 | + +## Providers + +`302ai`, `aihubmix`, `aimlapi`, `alibaba`, `amazon-bedrock`, `arcee`, `auriko`, `cerebras`, `clarifai`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `fastrouter`, `gmicloud`, `google-vertex`, `hyperbolic`, `inception`, `inferencenet` and 21 more + +## Free Code Models + +Code models available at zero cost — perfect for development and prototyping. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ----- | -------- | ------- | --------- | ---------- | ------------ | + +## Cheapest Code Models + +Best value code models for production use. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ------------------------------------------- | ------------ | ------- | --------- | ---------- | ------------ | +| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔧 🔓 | +| doubao-seed-code-preview-latest | 302ai | 131K | $0.0515 | $0.343 | 🔧 | +| qwen3-coder-30b-a3b-instruct | cortecs | 0 | $0.053 | $0.222 | 🧠 🔧 | +| qwen--qwen2.5-coder-32b-instruct | fastrouter | 32K | $0.06 | $0.15 | | +| qwen3-coder-flash | aihubmix | 0 | $0.068 | $0.272 | 🔧 📋 | +| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | 0 | $0.07 | $0.16 | 🔧 | +| Qwen3-Coder-30B-A3B-Instruct | ovhcloud | 262K | $0.07 | $0.26 | | +| qwen--qwen3-coder-30b-a3b-instruct | martian | 0 | $0.07 | $0.27 | | +| qwen--qwen3-coder-30b-a3b-instruct | novitaai | 160K | $0.07 | $0.27 | 🔧 📋 | +| qwen--qwen3-coder-30b-a3b-instruct | openrouter | 160K | $0.07 | $0.27 | 🔧 📋 | + +## Largest Context Code Models + +Code models with the largest context windows — for working with large codebases. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| -------------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| qwen--qwen3-coder | openrouter | 1M | $0.22 | $1.8 | 🔧 📋 | +| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | 🔧 | +| alibaba--qwen3-coder-flash | requesty | 1M | $0.3 | $1.5 | 🔧 | +| alibaba--qwen3-coder-plus | requesty | 1M | $1 | $5 | 🔧 | +| gpt-5-3-codex | meganova | 1M | $1.4 | $11.2 | 🔧 | +| qwen--qwen3-coder-flash | openrouter | 1M | $0.195 | $0.975 | 🔧 📋 | +| qwen--qwen3-coder-plus | openrouter | 1M | $0.65 | $3.25 | 🔧 📋 | +| gpt-5-codex | 302ai | 400K | $1.25 | $10 | 🔧 | +| gpt-5.1-codex | 302ai | 400K | $1.25 | $10 | 🔧 | + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Structured Output](structured-output.md) — 829 JSON-mode models +- [Free AI Models](free-models.md) — 81 free models by capability +- [Open-Weight Models](open-weights.md) — 527 models you can run yourself +- [Context Window Comparison](context-windows.md) — largest context windows +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/code-models.md b/docs/zh/code-models.md new file mode 100644 index 00000000..cf4df22b --- /dev/null +++ b/docs/zh/code-models.md @@ -0,0 +1,90 @@ +# 代码模型 + +[English](../code-models.md) + +专为代码生成、代码理解和软件开发任务设计或优化的 AI 模型。包含来自 41 个提供商的代码能力模型。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么代码模型很重要 + +代码模型针对以下场景优化： + +- **代码生成** — 根据描述编写函数、类和完整程序 +- **代码补全** — IDE 和编辑器中的自动补全建议 +- **代码审查** — 识别 bug、安全问题和风格违规 +- **代码解释** — 理解和文档化现有代码 +- **代码翻译** — 在编程语言之间转换代码 +- **测试生成** — 自动编写单元测试和集成测试 + +## 统计 + +| 指标 | 数量 | +| ---------------- | ---- | +| 代码模型 | 189 | +| 提供商 | 41 | +| 免费代码模型 | 0 | +| 开源权重代码模型 | 14 | +| 带推理 | 43 | +| 带工具调用 | 119 | +| 带结构化输出 | 43 | + +## 提供商 + +`302ai`, `aihubmix`, `aimlapi`, `alibaba`, `amazon-bedrock`, `arcee`, `auriko`, `cerebras`, `clarifai`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `fastrouter`, `gmicloud`, `google-vertex`, `hyperbolic`, `inception`, `inferencenet` 等 21 个 + +## 免费代码模型 + +零成本可用的代码模型 — 非常适合开发和原型验证。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---- | ------ | ------ | -------- | -------- | ---- | + +## 最便宜代码模型 + +生产环境使用的最佳性价比代码模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ------------------------------------------- | ------------ | ------ | -------- | -------- | ----- | +| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔧 🔓 | +| doubao-seed-code-preview-latest | 302ai | 131K | $0.0515 | $0.343 | 🔧 | +| qwen3-coder-30b-a3b-instruct | cortecs | 0 | $0.053 | $0.222 | 🧠 🔧 | +| qwen--qwen2.5-coder-32b-instruct | fastrouter | 32K | $0.06 | $0.15 | | +| qwen3-coder-flash | aihubmix | 0 | $0.068 | $0.272 | 🔧 📋 | +| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | 0 | $0.07 | $0.16 | 🔧 | +| Qwen3-Coder-30B-A3B-Instruct | ovhcloud | 262K | $0.07 | $0.26 | | +| qwen--qwen3-coder-30b-a3b-instruct | martian | 0 | $0.07 | $0.27 | | +| qwen--qwen3-coder-30b-a3b-instruct | novitaai | 160K | $0.07 | $0.27 | 🔧 📋 | +| qwen--qwen3-coder-30b-a3b-instruct | openrouter | 160K | $0.07 | $0.27 | 🔧 📋 | + +## 大上下文代码模型 + +上下文窗口最大的代码模型 — 适用于大型代码库。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| -------------------------- | ---------- | ------ | -------- | -------- | ----- | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| qwen--qwen3-coder | openrouter | 1M | $0.22 | $1.8 | 🔧 📋 | +| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | 🔧 | +| alibaba--qwen3-coder-flash | requesty | 1M | $0.3 | $1.5 | 🔧 | +| alibaba--qwen3-coder-plus | requesty | 1M | $1 | $5 | 🔧 | +| gpt-5-3-codex | meganova | 1M | $1.4 | $11.2 | 🔧 | +| qwen--qwen3-coder-flash | openrouter | 1M | $0.195 | $0.975 | 🔧 📋 | +| qwen--qwen3-coder-plus | openrouter | 1M | $0.65 | $3.25 | 🔧 📋 | +| gpt-5-codex | 302ai | 400K | $1.25 | $10 | 🔧 | +| gpt-5.1-codex | 302ai | 400K | $1.25 | $10 | 🔧 | + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 +- [上下文窗口对比](context-windows.md) — 最大上下文窗口 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/llms-full.txt b/llms-full.txt index 9d02af58..af68ee21 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -330,6 +330,7 @@ by_context = sorted( - [Vision Models](docs/vision-models.md) — 1,487 vision models - [Image Generation](docs/image-generation.md) — 28 image generation models - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models +- [Code Models](docs/code-models.md) — 189 code-focused models across 41 providers - [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video diff --git a/llms.txt b/llms.txt index 9d02af58..af68ee21 100644 --- a/llms.txt +++ b/llms.txt @@ -330,6 +330,7 @@ by_context = sorted( - [Vision Models](docs/vision-models.md) — 1,487 vision models - [Image Generation](docs/image-generation.md) — 28 image generation models - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models +- [Code Models](docs/code-models.md) — 189 code-focused models across 41 providers - [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video From 93c1e89f1db66785f83c8895e3f523ae005de8ec Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:36:51 +0800 Subject: [PATCH 158/311] Add code-models keyword to CITATION.cff --- CITATION.cff | 1 + 1 file changed, 1 insertion(+) diff --git a/CITATION.cff b/CITATION.cff index 1c094492..7c27bcc1 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -34,6 +34,7 @@ keywords: - cached-pricing - audio-models - video-models + - code-models - modality-matrix license: MIT version: 0.1.0 From a1971d2e8f79bc467cc2863b610274a9026908b5 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:37:13 +0800 Subject: [PATCH 159/311] Add code-models keyword to package.json (47 total) --- package.json | 1 + 1 file changed, 1 insertion(+) diff --git a/package.json b/package.json index b9537fd5..ea6bb2cc 100644 --- a/package.json +++ b/package.json @@ -15,6 +15,7 @@ "cerebras", "chatgpt", "claude", + "code-models", "context-window", "deepseek", "free-models", From 2b0c7cf22a506456a51e441de5d3020b8241c999 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:37:39 +0800 Subject: [PATCH 160/311] Add Code Models link to interactive catalog footer --- site/index.html | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/site/index.html b/site/index.html index c1394941..3a11fe8d 100644 --- a/site/index.html +++ b/site/index.html @@ -415,7 +415,10 @@

🤖 AI Models CatalogQuick Start · API · - FAQ + FAQ · + Code Models
From 5b27228a81d9534112775733d88a87a6bec998da Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:42:18 +0800 Subject: [PATCH 161/311] =?UTF-8?q?Redesign=20social-preview.svg=20?= =?UTF-8?q?=E2=80=94=20dark=20theme=20with=20gradient,=20stats=20cards,=20?= =?UTF-8?q?capability=20pills?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 5 stat cards: 4,587 models, 95 providers, 2,350 tool calling, 1,306 reasoning, 81 free - Capability pills: pricing, context windows, modalities, open weights, cached pricing - First-party data badge, npm install command - Dark GitHub-style theme with gradient accent --- social-preview.svg | 91 +++++++++++++++++++++++++++++++++------------- 1 file changed, 66 insertions(+), 25 deletions(-) diff --git a/social-preview.svg b/social-preview.svg index 4d3ea80a..923ef6cf 100644 --- a/social-preview.svg +++ b/social-preview.svg @@ -1,48 +1,89 @@ - - - + + + - - + + + + + + + - - + + + + + 🤖 - AI Models Catalog + AI Models Catalog - Structured YAML · First-Party Data · TypeScript Types + The most comprehensive structured catalog of AI models on GitHub + + + + + 4,587 + Models + + + + 95 + Providers + + + + 2,350 + Tool Calling + + + + 1,306 + Reasoning + + + + 81 + Free + + + + Pricing + + + Context Windows - - 4,587 - Models + + Modalities - 95 - Providers + + Open Weights - 1,306 - Reasoning + + Cached Pricing - - + + + ✓ First-Party Data Only · Zod Validation - - Pricing · Context Windows · Modalities · Tool Calling · Open Weights + + npm · CDN · CSV · GitHub Action · Hugging Face - - - npm install ai-models + + + $ npm install ai-models - - + + \ No newline at end of file From 2d8a9920b7cb68bcdb5b1cc6f23eeaea34cb6b17 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:43:57 +0800 Subject: [PATCH 162/311] Update social-preview.ts to match new SVG design with stats cards --- scripts/social-preview.ts | 89 +++++++++++++++++++++++++++------------ 1 file changed, 62 insertions(+), 27 deletions(-) diff --git a/scripts/social-preview.ts b/scripts/social-preview.ts index dccdfb02..562db303 100644 --- a/scripts/social-preview.ts +++ b/scripts/social-preview.ts @@ -4,8 +4,7 @@ * Outputs an SVG that can be converted to PNG via any browser or tool. * * Usage: - * node scripts/social-preview.js > social-preview.svg - * # Or open in browser and screenshot + * npx tsx scripts/social-preview.ts > social-preview.svg */ import fs from "node:fs"; @@ -28,50 +27,86 @@ for (const p of fs.readdirSync(providersDir)) { const svg = ` - - - + + + - - + + + + + + + - - + + + + + 🤖 - AI Models Catalog + AI Models Catalog - Structured YAML · First-Party Data · TypeScript Types + The most comprehensive structured catalog of AI models on GitHub + + + + ${modelCount.toLocaleString()} + Models + + + ${providerCount} + Providers + + + 2,350 + Tool Calling + + + 1,306 + Reasoning + + + 81 + Free + + + + Pricing + + + Context Windows - - ${modelCount.toLocaleString()} - Models + + Modalities - ${providerCount} - Providers + + Open Weights - 1,306 - Reasoning + + Cached Pricing - - + + + ✓ First-Party Data Only · Zod Validation - - Pricing · Context Windows · Modalities · Tool Calling · Open Weights + + npm · CDN · CSV · GitHub Action · Hugging Face - - - npm install ai-models + + + $ npm install ai-models - - + + `; process.stdout.write(svg); From d3a58e0107b921a327357d71309179f4d4dbda12 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:45:02 +0800 Subject: [PATCH 163/311] Add Contributors section to README with all-contributors placeholder --- README.md | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/README.md b/README.md index e98105cb..8682859f 100644 --- a/README.md +++ b/README.md @@ -536,6 +536,15 @@ Built something with this catalog? [Open a PR](https://github.com/i-need-token/a | **Data dashboards** | Import CSV into Excel/Google Sheets for visual pricing analysis | | **Chatbot builders** | Pick the right model by context window, modality, and budget | +## Contributors + +Thanks to everyone who has contributed to this catalog! + + + + +Want to contribute? Check out [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines. + ## License [MIT](LICENSE) From 40dc431124a36494617c7681c1b92e241e2f390c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:45:39 +0800 Subject: [PATCH 164/311] Add Changelog and Security Policy links to README Contributors section --- README.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/README.md b/README.md index 8682859f..31f4c9f9 100644 --- a/README.md +++ b/README.md @@ -545,6 +545,9 @@ Thanks to everyone who has contributed to this catalog! Want to contribute? Check out [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines. +- 📋 [Changelog](CHANGELOG.md) — what's new in each release +- 🔒 [Security Policy](SECURITY.md) — reporting vulnerabilities + ## License [MIT](LICENSE) From cd3e91e4e81ec1c27639ab2b6621e11ea98294c7 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:49:17 +0800 Subject: [PATCH 165/311] Add Agentic Models to README Use Cases table (EN + ZH) --- docs/agentic-models.md | 91 +++++++++++++++++++++++++++++++++++++++ docs/zh/agentic-models.md | 91 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 182 insertions(+) create mode 100644 docs/agentic-models.md create mode 100644 docs/zh/agentic-models.md diff --git a/docs/agentic-models.md b/docs/agentic-models.md new file mode 100644 index 00000000..0741fd7a --- /dev/null +++ b/docs/agentic-models.md @@ -0,0 +1,91 @@ +# Agentic Models + +[中文](zh/agentic-models.md) + +AI models with both **tool calling** and **reasoning** capabilities — the key requirements for building AI agents. These models can plan, reason through multi-step tasks, and use external tools to accomplish goals. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Agentic Models Matter + +Agentic models combine two critical capabilities: + +- **Tool calling** — invoke functions, APIs, and external tools to take actions +- **Reasoning** — chain-of-thought thinking to plan and decompose complex tasks + +Together, these enable: + +- **Autonomous agents** — models that can plan, act, and iterate independently +- **Multi-step workflows** — break complex tasks into subtasks with tool use +- **Self-correction** — reason about failures and retry with different approaches +- **Code agents** — write, execute, and debug code autonomously +- **Research agents** — search, synthesize, and summarize information + +## Stats + +| Metric | Count | +| -------------------------------------- | ----- | +| Agentic models (tool call + reasoning) | 1080 | +| Providers | 51 | +| Free agentic models | 0 | +| Open-weight agentic models | 65 | +| With structured output | 455 | + +## Providers + +`302ai`, `aihubmix`, `alibaba`, `amazon`, `anthropic`, `arcee`, `auriko`, `baidu`, `baseten`, `bytedance`, `chutes`, `clarifai`, `cloudflare`, `cortecs`, `deepseek`, `digitalocean`, `dinference`, `fastrouter`, `fireworks`, `google`, `hpc-ai`, `inclusionai`, `inferencenet`, `klusterai`, `llmgateway` and 26 more + +## Free Agentic Models + +Free models with both tool calling and reasoning — zero-cost agents. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ----- | -------- | ------- | --------- | ---------- | ------------ | + +## Cheapest Agentic Models + +Best value agentic models for production agents. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ----------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | | +| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | | +| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | | +| openai--gpt-5-nano:flex | requesty | 400K | $0.025 | $0.2 | | +| gpt-5-nano | aihubmix | 0 | $0.025 | $0.2 | 📋 | +| gpt-oss-20b | cortecs | 0 | $0.027 | $0.124 | | +| qwen--qwen3-4b-fp8 | novitaai | 128K | $0.03 | $0.03 | | +| openai--gpt-oss-20b | openrouter | 131K | $0.03 | $0.14 | 📋 | +| qwen-3.5-4b | auriko | 262K | $0.03 | $0.15 | | +| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | 📋 | + +## Largest Context Agentic Models + +Agentic models with the largest context windows — for complex multi-step tasks. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ----------------------------- | -------- | ------- | --------- | ---------- | ------------ | +| grok-4-20 | venice | 2M | $1.42 | $2.83 | 📋 | +| grok-4.20-beta-0309-reasoning | 302ai | 2M | $2 | $6 | | +| grok-4-1-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | +| grok-4-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | +| openai--gpt-5.5 | requesty | 1M | $5 | $30 | | +| openai-responses--gpt-5.4 | requesty | 1M | $2.5 | $15 | | +| openai-responses--gpt-5.5 | requesty | 1M | $5 | $30 | | +| openai--gpt-5.4 | requesty | 1M | $2.5 | $15 | | +| openai-responses--gpt-5.4-pro | requesty | 1M | $30 | $180 | | +| openai-responses--gpt-5.5-pro | requesty | 1M | $30 | $180 | | + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Code Models](code-models.md) — 189 code-focused models +- [Structured Output](structured-output.md) — 829 JSON-mode models +- [Free AI Models](free-models.md) — 81 free models by capability +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/agentic-models.md b/docs/zh/agentic-models.md new file mode 100644 index 00000000..714ff25d --- /dev/null +++ b/docs/zh/agentic-models.md @@ -0,0 +1,91 @@ +# 智能体模型 + +[English](../agentic-models.md) + +同时具备**工具调用**和**推理**能力的 AI 模型 — 构建 AI Agent 的关键要求。这些模型可以规划、推理多步骤任务，并使用外部工具完成目标。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么智能体模型很重要 + +智能体模型结合了两个关键能力： + +- **工具调用** — 调用函数、API 和外部工具来执行操作 +- **推理** — 链式思维思考，规划和分解复杂任务 + +两者结合可以实现： + +- **自主智能体** — 独立规划、行动和迭代的模型 +- **多步骤工作流** — 将复杂任务分解为子任务并使用工具 +- **自我纠正** — 推理失败原因并使用不同方法重试 +- **代码智能体** — 自主编写、执行和调试代码 +- **研究智能体** — 搜索、综合和总结信息 + +## 统计 + +| 指标 | 数量 | +| ----------------------------- | ---- | +| 智能体模型（工具调用 + 推理） | 1080 | +| 提供商 | 51 | +| 免费智能体模型 | 0 | +| 开源权重智能体模型 | 65 | +| 带结构化输出 | 455 | + +## 提供商 + +`302ai`, `aihubmix`, `alibaba`, `amazon`, `anthropic`, `arcee`, `auriko`, `baidu`, `baseten`, `bytedance`, `chutes`, `clarifai`, `cloudflare`, `cortecs`, `deepseek`, `digitalocean`, `dinference`, `fastrouter`, `fireworks`, `google`, `hpc-ai`, `inclusionai`, `inferencenet`, `klusterai`, `llmgateway` 等 26 个 + +## 免费智能体模型 + +同时具备工具调用和推理的免费模型 — 零成本 Agent。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---- | ------ | ------ | -------- | -------- | ---- | + +## 最便宜智能体模型 + +生产环境 Agent 的最佳性价比模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ----------------------- | ---------- | ------ | -------- | -------- | ---- | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | | +| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | | +| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | | +| openai--gpt-5-nano:flex | requesty | 400K | $0.025 | $0.2 | | +| gpt-5-nano | aihubmix | 0 | $0.025 | $0.2 | 📋 | +| gpt-oss-20b | cortecs | 0 | $0.027 | $0.124 | | +| qwen--qwen3-4b-fp8 | novitaai | 128K | $0.03 | $0.03 | | +| openai--gpt-oss-20b | openrouter | 131K | $0.03 | $0.14 | 📋 | +| qwen-3.5-4b | auriko | 262K | $0.03 | $0.15 | | +| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | 📋 | + +## 大上下文智能体模型 + +上下文窗口最大的智能体模型 — 适用于复杂多步骤任务。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ----------------------------- | -------- | ------ | -------- | -------- | ---- | +| grok-4-20 | venice | 2M | $1.42 | $2.83 | 📋 | +| grok-4.20-beta-0309-reasoning | 302ai | 2M | $2 | $6 | | +| grok-4-1-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | +| grok-4-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | +| openai--gpt-5.5 | requesty | 1M | $5 | $30 | | +| openai-responses--gpt-5.4 | requesty | 1M | $2.5 | $15 | | +| openai-responses--gpt-5.5 | requesty | 1M | $5 | $30 | | +| openai--gpt-5.4 | requesty | 1M | $2.5 | $15 | | +| openai-responses--gpt-5.4-pro | requesty | 1M | $30 | $180 | | +| openai-responses--gpt-5.5-pro | requesty | 1M | $30 | $180 | | + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [代码模型](code-models.md) — 189 个代码模型 +- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 From a3cae100dfdcad1a4d271091df3ad692372c7ca0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:51:07 +0800 Subject: [PATCH 166/311] Add agentic-models.md to all index files, sitemap, CITATION.cff, package.json - AGENTS.md, llms.txt, llms-full.txt - sitemap.xml (regenerated with all 26 docs) - site/index.html footer - CITATION.cff keywords - package.json keywords (agentic-models, ai-agents) --- AGENTS.md | 1 + CITATION.cff | 1 + llms-full.txt | 1 + llms.txt | 1 + package.json | 2 + site/sitemap.xml | 105 +++++++++++++++++++++++++++-------------------- 6 files changed, 66 insertions(+), 45 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index e7965624..b2885500 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -20,6 +20,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/image-generation.md`](docs/image-generation.md) — 28 image generation models ([中文](docs/zh/image-generation.md)) - [`docs/audio-models.md`](docs/audio-models.md) — 118 audio input + 34 audio output models ([中文](docs/zh/audio-models.md)) - [`docs/code-models.md`](docs/code-models.md) — 189 code-focused models across 41 providers ([中文](docs/zh/code-models.md)) +- [`docs/agentic-models.md`](docs/agentic-models.md) — Models with tool calling + reasoning for AI agents ([中文](docs/zh/agentic-models.md)) - [`docs/structured-output.md`](docs/structured-output.md) — 829 structured output models ([中文](docs/zh/structured-output.md)) - [`docs/modality-matrix.md`](docs/modality-matrix.md) — Model capabilities matrix ([中文](docs/zh/modality-matrix.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) diff --git a/CITATION.cff b/CITATION.cff index 7c27bcc1..1ccac0ae 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -35,6 +35,7 @@ keywords: - audio-models - video-models - code-models + - agentic-models - modality-matrix license: MIT version: 0.1.0 diff --git a/llms-full.txt b/llms-full.txt index af68ee21..6b807eb3 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -331,6 +331,7 @@ by_context = sorted( - [Image Generation](docs/image-generation.md) — 28 image generation models - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models - [Code Models](docs/code-models.md) — 189 code-focused models across 41 providers +- [Agentic Models](docs/agentic-models.md) — models with tool calling + reasoning for AI agents - [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video diff --git a/llms.txt b/llms.txt index af68ee21..6b807eb3 100644 --- a/llms.txt +++ b/llms.txt @@ -331,6 +331,7 @@ by_context = sorted( - [Image Generation](docs/image-generation.md) — 28 image generation models - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models - [Code Models](docs/code-models.md) — 189 code-focused models across 41 providers +- [Agentic Models](docs/agentic-models.md) — models with tool calling + reasoning for AI agents - [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video diff --git a/package.json b/package.json index ea6bb2cc..cffe2562 100644 --- a/package.json +++ b/package.json @@ -3,6 +3,8 @@ "version": "0.1.0", "description": "Structured YAML catalog of 4,587 AI models across 95 providers — pricing, context windows, modalities, capabilities. First-party data with TypeScript types and Zod validation.", "keywords": [ + "agentic-models", + "ai-agents", "ai-catalog", "ai-directory", "ai-model-pricing", diff --git a/site/sitemap.xml b/site/sitemap.xml index 56619a4e..5e510d5b 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -2,122 +2,137 @@ https://i-need-token.github.io/ai-models/ - weekly + daily 1.0 - https://github.com/i-need-token/ai-models/blob/main/docs/quick-start.md - monthly - 0.9 + https://i-need-token.github.io/ai-models/docs/agentic-models + weekly + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/api.md - monthly - 0.9 + https://i-need-token.github.io/ai-models/docs/api + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/audio-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/cached-pricing + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/code-examples + weekly + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/faq.md - monthly - 0.8 + https://i-need-token.github.io/ai-models/docs/code-models + weekly + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/glossary.md - monthly - 0.8 + https://i-need-token.github.io/ai-models/docs/context-windows + weekly + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/data-schema.md - monthly + https://i-need-token.github.io/ai-models/docs/data-acquisition + weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/providers.md + https://i-need-token.github.io/ai-models/docs/data-schema weekly - 0.8 + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/model-comparison.md + https://i-need-token.github.io/ai-models/docs/faq weekly - 0.8 + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/pricing-comparison.md + https://i-need-token.github.io/ai-models/docs/free-models weekly - 0.8 + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/context-windows.md + https://i-need-token.github.io/ai-models/docs/glossary weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/modality-matrix.md + https://i-need-token.github.io/ai-models/docs/image-generation weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/reasoning-models.md + https://i-need-token.github.io/ai-models/docs/lessons-learned weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/tool-calling.md + https://i-need-token.github.io/ai-models/docs/modality-matrix weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/structured-output.md + https://i-need-token.github.io/ai-models/docs/model-comparison weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/vision-models.md + https://i-need-token.github.io/ai-models/docs/model-selection weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/video-models.md + https://i-need-token.github.io/ai-models/docs/open-weights weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/audio-models.md + https://i-need-token.github.io/ai-models/docs/pricing-comparison weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/image-generation.md + https://i-need-token.github.io/ai-models/docs/providers weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/free-models.md + https://i-need-token.github.io/ai-models/docs/quick-start weekly - 0.8 + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/open-weights.md + https://i-need-token.github.io/ai-models/docs/reasoning-models weekly - 0.8 + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/code-examples.md - monthly + https://i-need-token.github.io/ai-models/docs/structured-output + weekly 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/data-acquisition.md - monthly - 0.6 + https://i-need-token.github.io/ai-models/docs/tool-calling + weekly + 0.7 - https://github.com/i-need-token/ai-models/blob/main/docs/lessons-learned.md - monthly - 0.6 + https://i-need-token.github.io/ai-models/docs/video-models + weekly + 0.7 - https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md - monthly - 0.5 + https://i-need-token.github.io/ai-models/docs/vision-models + weekly + 0.7 From 90f85af668cd45aba7f20813acca5e1bb01cd064 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:56:20 +0800 Subject: [PATCH 167/311] Add agentic-models.md to README documentation table --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 31f4c9f9..730224ba 100644 --- a/README.md +++ b/README.md @@ -423,6 +423,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Image Generation](docs/image-generation.md) | 28 image generation models — DALL·E, Imagen, GPT-5 Image | | [Audio Models](docs/audio-models.md) | 118 audio input + 34 audio output models | | [Code Models](docs/code-models.md) | 189 code-focused models across 41 providers | +| [Agentic Models](docs/agentic-models.md) | 1,080 models with tool calling + reasoning for AI agents | | [Video Models](docs/video-models.md) | 167 video input + 4 video output models | | [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | From 5bed0701fdfca7f657938842116208dce6d9bf82 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:58:44 +0800 Subject: [PATCH 168/311] Expand interactive catalog footer with 14 key doc links Added: Reasoning, Agentic Models, Cached Pricing, Context Windows, Vision --- site/index.html | 35 +++++++++++++++++++++++++++++------ 1 file changed, 29 insertions(+), 6 deletions(-) diff --git a/site/index.html b/site/index.html index 3a11fe8d..56a099c3 100644 --- a/site/index.html +++ b/site/index.html @@ -395,6 +395,10 @@

🤖 AI Models Catalognpm · Hugging Face
+ Quick Start + · Model Selection @@ -407,18 +411,37 @@

🤖 AI Models CatalogTool Calling · - PricingReasoning · - Quick StartAgentic Models - · API · - FAQ · + · Code Models + · + Pricing + · + Cached Pricing + · + Context Windows + · + Vision + · + API + · + FAQ
From 8d7b81c8560e54af9355eb888f43d05016a3f343 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 16:59:41 +0800 Subject: [PATCH 169/311] Add Sponsors section to README with GitHub Sponsors link --- README.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/README.md b/README.md index 730224ba..45b0740a 100644 --- a/README.md +++ b/README.md @@ -549,6 +549,10 @@ Want to contribute? Check out [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines. - 📋 [Changelog](CHANGELOG.md) — what's new in each release - 🔒 [Security Policy](SECURITY.md) — reporting vulnerabilities +## Sponsors + +Support this project by [sponsoring us on GitHub](https://github.com/sponsors/i-need-token). Your sponsorship helps maintain and expand the catalog. + ## License [MIT](LICENSE) From 71718b958dee02a7713219e25965e3757539d861 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 17:00:15 +0800 Subject: [PATCH 170/311] Add Star History chart image to README --- README.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/README.md b/README.md index 45b0740a..e2da296f 100644 --- a/README.md +++ b/README.md @@ -549,6 +549,10 @@ Want to contribute? Check out [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines. - 📋 [Changelog](CHANGELOG.md) — what's new in each release - 🔒 [Security Policy](SECURITY.md) — reporting vulnerabilities +## Star History + +[![Star History Chart](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) + ## Sponsors Support this project by [sponsoring us on GitHub](https://github.com/sponsors/i-need-token). Your sponsorship helps maintain and expand the catalog. From c415d53d3e1d9a4806e75235500ac026f9a1bdb3 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 17:01:04 +0800 Subject: [PATCH 171/311] =?UTF-8?q?Add=20Ecosystem=20&=20Integrations=20se?= =?UTF-8?q?ction=20to=20README=20=E2=80=94=20npm,=20CDN,=20Action,=20HF,?= =?UTF-8?q?=20CSV,=20catalog?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/README.md b/README.md index e2da296f..9eb9089c 100644 --- a/README.md +++ b/README.md @@ -514,6 +514,17 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. - **Bilingual docs** — 24 English + 24 Chinese documentation pages - **Open data** — all model data is open and programmatically accessible +## Ecosystem & Integrations + +| Integration | Description | Link | +| ----------------------- | --------------------------- | -------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, sort, filter models | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | + ## Roadmap - 🔜 **More providers** — continuously adding new AI model providers From aa6a0be51ed9c46b7c86127d50f0cc910817eb3c Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 17:32:17 +0800 Subject: [PATCH 172/311] Fix stats workflow: add oxfmt formatting step before commit --- .github/workflows/stats.yml | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/.github/workflows/stats.yml b/.github/workflows/stats.yml index a715ddc0..4c0c166e 100644 --- a/.github/workflows/stats.yml +++ b/.github/workflows/stats.yml @@ -5,6 +5,9 @@ on: schedule: - cron: "0 0 * * 0" # weekly on Sunday +permissions: + contents: write + jobs: stats: runs-on: ubuntu-latest @@ -36,6 +39,9 @@ jobs: {"schemaVersion":1,"label":"providers","message":"$PROVIDERS","color":"blue"} EOF + - name: Format badges + run: npx oxfmt --write .github/badges/*.json + - name: Commit badges run: | git config user.name "github-actions[bot]" From f53b9adde7935f05006e8f9b6faa0649ef80c9bb Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 17:41:16 +0800 Subject: [PATCH 173/311] =?UTF-8?q?Add=20Model=20Picks=20section=20?= =?UTF-8?q?=E2=80=94=20curated=20recommendations=20for=2010=20common=20use?= =?UTF-8?q?=20cases?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/README.md b/README.md index 9eb9089c..6e0af22d 100644 --- a/README.md +++ b/README.md @@ -90,6 +90,23 @@ Machine-readable YAML catalog of every major AI model provider and their models +## 🏆 Model Picks + +> Curated recommendations for common use cases — from [4,587 models](docs/model-comparison.md) across 95 providers + +| Use Case | Model | Why | Input $/M | Context | +| --------------------- | ---------------- | --------------------------------------- | ------------ | ------- | +| **Coding** | gpt-4.1 | Best code generation + 1M context | | 1M | +| **Coding (cheap)** | gpt-4.1-nano | 20× cheaper, great for autocomplete | /bin/bash.10 | 1M | +| **Reasoning** | o4-mini | Best cost-effective reasoning | .10 | 200K | +| **Reasoning (power)** | claude-opus-4 | Deepest reasoning for hard problems | 5 | 200K | +| **Agents** | claude-sonnet-4 | Best tool use + reasoning balance | | 200K | +| **Agents (cheap)** | gemini-2.5-flash | Fastest agent loop under | /bin/bash.15 | 1M | +| **Vision** | gemini-2.5-pro | Best multimodal understanding | .25 | 1M | +| **Free** | llama-4-scout | 10M context, open weights, free on Groq | /bin/bash | 10M | +| **Open weights** | deepseek-r1 | Best open reasoning model | /bin/bash.55 | 128K | +| **Large context** | gemini-2.5-flash | 1M context at lowest price | /bin/bash.15 | 1M | + ## Use Cases | Use Case | How This Catalog Helps | From 4458e1bc48be74b53d8cecb24cbe136cf4c613e7 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 17:41:44 +0800 Subject: [PATCH 174/311] Add Model Picks to README table of contents --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 6e0af22d..667a5bcb 100644 --- a/README.md +++ b/README.md @@ -42,6 +42,7 @@ Machine-readable YAML catalog of every major AI model provider and their models ## Contents - [Quick Compare](#quick-compare) +- [🏆 Model Picks](#-model-picks) - [Use Cases](#use-cases) - [Quick Numbers](#quick-numbers) - [Example Model](#example-model) From 75357c7edaca7906a21a4b7dfe34ae7b41cf0a23 Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" Date: Thu, 21 May 2026 09:42:36 +0000 Subject: [PATCH 175/311] chore: update dynamic badges [skip ci] --- .github/badges/models.json | 7 +------ .github/badges/providers.json | 7 +------ 2 files changed, 2 insertions(+), 12 deletions(-) diff --git a/.github/badges/models.json b/.github/badges/models.json index 287af461..4f27c22a 100644 --- a/.github/badges/models.json +++ b/.github/badges/models.json @@ -1,6 +1 @@ -{ - "schemaVersion": 1, - "label": "models", - "message": "4,587", - "color": "green" -} +{ "schemaVersion": 1, "label": "models", "message": "4587", "color": "green" } diff --git a/.github/badges/providers.json b/.github/badges/providers.json index db0a9530..89c832bf 100644 --- a/.github/badges/providers.json +++ b/.github/badges/providers.json @@ -1,6 +1 @@ -{ - "schemaVersion": 1, - "label": "providers", - "message": "95", - "color": "blue" -} +{ "schemaVersion": 1, "label": "providers", "message": "87", "color": "blue" } From d0beb809f780e4a8c748dfc640c24cdc460f4891 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:00:48 +0800 Subject: [PATCH 176/311] =?UTF-8?q?Add=20concrete=20roadmap=20items=20to?= =?UTF-8?q?=20README=20=E2=80=94=20embedding=20models,=20provider=20compar?= =?UTF-8?q?ison,=20benchmarks?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/README.md b/README.md index 667a5bcb..f37eaf67 100644 --- a/README.md +++ b/README.md @@ -545,6 +545,22 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Roadmap +- [ ] Embedding models documentation page +- [ ] Provider comparison (side-by-side) +- [ ] Model benchmarking data integration +- [ ] Streaming support metadata +- [ ] Fine-tuning availability tracking +- [ ] Regional availability data +- [ ] Community-contributed model reviews + +- [ ] Add embedding models documentation +- [ ] Add provider comparison page +- [ ] Add model benchmarking data +- [ ] Add streaming support metadata +- [ ] Add fine-tuning availability tracking +- [ ] Add regional availability data +- [ ] Community-contributed model reviews + - 🔜 **More providers** — continuously adding new AI model providers - 🔜 **REST API** — hosted API for querying the catalog - 🔜 **Historical pricing** — track pricing changes over time From 2e34fae09bc239cdd23a3f3b2b2ed39fca11efc1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:12:09 +0800 Subject: [PATCH 177/311] Add 3 new doc pages: chat-models, multimodal-models, embedding-models (EN + ZH) - chat-models.md: 2,350 models with tool calling across 67 providers - multimodal-models.md: 1,519 models with image/audio/video input - embedding-models.md: 5 embedding models for search/RAG/similarity - All 3 added to: README Use Cases + Documentation tables, AGENTS.md, llms.txt, llms-full.txt, sitemap.xml, site/index.html footer - CITATION.cff + package.json keywords updated (52 total) - Bilingual (EN + ZH) with cross-links --- .gitignore | 1 + AGENTS.md | 3 ++ CITATION.cff | 3 ++ README.md | 6 +++ docs/chat-models.md | 99 ++++++++++++++++++++++++++++++++++ docs/embedding-models.md | 71 ++++++++++++++++++++++++ docs/multimodal-models.md | 101 +++++++++++++++++++++++++++++++++++ docs/zh/chat-models.md | 99 ++++++++++++++++++++++++++++++++++ docs/zh/embedding-models.md | 71 ++++++++++++++++++++++++ docs/zh/multimodal-models.md | 101 +++++++++++++++++++++++++++++++++++ llms-full.txt | 3 ++ llms.txt | 3 ++ package.json | 3 ++ site/index.html | 12 +++++ site/sitemap.xml | 15 ++++++ social-preview.svg | 5 -- 16 files changed, 591 insertions(+), 5 deletions(-) create mode 100644 docs/chat-models.md create mode 100644 docs/embedding-models.md create mode 100644 docs/multimodal-models.md create mode 100644 docs/zh/chat-models.md create mode 100644 docs/zh/embedding-models.md create mode 100644 docs/zh/multimodal-models.md diff --git a/.gitignore b/.gitignore index def2133e..cde8bdb4 100644 --- a/.gitignore +++ b/.gitignore @@ -7,3 +7,4 @@ dist/ dist/ models.csv stats.json +social-preview.png diff --git a/AGENTS.md b/AGENTS.md index b2885500..263d4e6e 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -21,6 +21,9 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/audio-models.md`](docs/audio-models.md) — 118 audio input + 34 audio output models ([中文](docs/zh/audio-models.md)) - [`docs/code-models.md`](docs/code-models.md) — 189 code-focused models across 41 providers ([中文](docs/zh/code-models.md)) - [`docs/agentic-models.md`](docs/agentic-models.md) — Models with tool calling + reasoning for AI agents ([中文](docs/zh/agentic-models.md)) +- [`docs/chat-models.md`](docs/chat-models.md) — 2,350 models with tool calling for chat applications ([中文](docs/zh/chat-models.md)) +- [`docs/multimodal-models.md`](docs/multimodal-models.md) — 1,519 models with image/audio/video input ([中文](docs/zh/multimodal-models.md)) +- [`docs/embedding-models.md`](docs/embedding-models.md) — 5 embedding models for search, RAG, similarity ([中文](docs/zh/embedding-models.md)) - [`docs/structured-output.md`](docs/structured-output.md) — 829 structured output models ([中文](docs/zh/structured-output.md)) - [`docs/modality-matrix.md`](docs/modality-matrix.md) — Model capabilities matrix ([中文](docs/zh/modality-matrix.md)) - [`docs/providers.md`](docs/providers.md) — Provider overview by type and market ([中文](docs/zh/providers.md)) diff --git a/CITATION.cff b/CITATION.cff index 1ccac0ae..bfc2ceae 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -37,6 +37,9 @@ keywords: - code-models - agentic-models - modality-matrix + - chat-models + - multimodal-models + - embedding-models license: MIT version: 0.1.0 date-released: "2026-05-21" diff --git a/README.md b/README.md index f37eaf67..001f6b53 100644 --- a/README.md +++ b/README.md @@ -122,6 +122,9 @@ Machine-readable YAML catalog of every major AI model provider and their models | 🎯 **Choose the right model** | [Model selection guide](docs/model-selection.md) — decision framework | | 💸 **Optimize API costs** | [Cached pricing](docs/cached-pricing.md) — 1,374 models with 50-90% savings | | 🧪 **Prototype for free** | [Free models](docs/free-models.md) — 81 models at zero cost | +| 💬 **Build chat apps** | [Chat models](docs/chat-models.md) — 2,350 models with tool calling | +| 🖼️ **Process images/audio** | [Multimodal models](docs/multimodal-models.md) — 1,519 models with vision/audio/video | +| 🔎 **Power semantic search** | [Embedding models](docs/embedding-models.md) — vector search & RAG | ## Quick Numbers @@ -442,6 +445,9 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Audio Models](docs/audio-models.md) | 118 audio input + 34 audio output models | | [Code Models](docs/code-models.md) | 189 code-focused models across 41 providers | | [Agentic Models](docs/agentic-models.md) | 1,080 models with tool calling + reasoning for AI agents | +| [Chat Models](docs/chat-models.md) | 2,350 models with tool calling for chat applications | +| [Multimodal Models](docs/multimodal-models.md) | 1,519 models with image/audio/video input | +| [Embedding Models](docs/embedding-models.md) | 5 embedding models for search, RAG, similarity | | [Video Models](docs/video-models.md) | 167 video input + 4 video output models | | [Structured Output](docs/structured-output.md) | 829 JSON-mode models — cheapest, free, with tool calling | | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | diff --git a/docs/chat-models.md b/docs/chat-models.md new file mode 100644 index 00000000..e4bdd4cd --- /dev/null +++ b/docs/chat-models.md @@ -0,0 +1,99 @@ +# Chat Models + +[中文](zh/chat-models.md) + +AI models with **tool calling** support — the essential capability for building chat-based applications, AI assistants, and conversational agents. These models can understand natural language, generate responses, and invoke external tools. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Chat Models Matter + +Chat models are the backbone of modern AI applications: + +- **Conversational AI** — natural language dialogue with context +- **AI Assistants** — task-oriented chat with tool use +- **Customer Support** — automated support with knowledge base access +- **Content Generation** — writing, summarization, translation +- **Data Analysis** — natural language queries over structured data + +Tool calling is the key differentiator — it allows models to go beyond text generation and take actions in the real world. + +## Stats + +| Metric | Count | +| ------------------------------- | ----- | +| Chat models (with tool calling) | 2350 | +| Providers | 71 | +| Free chat models | 54 | +| Open-weight chat models | 278 | +| With reasoning | 1080 | +| With structured output | 758 | + +## Providers + +`01ai`, `302ai`, `aihubmix`, `aimlapi`, `alibaba`, `amazon`, `amazon-bedrock`, `anthropic`, `arcee`, `auriko`, `baidu`, `baseten`, `berget`, `bytedance`, `cerebras`, `chutes`, `clarifai`, `cloudferro-sherlock`, `cloudflare`, `cortecs`, `databricks`, `deepseek`, `digitalocean`, `dinference`, `evroc` and 46 more + +## Free Chat Models + +Free models with tool calling — zero-cost chat applications. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ---------------------------------------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| openrouter--owl-alpha | openrouter | 1M | Free | Free | 📋 | +| deepseek--deepseek-v4-flash--free | openrouter | 1M | Free | Free | 🧠 | +| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | | +| nvidia--nemotron-3-super-120b-a12b--free | openrouter | 1M | Free | Free | 🧠 📋 | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🧠 📋 | +| arcee-ai--trinity-large-thinking--free | openrouter | 262K | Free | Free | 🧠 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🧠 📋 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🧠 📋 | +| gemma-4-31b-it | auriko | 262K | Free | Free | 🧠 📋 | +| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | 256K | Free | Free | 🧠 | + +## Cheapest Chat Models + +Best value chat models for production. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ------------------------------------------- | ------------ | ------- | --------- | ---------- | ------------ | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 📋 | +| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔓 | +| inclusionai--ling-2.6-flash | openrouter | 262K | $0.01 | $0.03 | 📋 | +| ling-2.6-flash | inclusionai | 262K | $0.01 | $0.03 | | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 | +| qwen3-vl-flash-2026-01-22 | aihubmix | 0 | $0.0103 | $0.103 | 📋 | +| qwen3-vl-flash | aihubmix | 0 | $0.0103 | $0.103 | 📋 | +| klusterai--Meta-Llama-3.1-8B-Instruct-Turbo | klusterai | 131K | $0.015 | $0.02 | | +| granite-4.0-h-micro | cloudflare | 131K | $0.017 | $0.112 | 🔓 | +| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | 🧠 | + +## Largest Context Chat Models + +Chat models with the largest context windows — for long conversations and document analysis. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ---------------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 📋 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 📋 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | | + +## Related Documentation + +- [Agentic Models](agentic-models.md) — 1,080 models with tool calling + reasoning +- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [Code Models](code-models.md) — 189 code-focused models +- [Free AI Models](free-models.md) — 81 free models by capability +- [Structured Output](structured-output.md) — 829 JSON-mode models +- [Model Selection Guide](model-selection.md) — decision framework +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/embedding-models.md b/docs/embedding-models.md new file mode 100644 index 00000000..bce44d64 --- /dev/null +++ b/docs/embedding-models.md @@ -0,0 +1,71 @@ +# Embedding Models + +[中文](zh/embedding-models.md) + +AI models that generate **vector embeddings** — numerical representations of text, images, and other data. Essential for semantic search, RAG (Retrieval-Augmented Generation), clustering, and similarity tasks. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Embedding Models Matter + +Embedding models are the foundation of many AI systems: + +- **Semantic Search** — find relevant documents by meaning, not keywords +- **RAG** — retrieve context for LLMs to generate grounded answers +- **Clustering** — group similar items together +- **Similarity** — find duplicates, recommendations, and related content +- **Classification** — zero-shot and few-shot classification via embeddings + +## Stats + +| Metric | Count | +| ---------------------------- | ----- | +| Embedding models | 5 | +| Providers | 3 | +| Free embedding models | 0 | +| Open-weight embedding models | 1 | + +## Providers + +`openai`, `tencent`, `upstage` + +## Free Embedding Models + +Free embedding models — zero-cost semantic search and RAG. + +| Model | Provider | Context | Input $/M | +| ----- | -------- | ------- | --------- | + +## Cheapest Embedding Models + +Best value embedding models for production. + +| Model | Provider | Context | Input $/M | +| ----------------------- | -------- | ------- | --------- | --- | +| text-embedding-3-small | openai | 8K | $0.02 | | +| solar-embedding-1-large | upstage | 0 | $0.1 | 🔓 | +| text-embedding-ada-002 | openai | 8K | $0.1 | | +| text-embedding-3-large | openai | 8K | $0.13 | | +| hunyuan-embedding | tencent | 0 | $0.7 | | + +## Largest Context Embedding Models + +Embedding models with the largest context windows — for embedding long documents. + +| Model | Provider | Context | Input $/M | +| ---------------------- | -------- | ------- | --------- | --- | +| text-embedding-ada-002 | openai | 8K | $0.1 | | +| text-embedding-3-small | openai | 8K | $0.02 | | +| text-embedding-3-large | openai | 8K | $0.13 | | + +## Related Documentation + +- [Free AI Models](free-models.md) — 81 free models by capability +- [Open Weights](open-weights.md) — 527 open-weight models +- [Model Selection Guide](model-selection.md) — decision framework +- [API Reference](api.md) — programmatic access +- [Quick Start](quick-start.md) — get started in 5 minutes + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/multimodal-models.md b/docs/multimodal-models.md new file mode 100644 index 00000000..1745d8f0 --- /dev/null +++ b/docs/multimodal-models.md @@ -0,0 +1,101 @@ +# Multimodal Models + +[中文](zh/multimodal-models.md) + +AI models that can process **multiple input modalities** — images, audio, and video alongside text. These models power visual Q&A, document analysis, video understanding, and audio transcription. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Multimodal Models Matter + +Multimodal models break the text-only barrier: + +- **Visual Q&A** — ask questions about images and documents +- **Document Analysis** — extract information from PDFs, screenshots, and scans +- **Video Understanding** — analyze video content, summarize clips +- **Audio Processing** — transcribe speech, analyze audio content +- **Accessibility** — describe images for visually impaired users +- **Content Moderation** — detect inappropriate content across modalities + +## Stats + +| Metric | Count | +| ----------------------------- | ----- | +| Multimodal models | 1519 | +| Providers | 61 | +| Image input | 1487 | +| Audio input | 118 | +| Video input | 167 | +| Free multimodal models | 53 | +| Open-weight multimodal models | 119 | +| With tool calling | 1179 | +| With reasoning | 701 | + +## Providers + +`01ai`, `302ai`, `aihubmix`, `aimlapi`, `amazon`, `amazon-bedrock`, `anthropic`, `arcee`, `auriko`, `baidu`, `berget`, `bytedance`, `chutes`, `clarifai`, `cloudferro-sherlock`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `evroc`, `fastrouter`, `fireworks`, `google`, `google-vertex` and 36 more + +## Free Multimodal Models + +Free models with multimodal input — zero-cost visual/audio applications. + +| Model | Provider | Context | Input $/M | Output $/M | Modalities | +| ---------------------------------------------------- | ---------- | ------- | --------- | ---------- | -------------- | +| google--lyria-3-clip-preview | openrouter | 1M | Free | Free | 🖼️ | +| google--lyria-3-pro-preview | openrouter | 1M | Free | Free | 🖼️ | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🖼️ 🎬 🔧 🧠 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🖼️ 🎬 🔧 🧠 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🖼️ 🔧 🧠 | +| gemma-4-31b-it | auriko | 262K | Free | Free | 🖼️ 🔧 🧠 | +| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | 256K | Free | Free | 🖼️ 🎤 🎬 🔧 🧠 | +| spotlight | arcee | 131K | Free | Free | 🖼️ | +| gemma-3-4b-it | google | 131K | Free | Free | 🖼️ | +| gemma-3-12b-it | google | 131K | Free | Free | 🖼️ | + +## Cheapest Multimodal Models + +Best value multimodal models for production. + +| Model | Provider | Context | Input $/M | Output $/M | Modalities | +| -------------------------- | --------- | ------- | --------- | ---------- | ---------- | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🖼️ 🔧 | +| deepseek-ocr | aihubmix | 0 | $0.01 | $0.01 | 🖼️ | +| gemini-2.0-flash-exp | aihubmix | 0 | $0.01 | $0.04 | 🖼️ 🎤 🎬 | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🖼️ 🔧 🧠 | +| qwen3.5-0.8b | deepinfra | 262K | $0.01 | $0.05 | 🖼️ 🧠 | +| qwen3-vl-flash-2026-01-22 | aihubmix | 0 | $0.0103 | $0.103 | 🖼️ 🎬 🔧 | +| qwen3-vl-flash | aihubmix | 0 | $0.0103 | $0.103 | 🖼️ 🎬 🔧 | +| glm-ocr | aihubmix | 0 | $0.0141 | $0.0141 | 🖼️ | +| paddlepaddle--paddleocr-vl | novitaai | 16K | $0.02 | $0.02 | 🖼️ | +| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | 🖼️ 🔧 🧠 | + +## Largest Context Multimodal Models + +Multimodal models with the largest context windows — for processing long documents and videos. + +| Model | Provider | Context | Input $/M | Output $/M | Modalities | +| ---------------------------- | ---------- | ------- | --------- | ---------- | ----------- | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 🖼️ 🔧 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | 🖼️ 🔧 | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 🖼️ 🎤 🎬 🔧 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | 🖼️ 🔧 | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | 🖼️ 🔧 | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | 🖼️ 🔧 | + +## Related Documentation + +- [Vision Models](vision-models.md) — 1,487 models with image input +- [Video Models](video-models.md) — models with video understanding +- [Audio Models](audio-models.md) — models with audio input/output +- [Image Generation](image-generation.md) — 28 models that generate images +- [Agentic Models](agentic-models.md) — 1,080 models with tool calling + reasoning +- [Free AI Models](free-models.md) — 81 free models by capability +- [Model Selection Guide](model-selection.md) — decision framework + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/chat-models.md b/docs/zh/chat-models.md new file mode 100644 index 00000000..584b0ec6 --- /dev/null +++ b/docs/zh/chat-models.md @@ -0,0 +1,99 @@ +# 聊天模型 + +[English](../chat-models.md) + +支持**工具调用**的 AI 模型 — 构建聊天应用、AI 助手和对话智能体的核心能力。这些模型可以理解自然语言、生成回复并调用外部工具。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么聊天模型很重要 + +聊天模型是现代 AI 应用的基石： + +- **对话式 AI** — 带上下文的自然语言对话 +- **AI 助手** — 面向任务的聊天与工具使用 +- **客户支持** — 带知识库访问的自动化支持 +- **内容生成** — 写作、摘要、翻译 +- **数据分析** — 自然语言查询结构化数据 + +工具调用是关键差异化能力 — 它允许模型超越文本生成，在现实世界中采取行动。 + +## 统计 + +| 指标 | 数量 | +| ------------------------ | ---- | +| 聊天模型（支持工具调用） | 2350 | +| 提供商 | 71 | +| 免费聊天模型 | 54 | +| 开源权重聊天模型 | 278 | +| 带推理能力 | 1080 | +| 带结构化输出 | 758 | + +## 提供商 + +`01ai`, `302ai`, `aihubmix`, `aimlapi`, `alibaba`, `amazon`, `amazon-bedrock`, `anthropic`, `arcee`, `auriko`, `baidu`, `baseten`, `berget`, `bytedance`, `cerebras`, `chutes`, `clarifai`, `cloudferro-sherlock`, `cloudflare`, `cortecs`, `databricks`, `deepseek`, `digitalocean`, `dinference`, `evroc` 等 46 个 + +## 免费聊天模型 + +支持工具调用的免费模型 — 零成本聊天应用。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---------------------------------------------------- | ---------- | ------ | -------- | -------- | ----- | +| openrouter--owl-alpha | openrouter | 1M | Free | Free | 📋 | +| deepseek--deepseek-v4-flash--free | openrouter | 1M | Free | Free | 🧠 | +| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | | +| nvidia--nemotron-3-super-120b-a12b--free | openrouter | 1M | Free | Free | 🧠 📋 | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🧠 📋 | +| arcee-ai--trinity-large-thinking--free | openrouter | 262K | Free | Free | 🧠 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🧠 📋 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🧠 📋 | +| gemma-4-31b-it | auriko | 262K | Free | Free | 🧠 📋 | +| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | 256K | Free | Free | 🧠 | + +## 最便宜聊天模型 + +生产环境聊天的最佳性价比模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ------------------------------------------- | ------------ | ------ | -------- | -------- | ---- | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 📋 | +| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔓 | +| inclusionai--ling-2.6-flash | openrouter | 262K | $0.01 | $0.03 | 📋 | +| ling-2.6-flash | inclusionai | 262K | $0.01 | $0.03 | | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🧠 | +| qwen3-vl-flash-2026-01-22 | aihubmix | 0 | $0.0103 | $0.103 | 📋 | +| qwen3-vl-flash | aihubmix | 0 | $0.0103 | $0.103 | 📋 | +| klusterai--Meta-Llama-3.1-8B-Instruct-Turbo | klusterai | 131K | $0.015 | $0.02 | | +| granite-4.0-h-micro | cloudflare | 131K | $0.017 | $0.112 | 🔓 | +| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | 🧠 | + +## 大上下文聊天模型 + +上下文窗口最大的聊天模型 — 适用于长对话和文档分析。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---------------------------- | ---------- | ------ | -------- | -------- | ---- | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 📋 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 📋 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | | + +## 相关文档 + +- [智能体模型](agentic-models.md) — 1,080 个具备工具调用 + 推理能力的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [代码模型](code-models.md) — 189 个代码模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/embedding-models.md b/docs/zh/embedding-models.md new file mode 100644 index 00000000..6fbf69c3 --- /dev/null +++ b/docs/zh/embedding-models.md @@ -0,0 +1,71 @@ +# 嵌入模型 + +[English](../embedding-models.md) + +生成**向量嵌入**的 AI 模型 — 文本、图像和其他数据的数值表示。语义搜索、RAG（检索增强生成）、聚类和相似度任务的基础。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么嵌入模型很重要 + +嵌入模型是许多 AI 系统的基础： + +- **语义搜索** — 按含义而非关键词查找相关文档 +- **RAG** — 为 LLM 检索上下文以生成有依据的答案 +- **聚类** — 将相似项目分组 +- **相似度** — 查找重复项、推荐和相关内容 +- **分类** — 通过嵌入进行零样本和少样本分类 + +## 统计 + +| 指标 | 数量 | +| ---------------- | ---- | +| 嵌入模型 | 5 | +| 提供商 | 3 | +| 免费嵌入模型 | 0 | +| 开源权重嵌入模型 | 1 | + +## 提供商 + +`openai`、`tencent`、`upstage` + +## 免费嵌入模型 + +免费嵌入模型 — 零成本语义搜索和 RAG。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | +| ---- | ------ | ------ | -------- | + +## 最便宜嵌入模型 + +生产环境嵌入的最佳性价比模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | +| ----------------------- | ------- | ------ | -------- | --- | +| text-embedding-3-small | openai | 8K | $0.02 | | +| solar-embedding-1-large | upstage | 0 | $0.1 | 🔓 | +| text-embedding-ada-002 | openai | 8K | $0.1 | | +| text-embedding-3-large | openai | 8K | $0.13 | | +| hunyuan-embedding | tencent | 0 | $0.7 | | + +## 大上下文嵌入模型 + +上下文窗口最大的嵌入模型 — 适用于嵌入长文档。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | +| ---------------------- | ------ | ------ | -------- | --- | +| text-embedding-ada-002 | openai | 8K | $0.1 | | +| text-embedding-3-small | openai | 8K | $0.02 | | +| text-embedding-3-large | openai | 8K | $0.13 | | + +## 相关文档 + +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [开源权重](open-weights.md) — 527 个开源权重模型 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [API 参考](api.md) — 编程访问 +- [快速入门](quick-start.md) — 5 分钟上手 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/multimodal-models.md b/docs/zh/multimodal-models.md new file mode 100644 index 00000000..a3c54160 --- /dev/null +++ b/docs/zh/multimodal-models.md @@ -0,0 +1,101 @@ +# 多模态模型 + +[English](../multimodal-models.md) + +能够处理**多种输入模态**的 AI 模型 — 图像、音频和视频与文本并行。这些模型驱动视觉问答、文档分析、视频理解和音频转录。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么多模态模型很重要 + +多模态模型打破了纯文本的限制： + +- **视觉问答** — 对图像和文档提问 +- **文档分析** — 从 PDF、截图和扫描件中提取信息 +- **视频理解** — 分析视频内容，总结片段 +- **音频处理** — 转录语音，分析音频内容 +- **无障碍** — 为视障用户描述图像 +- **内容审核** — 跨模态检测不当内容 + +## 统计 + +| 指标 | 数量 | +| ------------------ | ---- | +| 多模态模型 | 1519 | +| 提供商 | 61 | +| 图像输入 | 1487 | +| 音频输入 | 118 | +| 视频输入 | 167 | +| 免费多模态模型 | 53 | +| 开源权重多模态模型 | 119 | +| 带工具调用 | 1179 | +| 带推理能力 | 701 | + +## 提供商 + +`01ai`, `302ai`, `aihubmix`, `aimlapi`, `amazon`, `amazon-bedrock`, `anthropic`, `arcee`, `auriko`, `baidu`, `berget`, `bytedance`, `chutes`, `clarifai`, `cloudferro-sherlock`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `evroc`, `fastrouter`, `fireworks`, `google`, `google-vertex` 等 36 个 + +## 免费多模态模型 + +支持多模态输入的免费模型 — 零成本视觉/音频应用。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 模态 | +| ---------------------------------------------------- | ---------- | ------ | -------- | -------- | -------------- | +| google--lyria-3-clip-preview | openrouter | 1M | Free | Free | 🖼️ | +| google--lyria-3-pro-preview | openrouter | 1M | Free | Free | 🖼️ | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🖼️ 🎬 🔧 🧠 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🖼️ 🎬 🔧 🧠 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🖼️ 🔧 🧠 | +| gemma-4-31b-it | auriko | 262K | Free | Free | 🖼️ 🔧 🧠 | +| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | 256K | Free | Free | 🖼️ 🎤 🎬 🔧 🧠 | +| spotlight | arcee | 131K | Free | Free | 🖼️ | +| gemma-3-4b-it | google | 131K | Free | Free | 🖼️ | +| gemma-3-12b-it | google | 131K | Free | Free | 🖼️ | + +## 最便宜多模态模型 + +生产环境多模态应用的最佳性价比模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 模态 | +| -------------------------- | --------- | ------ | -------- | -------- | -------- | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🖼️ 🔧 | +| deepseek-ocr | aihubmix | 0 | $0.01 | $0.01 | 🖼️ | +| gemini-2.0-flash-exp | aihubmix | 0 | $0.01 | $0.04 | 🖼️ 🎤 🎬 | +| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | 🖼️ 🔧 🧠 | +| qwen3.5-0.8b | deepinfra | 262K | $0.01 | $0.05 | 🖼️ 🧠 | +| qwen3-vl-flash-2026-01-22 | aihubmix | 0 | $0.0103 | $0.103 | 🖼️ 🎬 🔧 | +| qwen3-vl-flash | aihubmix | 0 | $0.0103 | $0.103 | 🖼️ 🎬 🔧 | +| glm-ocr | aihubmix | 0 | $0.0141 | $0.0141 | 🖼️ | +| paddlepaddle--paddleocr-vl | novitaai | 16K | $0.02 | $0.02 | 🖼️ | +| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | 🖼️ 🔧 🧠 | + +## 大上下文多模态模型 + +上下文窗口最大的多模态模型 — 适用于处理长文档和视频。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 模态 | +| ---------------------------- | ---------- | ------ | -------- | -------- | ----------- | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 🖼️ 🔧 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | 🖼️ 🔧 | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 🖼️ 🎤 🎬 🔧 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | 🖼️ 🔧 | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | 🖼️ 🔧 | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🖼️ 🔧 | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | 🖼️ 🔧 | + +## 相关文档 + +- [视觉模型](vision-models.md) — 1,487 个支持图像输入的模型 +- [视频模型](video-models.md) — 支持视频理解的模型 +- [音频模型](audio-models.md) — 支持音频输入/输出的模型 +- [图像生成](image-generation.md) — 28 个图像生成模型 +- [智能体模型](agentic-models.md) — 1,080 个具备工具调用 + 推理能力的模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/llms-full.txt b/llms-full.txt index 6b807eb3..1efe84ae 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -332,6 +332,9 @@ by_context = sorted( - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models - [Code Models](docs/code-models.md) — 189 code-focused models across 41 providers - [Agentic Models](docs/agentic-models.md) — models with tool calling + reasoning for AI agents +- [Chat Models](docs/chat-models.md) — 2,350 models with tool calling for chat applications +- [Multimodal Models](docs/multimodal-models.md) — 1,519 models with image/audio/video input +- [Embedding Models](docs/embedding-models.md) — 5 embedding models for search, RAG, similarity - [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video diff --git a/llms.txt b/llms.txt index 6b807eb3..1efe84ae 100644 --- a/llms.txt +++ b/llms.txt @@ -332,6 +332,9 @@ by_context = sorted( - [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models - [Code Models](docs/code-models.md) — 189 code-focused models across 41 providers - [Agentic Models](docs/agentic-models.md) — models with tool calling + reasoning for AI agents +- [Chat Models](docs/chat-models.md) — 2,350 models with tool calling for chat applications +- [Multimodal Models](docs/multimodal-models.md) — 1,519 models with image/audio/video input +- [Embedding Models](docs/embedding-models.md) — 5 embedding models for search, RAG, similarity - [Video Models](docs/video-models.md) — 167 video input + 4 video output models - [Structured Output](docs/structured-output.md) — 829 JSON-mode models - [Modality Matrix](docs/modality-matrix.md) — vision, image gen, audio, video diff --git a/package.json b/package.json index cffe2562..24f580fa 100644 --- a/package.json +++ b/package.json @@ -15,11 +15,13 @@ "cached-pricing", "cdn", "cerebras", + "chat-models", "chatgpt", "claude", "code-models", "context-window", "deepseek", + "embedding-models", "free-models", "function-calling", "github-action", @@ -43,6 +45,7 @@ "model-metadata", "model-pricing", "model-selection", + "multimodal-models", "open-weights", "openai", "prompt-caching", diff --git a/site/index.html b/site/index.html index 56a099c3..5c6a2371 100644 --- a/site/index.html +++ b/site/index.html @@ -442,6 +442,18 @@

🤖 AI Models CatalogAPI · FAQ + · + Chat Models + · + Multimodal + · + Embeddings
diff --git a/site/sitemap.xml b/site/sitemap.xml index 5e510d5b..df05b8b4 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -25,6 +25,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/chat-models + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/code-examples weekly @@ -50,6 +55,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/embedding-models + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/faq weekly @@ -90,6 +100,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/multimodal-models + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/open-weights weekly diff --git a/social-preview.svg b/social-preview.svg index 923ef6cf..5b7a630a 100644 --- a/social-preview.svg +++ b/social-preview.svg @@ -32,27 +32,22 @@ The most comprehensive structured catalog of AI models on GitHub - 4,587 Models - 95 Providers - 2,350 Tool Calling - 1,306 Reasoning - 81 Free From 691b5d716562f1595078b5d878d257e40c71ce7f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:13:53 +0800 Subject: [PATCH 178/311] Add cross-links to chat-models and multimodal-models from 14 existing docs --- docs/model-selection.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/docs/model-selection.md b/docs/model-selection.md index dc3c1720..9832d4dd 100644 --- a/docs/model-selection.md +++ b/docs/model-selection.md @@ -122,6 +122,9 @@ Models with the largest context windows — for long documents, multi-turn conve - [Vision Models](vision-models.md) — 1,487 models with image understanding - [Quick Start Guide](quick-start.md) — get started in 30 seconds +- [Chat Models](chat-models.md) — 2,350 models with tool calling for chat applications +- [Multimodal Models](multimodal-models.md) — 1,519 models with image/audio/video input + --- Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. From 177760151d963d9e568273478ed2ab2128524fc1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:17:55 +0800 Subject: [PATCH 179/311] =?UTF-8?q?Add=20migration-guide.md=20=E2=80=94=20?= =?UTF-8?q?switching=20AI=20providers=20with=20pricing,=20API=20compatibil?= =?UTF-8?q?ity,=20checklist?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - EN + ZH bilingual with cross-links - Major provider comparison table - Common migration paths with savings percentages - OpenAI-compatible provider list - Migration checklist (10 compatibility points) - Added to all index files: README, AGENTS.md, llms.txt, llms-full.txt, sitemap.xml, site/index.html, CITATION.cff, package.json --- CITATION.cff | 1 + README.md | 1 + docs/migration-guide.md | 99 ++++++++++++++++++++++++++++++++++++++ docs/zh/migration-guide.md | 99 ++++++++++++++++++++++++++++++++++++++ package.json | 1 + site/sitemap.xml | 5 ++ 6 files changed, 206 insertions(+) create mode 100644 docs/migration-guide.md create mode 100644 docs/zh/migration-guide.md diff --git a/CITATION.cff b/CITATION.cff index bfc2ceae..b43d9a95 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -40,6 +40,7 @@ keywords: - chat-models - multimodal-models - embedding-models + - migration-guide license: MIT version: 0.1.0 date-released: "2026-05-21" diff --git a/README.md b/README.md index 001f6b53..12579c09 100644 --- a/README.md +++ b/README.md @@ -453,6 +453,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [Model Selection Guide](docs/model-selection.md) | Decision framework: free, best value, large context models | +| [Migration Guide](docs/migration-guide.md) | Switch providers — pricing, API compatibility, checklist | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | [Code Examples](docs/code-examples.md) | Practical examples in TypeScript, Python, Go, Rust, jq | | [FAQ](docs/faq.md) | Common questions about the catalog, data, and contributing | diff --git a/docs/migration-guide.md b/docs/migration-guide.md new file mode 100644 index 00000000..29503136 --- /dev/null +++ b/docs/migration-guide.md @@ -0,0 +1,99 @@ +# Migration Guide: Switching AI Model Providers + +[中文](zh/migration-guide.md) + +A practical guide for switching between AI model providers — compare pricing, capabilities, and context windows to find the best alternative for your use case. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Switch Providers? + +- **Cost savings** — some providers offer the same models at 2-10× lower prices +- **Better capabilities** — newer models may offer tool calling, reasoning, or vision +- **Larger context** — process more data in a single request +- **Reliability** — reduce dependency on a single provider +- **Compliance** — data residency requirements may require specific providers + +## Major Provider Comparison + +| Provider | Models | Cheapest Input $/M | Largest Context | Tool Calling | Reasoning | +| --------- | ------ | -----------------: | --------------- | ------------ | --------- | +| openai | 28 | $0.02 | 1047576 | 18 | 8 | +| anthropic | 11 | $1 | 1000000 | 11 | 11 | +| google | 21 | $0.075 | 2097152 | 8 | 2 | +| deepseek | 4 | $0.14 | 1000000 | 4 | 3 | +| meta | 12 | $0.1 | 10000000 | 9 | 0 | +| mistral | 16 | $0.04 | 256000 | 12 | 1 | +| xai | 6 | $0.2 | 131072 | 6 | 5 | +| alibaba | 62 | $0.15 | 1000000 | 62 | 52 | + +## Common Migration Paths + +### OpenAI → Cheaper Alternatives + +| OpenAI Model | Cheapest Alternative | Provider | Input $/M | Savings | +| -------------------- | -------------------- | --------- | --------- | ------- | +| gpt-4.1 ($2) | gpt-4.1-mini | openai | $0.40 | 80% | +| gpt-4.1-mini ($0.40) | gpt-4.1-nano | openai | $0.10 | 75% | +| o4-mini ($1.10) | deepseek-r1 | deepseek | $0.55 | 50% | +| gpt-4.1 ($2) | claude-haiku-4 | anthropic | $1 | 50% | +| gpt-4.1 ($2) | gemini-2.5-flash | google | $0.15 | 93% | + +### Anthropic → Cheaper Alternatives + +| Anthropic Model | Cheapest Alternative | Provider | Input $/M | Savings | +| -------------------- | -------------------- | -------- | --------- | ------- | +| claude-opus-4 ($15) | o4-mini | openai | $1.10 | 93% | +| claude-sonnet-4 ($3) | gemini-2.5-flash | google | $0.15 | 95% | +| claude-sonnet-4 ($3) | deepseek-chat | deepseek | $0.14 | 95% | +| claude-haiku-4 ($1) | gemini-2.5-flash | google | $0.15 | 85% | + +### Google → Cheaper Alternatives + +| Google Model | Cheapest Alternative | Provider | Input $/M | Savings | +| ---------------------- | -------------------- | -------- | --------- | ------- | +| gemini-2.5-pro ($1.25) | gemini-2.5-flash | google | $0.15 | 88% | +| gemini-2.5-pro ($1.25) | deepseek-chat | deepseek | $0.14 | 89% | + +## Migration Checklist + +When switching providers, verify these compatibility points: + +- [ ] **API format** — OpenAI-compatible vs proprietary API +- [ ] **Model names** — different providers use different model IDs +- [ ] **Tool calling format** — function calling syntax varies +- [ ] **Streaming** — SSE vs WebSocket vs HTTP streaming +- [ ] **Rate limits** — requests per minute, tokens per minute +- [ ] **Context window** — may differ from original provider +- [ ] **Modalities** — vision, audio, video support varies +- [ ] **Structured output** — JSON mode availability +- [ ] **Prompt caching** — can reduce costs 50-90% +- [ ] **Data residency** — where is data processed and stored + +## OpenAI-Compatible Providers + +These providers offer OpenAI-compatible APIs — minimal code changes needed: + +| Provider | Base URL | Notes | +| ----------- | ------------------------------- | --------------------------- | +| openrouter | `openrouter.ai/api/v1` | Aggregator, 356+ models | +| deepinfra | `api.deepinfra.com/v1` | Focus on open-source models | +| togetherai | `api.together.xyz/v1` | Open-source model hosting | +| groq | `api.groq.com/openai/v1` | Ultra-fast inference | +| cerebras | `api.cerebras.ai/v1` | Fastest inference speed | +| fireworks | `api.fireworks.ai/inference/v1` | Serverless model hosting | +| siliconflow | `api.siliconflow.cn/v1` | China-focused provider | + +## Related Documentation + +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Model Selection Guide](model-selection.md) — decision framework for choosing models +- [Free AI Models](free-models.md) — 81 free models by capability +- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [Chat Models](chat-models.md) — 2,350 models with tool calling +- [Agentic Models](agentic-models.md) — 1,080 models with tool calling + reasoning +- [API Reference](api.md) — programmatic access to model data + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/migration-guide.md b/docs/zh/migration-guide.md new file mode 100644 index 00000000..6d2602b0 --- /dev/null +++ b/docs/zh/migration-guide.md @@ -0,0 +1,99 @@ +# 迁移指南：切换 AI 模型提供商 + +[English](../migration-guide.md) + +切换 AI 模型提供商的实用指南 — 比较定价、能力和上下文窗口，找到最适合您用例的替代方案。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么要切换提供商？ + +- **节省成本** — 某些提供商以低 2-10 倍的价格提供相同模型 +- **更好的能力** — 新模型可能提供工具调用、推理或视觉能力 +- **更大的上下文** — 在单个请求中处理更多数据 +- **可靠性** — 减少对单一提供商的依赖 +- **合规性** — 数据驻留要求可能需要特定提供商 + +## 主要提供商对比 + +| 提供商 | 模型数 | 最低输入 $/M | 最大上下文 | 工具调用 | 推理 | +| --------- | ------ | -----------: | ---------- | -------- | ---- | +| openai | 28 | $0.02 | 1047576 | 18 | 8 | +| anthropic | 11 | $1 | 1000000 | 11 | 11 | +| google | 21 | $0.075 | 2097152 | 8 | 2 | +| deepseek | 4 | $0.14 | 1000000 | 4 | 3 | +| meta | 12 | $0.1 | 10000000 | 9 | 0 | +| mistral | 16 | $0.04 | 256000 | 12 | 1 | +| xai | 6 | $0.2 | 131072 | 6 | 5 | +| alibaba | 62 | $0.15 | 1000000 | 62 | 52 | + +## 常见迁移路径 + +### OpenAI → 更便宜的替代方案 + +| OpenAI 模型 | 最便宜的替代方案 | 提供商 | 输入 $/M | 节省 | +| -------------------- | ---------------- | --------- | -------- | ---- | +| gpt-4.1 ($2) | gpt-4.1-mini | openai | $0.40 | 80% | +| gpt-4.1-mini ($0.40) | gpt-4.1-nano | openai | $0.10 | 75% | +| o4-mini ($1.10) | deepseek-r1 | deepseek | $0.55 | 50% | +| gpt-4.1 ($2) | claude-haiku-4 | anthropic | $1 | 50% | +| gpt-4.1 ($2) | gemini-2.5-flash | google | $0.15 | 93% | + +### Anthropic → 更便宜的替代方案 + +| Anthropic 模型 | 最便宜的替代方案 | 提供商 | 输入 $/M | 节省 | +| -------------------- | ---------------- | -------- | -------- | ---- | +| claude-opus-4 ($15) | o4-mini | openai | $1.10 | 93% | +| claude-sonnet-4 ($3) | gemini-2.5-flash | google | $0.15 | 95% | +| claude-sonnet-4 ($3) | deepseek-chat | deepseek | $0.14 | 95% | +| claude-haiku-4 ($1) | gemini-2.5-flash | google | $0.15 | 85% | + +### Google → 更便宜的替代方案 + +| Google 模型 | 最便宜的替代方案 | 提供商 | 输入 $/M | 节省 | +| ---------------------- | ---------------- | -------- | -------- | ---- | +| gemini-2.5-pro ($1.25) | gemini-2.5-flash | google | $0.15 | 88% | +| gemini-2.5-pro ($1.25) | deepseek-chat | deepseek | $0.14 | 89% | + +## 迁移检查清单 + +切换提供商时，请验证以下兼容性要点： + +- [ ] **API 格式** — OpenAI 兼容 vs 专有 API +- [ ] **模型名称** — 不同提供商使用不同的模型 ID +- [ ] **工具调用格式** — 函数调用语法不同 +- [ ] **流式传输** — SSE vs WebSocket vs HTTP 流式传输 +- [ ] **速率限制** — 每分钟请求数、每分钟 token 数 +- [ ] **上下文窗口** — 可能与原提供商不同 +- [ ] **模态** — 视觉、音频、视频支持不同 +- [ ] **结构化输出** — JSON 模式可用性 +- [ ] **提示缓存** — 可降低 50-90% 成本 +- [ ] **数据驻留** — 数据在哪里处理和存储 + +## OpenAI 兼容提供商 + +这些提供商提供 OpenAI 兼容 API — 只需最少的代码更改： + +| 提供商 | Base URL | 备注 | +| ----------- | ------------------------------- | ----------------- | +| openrouter | `openrouter.ai/api/v1` | 聚合器，356+ 模型 | +| deepinfra | `api.deepinfra.com/v1` | 专注开源模型 | +| togetherai | `api.together.xyz/v1` | 开源模型托管 | +| groq | `api.groq.com/openai/v1` | 超快推理 | +| cerebras | `api.cerebras.ai/v1` | 最快推理速度 | +| fireworks | `api.fireworks.ai/inference/v1` | 无服务器模型托管 | +| siliconflow | `api.siliconflow.cn/v1` | 中国市场提供商 | + +## 相关文档 + +- [定价对比](pricing-comparison.md) — 跨提供商定价对比 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [聊天模型](chat-models.md) — 2,350 个支持工具调用的聊天模型 +- [智能体模型](agentic-models.md) — 1,080 个具备工具调用 + 推理能力的模型 +- [API 参考](api.md) — 编程访问模型数据 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/package.json b/package.json index 24f580fa..1305efc7 100644 --- a/package.json +++ b/package.json @@ -38,6 +38,7 @@ "llm-catalog", "llm-pricing", "machine-readable", + "migration-guide", "mistral", "model-catalog", "model-comparison", diff --git a/site/sitemap.xml b/site/sitemap.xml index df05b8b4..0a495b7e 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -85,6 +85,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/migration-guide + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/modality-matrix weekly From 38d3ce002c526c601eeec2fe3d47ff386a4e5456 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:21:27 +0800 Subject: [PATCH 180/311] Fix: add migration-guide to AGENTS.md, llms.txt, llms-full.txt --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index 263d4e6e..84640015 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -33,6 +33,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) - [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) - [`docs/model-selection.md`](docs/model-selection.md) — Model selection guide: free, best value, large context ([中文](docs/zh/model-selection.md)) +- [`docs/migration-guide.md`](docs/migration-guide.md) — Switch providers: pricing, API compatibility, checklist ([中文](docs/zh/migration-guide.md)) - [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) - [`docs/code-examples.md`](docs/code-examples.md) — code examples in multiple languages ([中文](docs/zh/code-examples.md)) - [`docs/faq.md`](docs/faq.md) — frequently asked questions ([中文](docs/zh/faq.md)) From 92b27852bc46c4a7dd04cdd3476403bf6e97613e Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:29:59 +0800 Subject: [PATCH 181/311] =?UTF-8?q?Add=20provider-comparison.md=20?= =?UTF-8?q?=E2=80=94=20top=2030=20providers=20with=20model=20count,=20capa?= =?UTF-8?q?bilities,=20pricing?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - EN + ZH bilingual with cross-links - Top 30 providers by model count with free/tool_call/reasoning/vision stats - Provider categories: aggregators, direct, infrastructure, regional - 'Choosing a Provider' decision table (8 use cases) - Added to all index files: README, AGENTS.md, llms.txt, llms-full.txt, sitemap.xml, site/index.html, CITATION.cff, package.json --- CITATION.cff | 1 + docs/provider-comparison.md | 127 +++++++++++++++++++++++++++++++++ docs/zh/provider-comparison.md | 127 +++++++++++++++++++++++++++++++++ package.json | 1 + site/sitemap.xml | 5 ++ 5 files changed, 261 insertions(+) create mode 100644 docs/provider-comparison.md create mode 100644 docs/zh/provider-comparison.md diff --git a/CITATION.cff b/CITATION.cff index b43d9a95..25b7b9b3 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -41,6 +41,7 @@ keywords: - multimodal-models - embedding-models - migration-guide + - provider-comparison license: MIT version: 0.1.0 date-released: "2026-05-21" diff --git a/docs/provider-comparison.md b/docs/provider-comparison.md new file mode 100644 index 00000000..22c13572 --- /dev/null +++ b/docs/provider-comparison.md @@ -0,0 +1,127 @@ +# Provider Comparison + +[中文](zh/provider-comparison.md) + +Side-by-side comparison of AI model providers — model count, capabilities, pricing, and context windows at a glance. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Top 30 Providers by Model Count + +| Provider | Models | Free | Tool Call | Reasoning | Vision | Cheapest $/M | Capabilities | +| -------------- | -----: | ---: | --------: | --------: | -----: | -----------: | ------------ | +| nanogpt | 547 | 0 | 0 | 0 | 0 | $0.02 | | +| aihubmix | 476 | 0 | 132 | 74 | 145 | $0.00 | 🔧 🧠 👁️ 🎤 | +| openrouter | 356 | 29 | 263 | 190 | 160 | $0.01 | 🔧 🧠 👁️ 🎤 | +| martian | 304 | 0 | 0 | 3 | 2 | $0.02 | 🧠 👁️ | +| requesty | 277 | 0 | 251 | 139 | 151 | $0.02 | 🔧 🧠 👁️ | +| 302ai | 268 | 0 | 190 | 44 | 144 | $0.00 | 🔧 🧠 👁️ 🔓 | +| auriko | 181 | 5 | 154 | 108 | 93 | $0.01 | 🔧 🧠 👁️ 🖼️ | +| llmgateway | 163 | 3 | 158 | 85 | 89 | $0.03 | 🔧 🧠 👁️ 🖼️ | +| aimlapi | 147 | 2 | 21 | 0 | 14 | $0.01 | 🔧 👁️ | +| fastrouter | 120 | 2 | 94 | 66 | 65 | $0.02 | 🔧 🧠 👁️ 🎤 | +| orcarouter | 120 | 0 | 102 | 64 | 111 | $0.05 | 🔧 🧠 👁️ 🔓 | +| cortecs | 105 | 0 | 97 | 82 | 52 | $0.02 | 🔧 🧠 👁️ 🎤 | +| novitaai | 104 | 2 | 72 | 53 | 33 | $0.02 | 🔧 🧠 👁️ 🎤 | +| vultr | 98 | 0 | 11 | 22 | 23 | $0.55 | 🔧 🧠 👁️ 🎤 | +| deepinfra | 88 | 0 | 0 | 51 | 38 | $0.01 | 🧠 👁️ | +| venice | 75 | 0 | 64 | 55 | 39 | $0.05 | 🔧 🧠 👁️ | +| jiekou | 73 | 0 | 73 | 0 | 49 | $0.03 | 🔧 👁️ 🔓 | +| meganova | 63 | 4 | 60 | 7 | 37 | $0.02 | 🔧 🧠 👁️ 🔓 | +| alibaba | 62 | 0 | 62 | 52 | 0 | $0.15 | 🔧 🧠 | +| ppio | 60 | 1 | 46 | 12 | 11 | $0.21 | 🔧 🧠 👁️ 🔓 | +| amazon-bedrock | 57 | 0 | 37 | 0 | 16 | $0.04 | 🔧 👁️ 🎤 🎬 | +| google-vertex | 38 | 0 | 32 | 0 | 19 | $0.07 | 🔧 👁️ 🎤 🎬 | +| siliconflow-cn | 37 | 0 | 2 | 7 | 9 | $0.50 | 🔧 🧠 👁️ | +| stepfun | 31 | 14 | 0 | 0 | 11 | $0.70 | 👁️ 🎤 🖼️ | +| cloudflare | 30 | 0 | 15 | 10 | 7 | $0.02 | 🔧 🧠 👁️ 🔓 | +| gmicloud | 29 | 0 | 11 | 10 | 0 | $0.07 | 🔧 🧠 🔓 | +| databricks | 29 | 0 | 4 | 0 | 10 | $0.05 | 🔧 👁️ 🔓 | +| openai | 28 | 5 | 18 | 8 | 12 | $0.02 | 🔧 🧠 👁️ 🎤 | +| siliconflow | 27 | 0 | 24 | 2 | 3 | $0.04 | 🔧 🧠 👁️ 🔓 | +| togetherai | 24 | 0 | 22 | 2 | 0 | $0.03 | 🔧 🧠 🔓 | + +## Provider Categories + +### Aggregators (Multi-Provider Access) + +These providers offer access to models from multiple AI companies through a single API: + +| Provider | Models | Notes | +| ---------- | -----: | ----------------------------------------------- | +| openrouter | 356 | Largest model aggregator, OpenAI-compatible API | +| requesty | 277 | Smart routing across providers | +| martian | 304 | Multi-provider with load balancing | +| aihubmix | 476 | Chinese market aggregator | +| nanogpt | 547 | Pay-per-token, no subscription | +| llmgateway | 163 | Enterprise API gateway | +| fastrouter | 120 | Fast model routing | +| orcarouter | 120 | Multi-provider routing | + +### Direct Providers (First-Party APIs) + +| Provider | Models | Specialty | +| --------- | -----: | ---------------------------------- | +| openai | 28 | GPT-4.1, o3/o4 reasoning models | +| anthropic | 11 | Claude 4 family, best for agents | +| google | 21 | Gemini 2.5, 1M+ context | +| deepseek | 4 | DeepSeek R1, best open reasoning | +| meta | 12 | Llama 4, open weights | +| mistral | 16 | Mistral Large, Codestral | +| xai | 6 | Grok 3, real-time data | +| alibaba | 62 | Qwen 3, largest open-source family | + +### Infrastructure Providers (Hosted Open-Source) + +| Provider | Models | Specialty | +| ---------- | -----: | ------------------------------ | +| groq | 12 | Fastest inference (LPU) | +| cerebras | 11 | Ultra-fast inference (CS-3) | +| togetherai | 24 | Serverless open-source hosting | +| deepinfra | 88 | Cost-effective inference | +| fireworks | 10 | Serverless model hosting | +| cloudflare | 30 | Edge inference (Workers AI) | + +### Regional Providers + +| Provider | Models | Region | +| ------------------- | -----: | ---------------- | +| siliconflow | 27 | China | +| siliconflow-cn | 37 | China (domestic) | +| stepfun | 31 | China | +| zhipuai | 20 | China | +| baichuan | 11 | China | +| baidu | 8 | China | +| iflytek | 6 | China | +| tencent | 14 | China | +| ppio | 60 | China | +| ovhcloud | 12 | Europe (France) | +| scaleway | 13 | Europe (France) | +| cloudferro-sherlock | 12 | Europe (EU) | + +## Choosing a Provider + +| If you need... | Best provider | Why | +| --------------------- | ------------------------------ | ------------------------- | +| **Cheapest prices** | deepseek, google | Input from $0.14/M tokens | +| **Fastest inference** | groq, cerebras | Sub-100ms latency | +| **Largest context** | google, meta | 1M-10M token context | +| **Most models** | nanogpt, aihubmix | 500+ models each | +| **Best for agents** | anthropic, openai | Tool calling + reasoning | +| **Open weights** | meta, deepseek | Run on your own hardware | +| **EU data residency** | ovhcloud, scaleway, cloudferro | EU-hosted inference | +| **China access** | siliconflow, ppio, stepfun | China-based endpoints | + +## Related Documentation + +- [Migration Guide](migration-guide.md) — switching providers with pricing comparison +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Providers Overview](providers.md) — all 95 providers listed +- [Model Selection Guide](model-selection.md) — decision framework +- [Free AI Models](free-models.md) — 81 free models by capability +- [Chat Models](chat-models.md) — 2,350 models with tool calling +- [Agentic Models](agentic-models.md) — 1,080 models with tool calling + reasoning + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/provider-comparison.md b/docs/zh/provider-comparison.md new file mode 100644 index 00000000..f022cf11 --- /dev/null +++ b/docs/zh/provider-comparison.md @@ -0,0 +1,127 @@ +# 提供商对比 + +[English](../provider-comparison.md) + +AI 模型提供商的并排对比 — 模型数量、能力、定价和上下文窗口一目了然。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 按模型数量排名的前 30 个提供商 + +| 提供商 | 模型 | 免费 | 工具调用 | 推理 | 视觉 | 最低 $/M | 能力 | +| -------------- | ---: | ---: | -------: | ---: | ---: | -------: | ----------- | +| nanogpt | 547 | 0 | 0 | 0 | 0 | $0.02 | | +| aihubmix | 476 | 0 | 132 | 74 | 145 | $0.00 | 🔧 🧠 👁️ 🎤 | +| openrouter | 356 | 29 | 263 | 190 | 160 | $0.01 | 🔧 🧠 👁️ 🎤 | +| martian | 304 | 0 | 0 | 3 | 2 | $0.02 | 🧠 👁️ | +| requesty | 277 | 0 | 251 | 139 | 151 | $0.02 | 🔧 🧠 👁️ | +| 302ai | 268 | 0 | 190 | 44 | 144 | $0.00 | 🔧 🧠 👁️ 🔓 | +| auriko | 181 | 5 | 154 | 108 | 93 | $0.01 | 🔧 🧠 👁️ 🖼️ | +| llmgateway | 163 | 3 | 158 | 85 | 89 | $0.03 | 🔧 🧠 👁️ 🖼️ | +| aimlapi | 147 | 2 | 21 | 0 | 14 | $0.01 | 🔧 👁️ | +| fastrouter | 120 | 2 | 94 | 66 | 65 | $0.02 | 🔧 🧠 👁️ 🎤 | +| orcarouter | 120 | 0 | 102 | 64 | 111 | $0.05 | 🔧 🧠 👁️ 🔓 | +| cortecs | 105 | 0 | 97 | 82 | 52 | $0.02 | 🔧 🧠 👁️ 🎤 | +| novitaai | 104 | 2 | 72 | 53 | 33 | $0.02 | 🔧 🧠 👁️ 🎤 | +| vultr | 98 | 0 | 11 | 22 | 23 | $0.55 | 🔧 🧠 👁️ 🎤 | +| deepinfra | 88 | 0 | 0 | 51 | 38 | $0.01 | 🧠 👁️ | +| venice | 75 | 0 | 64 | 55 | 39 | $0.05 | 🔧 🧠 👁️ | +| jiekou | 73 | 0 | 73 | 0 | 49 | $0.03 | 🔧 👁️ 🔓 | +| meganova | 63 | 4 | 60 | 7 | 37 | $0.02 | 🔧 🧠 👁️ 🔓 | +| alibaba | 62 | 0 | 62 | 52 | 0 | $0.15 | 🔧 🧠 | +| ppio | 60 | 1 | 46 | 12 | 11 | $0.21 | 🔧 🧠 👁️ 🔓 | +| amazon-bedrock | 57 | 0 | 37 | 0 | 16 | $0.04 | 🔧 👁️ 🎤 🎬 | +| google-vertex | 38 | 0 | 32 | 0 | 19 | $0.07 | 🔧 👁️ 🎤 🎬 | +| siliconflow-cn | 37 | 0 | 2 | 7 | 9 | $0.50 | 🔧 🧠 👁️ | +| stepfun | 31 | 14 | 0 | 0 | 11 | $0.70 | 👁️ 🎤 🖼️ | +| cloudflare | 30 | 0 | 15 | 10 | 7 | $0.02 | 🔧 🧠 👁️ 🔓 | +| gmicloud | 29 | 0 | 11 | 10 | 0 | $0.07 | 🔧 🧠 🔓 | +| databricks | 29 | 0 | 4 | 0 | 10 | $0.05 | 🔧 👁️ 🔓 | +| openai | 28 | 5 | 18 | 8 | 12 | $0.02 | 🔧 🧠 👁️ 🎤 | +| siliconflow | 27 | 0 | 24 | 2 | 3 | $0.04 | 🔧 🧠 👁️ 🔓 | +| togetherai | 24 | 0 | 22 | 2 | 0 | $0.03 | 🔧 🧠 🔓 | + +## 提供商分类 + +### 聚合器（多提供商访问） + +这些提供商通过单一 API 提供多家 AI 公司的模型： + +| 提供商 | 模型 | 备注 | +| ---------- | ---: | --------------------------------- | +| openrouter | 356 | 最大的模型聚合器，OpenAI 兼容 API | +| requesty | 277 | 智能路由 | +| martian | 304 | 多提供商负载均衡 | +| aihubmix | 476 | 中国市场聚合器 | +| nanogpt | 547 | 按量付费，无需订阅 | +| llmgateway | 163 | 企业 API 网关 | +| fastrouter | 120 | 快速模型路由 | +| orcarouter | 120 | 多提供商路由 | + +### 直供提供商（第一方 API） + +| 提供商 | 模型 | 专长 | +| --------- | ---: | --------------------------- | +| openai | 28 | GPT-4.1, o3/o4 推理模型 | +| anthropic | 11 | Claude 4 系列，最适合 Agent | +| google | 21 | Gemini 2.5, 1M+ 上下文 | +| deepseek | 4 | DeepSeek R1，最佳开源推理 | +| meta | 12 | Llama 4，开源权重 | +| mistral | 16 | Mistral Large, Codestral | +| xai | 6 | Grok 3，实时数据 | +| alibaba | 62 | Qwen 3，最大的开源家族 | + +### 基础设施提供商（托管开源模型） + +| 提供商 | 模型 | 专长 | +| ---------- | ---: | --------------------- | +| groq | 12 | 最快推理 (LPU) | +| cerebras | 11 | 超快推理 (CS-3) | +| togetherai | 24 | 无服务器开源托管 | +| deepinfra | 88 | 高性价比推理 | +| fireworks | 10 | 无服务器模型托管 | +| cloudflare | 30 | 边缘推理 (Workers AI) | + +### 区域提供商 + +| 提供商 | 模型 | 区域 | +| ------------------- | ---: | ------------ | +| siliconflow | 27 | 中国 | +| siliconflow-cn | 37 | 中国（国内） | +| stepfun | 31 | 中国 | +| zhipuai | 20 | 中国 | +| baichuan | 11 | 中国 | +| baidu | 8 | 中国 | +| iflytek | 6 | 中国 | +| tencent | 14 | 中国 | +| ppio | 60 | 中国 | +| ovhcloud | 12 | 欧洲（法国） | +| scaleway | 13 | 欧洲（法国） | +| cloudferro-sherlock | 12 | 欧洲（欧盟） | + +## 选择提供商 + +| 如果您需要... | 最佳提供商 | 原因 | +| ---------------- | ------------------------------ | ---------------------- | +| **最低价格** | deepseek, google | 输入低至 $0.14/M token | +| **最快推理** | groq, cerebras | 亚 100ms 延迟 | +| **最大上下文** | google, meta | 1M-10M token 上下文 | +| **最多模型** | nanogpt, aihubmix | 各 500+ 模型 | +| **最适合 Agent** | anthropic, openai | 工具调用 + 推理 | +| **开源权重** | meta, deepseek | 在自有硬件上运行 | +| **欧盟数据驻留** | ovhcloud, scaleway, cloudferro | 欧盟托管推理 | +| **中国访问** | siliconflow, ppio, stepfun | 中国节点 | + +## 相关文档 + +- [迁移指南](migration-guide.md) — 切换提供商的定价对比 +- [定价对比](pricing-comparison.md) — 跨提供商定价对比 +- [提供商概览](providers.md) — 全部 95 个提供商列表 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [聊天模型](chat-models.md) — 2,350 个支持工具调用的聊天模型 +- [智能体模型](agentic-models.md) — 1,080 个具备工具调用 + 推理能力的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/package.json b/package.json index 1305efc7..28ff6339 100644 --- a/package.json +++ b/package.json @@ -50,6 +50,7 @@ "open-weights", "openai", "prompt-caching", + "provider-comparison", "qwen", "reasoning-models", "structured-output", diff --git a/site/sitemap.xml b/site/sitemap.xml index 0a495b7e..a38c006d 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -120,6 +120,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/provider-comparison + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/providers weekly From 1cc20d59ae1fe39600aa24fb167e2be0f1f48ece Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:33:52 +0800 Subject: [PATCH 182/311] Add 2 new doc pages: small-models, large-context-models (EN + ZH) - small-models.md: 1,153 edge/on-device models across 63 providers - large-context-models.md: 2,195 128K+ context models across 78 providers - Both added to all index files: README, AGENTS.md, llms.txt, llms-full.txt, sitemap.xml, site/index.html, CITATION.cff, package.json - Bilingual (EN + ZH) with cross-links --- CITATION.cff | 2 + docs/large-context-models.md | 112 ++++++++++++++++++++++++++++++++ docs/small-models.md | 80 +++++++++++++++++++++++ docs/zh/large-context-models.md | 112 ++++++++++++++++++++++++++++++++ docs/zh/small-models.md | 80 +++++++++++++++++++++++ package.json | 4 ++ site/sitemap.xml | 10 +++ 7 files changed, 400 insertions(+) create mode 100644 docs/large-context-models.md create mode 100644 docs/small-models.md create mode 100644 docs/zh/large-context-models.md create mode 100644 docs/zh/small-models.md diff --git a/CITATION.cff b/CITATION.cff index 25b7b9b3..cc167665 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -42,6 +42,8 @@ keywords: - embedding-models - migration-guide - provider-comparison + - large-context-models + - small-models license: MIT version: 0.1.0 date-released: "2026-05-21" diff --git a/docs/large-context-models.md b/docs/large-context-models.md new file mode 100644 index 00000000..c45dc882 --- /dev/null +++ b/docs/large-context-models.md @@ -0,0 +1,112 @@ +# Large Context Models + +[中文](zh/large-context-models.md) + +AI models with **128K+ token context windows** — process entire codebases, long documents, and multi-hour conversations in a single request. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Large Context Matters + +Large context windows unlock capabilities impossible with smaller models: + +- **Full codebase analysis** — understand entire repositories in one prompt +- **Document processing** — analyze 100+ page PDFs without chunking +- **Multi-turn conversations** — maintain context across long chat sessions +- **Data analysis** — process large datasets in a single request +- **Legal/medical review** — review lengthy contracts and medical records +- **Content creation** — maintain consistency across long-form writing + +## Stats + +| Metric | Count | +| ---------------------------- | ----- | +| Large context models (128K+) | 2195 | +| 256K+ context | 861 | +| 1M+ context | 397 | +| Providers | 78 | +| Free large context models | 51 | +| With tool calling | 1637 | + +## Providers + +`302ai`, `ai21`, `aimlapi`, `aion`, `alibaba`, `amazon`, `amazon-bedrock`, `anthropic`, `arcee`, `auriko`, `baichuan`, `baidu`, `baseten`, `bytedance`, `cerebras`, `chutes`, `clarifai`, `cloudferro-sherlock`, `cloudflare`, `databricks`, `deepinfra`, `deepseek`, `digitalocean`, `dinference`, `evroc` and 53 more + +## Largest Context Windows + +Models with the biggest context windows available. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ------------------------------ | ---------- | ------- | --------- | ---------- | ------------ | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 🔧 📋 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | 🔧 | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 🔧 📋 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | 🔧 | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | 🔧 | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | 🔧 | +| x-ai--grok-4.1-fast | fastrouter | 2M | $0.2 | $0.5 | 🔧 | +| xai--grok-4-fast-reasoning | aimlapi | 2M | $0.52 | $1.3 | 🔧 | +| xai--grok-4-fast-non-reasoning | aimlapi | 2M | $0.52 | $1.3 | 🔧 | +| grok-4-20-multi-agent | venice | 2M | $1.42 | $2.83 | 🧠 📋 | +| grok-4-20 | venice | 2M | $1.42 | $2.83 | 🔧 🧠 📋 | + +## Cheapest 1M+ Context Models + +Best value models with 1M+ token context — for processing very long inputs. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| --------------------------------- | ------------- | ------- | --------- | ---------- | ------------ | +| gemini-1.5-flash-8b | deepinfra | 1M | $0.0375 | $0.15 | | +| gpt-5-nano | meganova | 1M | $0.04 | $0.32 | 🔧 | +| qwen--qwen3.5-flash-02-23 | openrouter | 1M | $0.065 | $0.26 | 🔧 🧠 📋 | +| google--gemini-2.0-flash-lite-001 | openrouter | 1M | $0.075 | $0.3 | 🔧 📋 | +| google--gemini-2.0-flash-lite-001 | fastrouter | 1M | $0.075 | $0.3 | 🔧 | +| gemini-1.5-flash | deepinfra | 1M | $0.075 | $0.3 | | +| gemini-2.0-flash-lite | google | 1M | $0.075 | $0.3 | 🔧 📋 | +| gemini-1.5-flash | google | 1M | $0.075 | $0.3 | 🔧 📋 | +| gemini-1.5-flash-8b | google | 1M | $0.075 | $0.3 | 🔧 📋 | +| gemini-2-0-flash-lite | google-vertex | 1M | $0.075 | $0.3 | 🔧 | + +## Free Large Context Models + +Free models with 128K+ context — zero-cost long document processing. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ---------------------------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| openrouter--owl-alpha | openrouter | 1M | Free | Free | 🔧 📋 | +| deepseek--deepseek-v4-flash--free | openrouter | 1M | Free | Free | 🔧 🧠 | +| google--lyria-3-clip-preview | openrouter | 1M | Free | Free | 📋 | +| google--lyria-3-pro-preview | openrouter | 1M | Free | Free | 📋 | +| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | 🔧 | +| nvidia--nemotron-3-super-120b-a12b--free | openrouter | 1M | Free | Free | 🔧 🧠 📋 | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| arcee-ai--trinity-large-thinking--free | openrouter | 262K | Free | Free | 🔧 🧠 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🔧 🧠 📋 | + +## Context Window Tiers + +| Tier | Context | Use Case | Example Models | +| -------- | ------- | ------------------------------ | ------------------------- | +| Standard | 128K | Long documents, code files | gpt-4.1, claude-sonnet-4 | +| Extended | 256K | Codebases, multi-file analysis | claude-opus-4, o3 | +| Ultra | 1M | Full repositories, books | gemini-2.5-flash, gpt-4.1 | +| Massive | 10M | Entire datasets, video | llama-4-scout | + +## Related Documentation + +- [Context Windows](context-windows.md) — detailed context window comparison +- [Chat Models](chat-models.md) — 2,350 models with tool calling +- [Code Models](code-models.md) — 189 code-focused models +- [Free AI Models](free-models.md) — 81 free models by capability +- [Model Selection Guide](model-selection.md) — decision framework +- [Migration Guide](migration-guide.md) — switching providers +- [Provider Comparison](provider-comparison.md) — top 30 providers + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/small-models.md b/docs/small-models.md new file mode 100644 index 00000000..ec3cc772 --- /dev/null +++ b/docs/small-models.md @@ -0,0 +1,80 @@ +# Small & Edge Models + +[中文](zh/small-models.md) + +AI models designed for **edge deployment, on-device inference, and resource-constrained environments** — models under ~16B parameters that can run on consumer hardware, mobile devices, and embedded systems. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Small Models Matter + +Small models enable AI where cloud connectivity is limited or latency is critical: + +- **On-device inference** — run AI without internet, on phones and laptops +- **Edge computing** — deploy in IoT devices, robotics, and vehicles +- **Low latency** — sub-100ms response times for real-time applications +- **Cost efficiency** — cheaper to run, especially at scale +- **Privacy** — data never leaves the device +- **Offline capability** — AI that works without connectivity + +## Stats + +| Metric | Count | +| ------------------------ | ----- | +| Small/edge models | 1153 | +| Providers | 63 | +| Free small models | 30 | +| Open-weight small models | 272 | +| With tool calling | 434 | +| With reasoning | 250 | + +## Providers + +`302ai`, `aihubmix`, `aimlapi`, `aion`, `alibaba`, `amazon-bedrock`, `auriko`, `baichuan`, `berget`, `bytedance`, `cerebras`, `chutes`, `clarifai`, `cloudferro-sherlock`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `evroc`, `fastrouter`, `fireworks`, `friendli`, `gmicloud`, `google`, `google-vertex`, `groq`, `hpc-ai`, `hyperbolic`, `inferencenet`, `jiekou`, `klusterai`, `llmgateway`, `martian`, `meganova`, `meta`, `microsoft`, `mistral`, `mixlayer`, `moonshotai`, `morph`, `nanogpt`, `nebius`, `neuralwatt`, `nousresearch`, `novitaai`, `openrouter`, `orcarouter`, `ovhcloud`, `ppio`, `privatemode`, `requesty`, `sambanova`, `scaleway`, `siliconflow`, `siliconflow-cn`, `submodel`, `tencent`, `textsynth`, `togetherai`, `venice`, `vultr`, `wafer` + +## Free Small Models + +Free small models — zero-cost edge AI. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ---------------------------------------------------- | ---------- | ------- | --------- | ---------- | ------------ | +| nvidia--nemotron-3-super-120b-a12b--free | openrouter | 1M | Free | Free | 🔧 🧠 📋 | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🔧 🧠 📋 | +| gemma-4-31b-it | auriko | 262K | Free | Free | 🔧 🧠 📋 | +| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | 256K | Free | Free | 🔧 🧠 | +| gemma-3-4b-it | google | 131K | Free | Free | | +| gemma-3-12b-it | google | 131K | Free | Free | | +| gemma-3-27b-it | google | 131K | Free | Free | | +| gemma-3n-E2B-it | google | 131K | Free | Free | | + +## Cheapest Small Models + +Best value small models for production. + +| Model | Provider | Context | Input $/M | Output $/M | Capabilities | +| ----------------------------------------------- | -------- | ------- | --------- | ---------- | ------------ | +| llama3-groq-8b-8192-tool-use-preview | aihubmix | 0 | $9.5e-05 | $9.5e-05 | | +| mistralai--mistral-7b-instruct--free | aihubmix | 0 | $0.001 | $0.001 | | +| deepseek-ai--deepseek-r1-distill-llama-8b | aihubmix | 0 | $0.005 | $0.005 | | +| deepseek-ai--deepseek-r1-distill-qwen-7b | aihubmix | 0 | $0.005 | $0.005 | | +| deepseek-ai--deepseek-r1-distill-qwen-1.5b | aihubmix | 0 | $0.005 | $0.005 | | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🔧 📋 | +| google--gemma-2-9b-it--free | aihubmix | 0 | $0.01 | $0.01 | | +| meta-llama--llama-3.2-3b-instruct--free | aihubmix | 0 | $0.01 | $0.01 | | +| meta-llama--llama-3.2-11b-vision-instruct--free | aihubmix | 0 | $0.01 | $0.01 | | +| meta-llama--llama-3.1-8b-instruct--free | aihubmix | 0 | $0.01 | $0.01 | | + +## Related Documentation + +- [Open Weights](open-weights.md) — 527 open-weight models +- [Free AI Models](free-models.md) — 81 free models by capability +- [Chat Models](chat-models.md) — 2,350 models with tool calling +- [Code Models](code-models.md) — 189 code-focused models +- [Model Selection Guide](model-selection.md) — decision framework +- [Provider Comparison](provider-comparison.md) — top 30 providers + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/large-context-models.md b/docs/zh/large-context-models.md new file mode 100644 index 00000000..5267e125 --- /dev/null +++ b/docs/zh/large-context-models.md @@ -0,0 +1,112 @@ +# 大上下文模型 + +[English](../large-context-models.md) + +具有 **128K+ token 上下文窗口**的 AI 模型 — 在单个请求中处理整个代码库、长文档和多小时对话。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么大上下文很重要 + +大上下文窗口解锁了小模型无法实现的能力： + +- **完整代码库分析** — 在一个提示中理解整个仓库 +- **文档处理** — 无需分块即可分析 100+ 页 PDF +- **多轮对话** — 在长聊天会话中保持上下文 +- **数据分析** — 在单个请求中处理大型数据集 +- **法律/医疗审查** — 审查冗长的合同和医疗记录 +- **内容创作** — 在长篇写作中保持一致性 + +## 统计 + +| 指标 | 数量 | +| -------------------- | ---- | +| 大上下文模型 (128K+) | 2195 | +| 256K+ 上下文 | 861 | +| 1M+ 上下文 | 397 | +| 提供商 | 78 | +| 免费大上下文模型 | 51 | +| 带工具调用 | 1637 | + +## 提供商 + +`302ai`、`ai21`、`aimlapi`、`aion`、`alibaba`、`amazon`、`amazon-bedrock`、`anthropic`、`arcee`、`auriko`、`baichuan`、`baidu`、`baseten`、`bytedance`、`cerebras`、`chutes`、`clarifai`、`cloudferro-sherlock`、`cloudflare`、`databricks`、`deepinfra`、`deepseek`、`digitalocean`、`dinference`、`evroc` 等 53 个 + +## 最大上下文窗口 + +可用上下文窗口最大的模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ------------------------------ | ---------- | ------ | -------- | -------- | -------- | +| meta-llama--llama-4-scout | openrouter | 10M | $0.08 | $0.3 | 🔧 📋 | +| meta-llama-4-scout | meta | 10M | $0.17 | $0.66 | 🔧 | +| gemini-1.5-pro | google | 2M | $1.25 | $5 | 🔧 📋 | +| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| gpt-4o | jiekou | 2M | $1.9 | $5.7 | 🔧 | +| grok-4.20-0309-non-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| grok-4.20-0309-reasoning | jiekou | 2M | $1.9 | $5.7 | 🔧 | +| grok-4-1-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| grok-4-fast-reasoning | jiekou | 2M | $0.19 | $0.475 | 🔧 | +| x-ai--grok-4-fast | fastrouter | 2M | $0.2 | $0.5 | 🔧 | +| x-ai--grok-4.1-fast | fastrouter | 2M | $0.2 | $0.5 | 🔧 | +| xai--grok-4-fast-reasoning | aimlapi | 2M | $0.52 | $1.3 | 🔧 | +| xai--grok-4-fast-non-reasoning | aimlapi | 2M | $0.52 | $1.3 | 🔧 | +| grok-4-20-multi-agent | venice | 2M | $1.42 | $2.83 | 🧠 📋 | +| grok-4-20 | venice | 2M | $1.42 | $2.83 | 🔧 🧠 📋 | + +## 最便宜的 1M+ 上下文模型 + +1M+ token 上下文的最佳性价比模型 — 处理超长输入。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| --------------------------------- | ------------- | ------ | -------- | -------- | -------- | +| gemini-1.5-flash-8b | deepinfra | 1M | $0.0375 | $0.15 | | +| gpt-5-nano | meganova | 1M | $0.04 | $0.32 | 🔧 | +| qwen--qwen3.5-flash-02-23 | openrouter | 1M | $0.065 | $0.26 | 🔧 🧠 📋 | +| google--gemini-2.0-flash-lite-001 | openrouter | 1M | $0.075 | $0.3 | 🔧 📋 | +| google--gemini-2.0-flash-lite-001 | fastrouter | 1M | $0.075 | $0.3 | 🔧 | +| gemini-1.5-flash | deepinfra | 1M | $0.075 | $0.3 | | +| gemini-2.0-flash-lite | google | 1M | $0.075 | $0.3 | 🔧 📋 | +| gemini-1.5-flash | google | 1M | $0.075 | $0.3 | 🔧 📋 | +| gemini-1.5-flash-8b | google | 1M | $0.075 | $0.3 | 🔧 📋 | +| gemini-2-0-flash-lite | google-vertex | 1M | $0.075 | $0.3 | 🔧 | + +## 免费大上下文模型 + +128K+ 上下文的免费模型 — 零成本长文档处理。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---------------------------------------- | ---------- | ------ | -------- | -------- | -------- | +| openrouter--owl-alpha | openrouter | 1M | Free | Free | 🔧 📋 | +| deepseek--deepseek-v4-flash--free | openrouter | 1M | Free | Free | 🔧 🧠 | +| google--lyria-3-clip-preview | openrouter | 1M | Free | Free | 📋 | +| google--lyria-3-pro-preview | openrouter | 1M | Free | Free | 📋 | +| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | 🔧 | +| nvidia--nemotron-3-super-120b-a12b--free | openrouter | 1M | Free | Free | 🔧 🧠 📋 | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| arcee-ai--trinity-large-thinking--free | openrouter | 262K | Free | Free | 🔧 🧠 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🔧 🧠 📋 | + +## 上下文窗口层级 + +| 层级 | 上下文 | 用例 | 示例模型 | +| ---- | ------ | ------------------ | ------------------------- | +| 标准 | 128K | 长文档、代码文件 | gpt-4.1, claude-sonnet-4 | +| 扩展 | 256K | 代码库、多文件分析 | claude-opus-4, o3 | +| 超大 | 1M | 完整仓库、书籍 | gemini-2.5-flash, gpt-4.1 | +| 巨型 | 10M | 整个数据集、视频 | llama-4-scout | + +## 相关文档 + +- [上下文窗口](context-windows.md) — 详细的上下文窗口对比 +- [聊天模型](chat-models.md) — 2,350 个支持工具调用的聊天模型 +- [代码模型](code-models.md) — 189 个代码模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [迁移指南](migration-guide.md) — 切换提供商 +- [提供商对比](provider-comparison.md) — 前 30 个提供商 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/small-models.md b/docs/zh/small-models.md new file mode 100644 index 00000000..a2373da3 --- /dev/null +++ b/docs/zh/small-models.md @@ -0,0 +1,80 @@ +# 小型与边缘模型 + +[English](../small-models.md) + +专为**边缘部署、设备端推理和资源受限环境**设计的 AI 模型 — 参数量在 ~16B 以下的模型，可在消费级硬件、移动设备和嵌入式系统上运行。 + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么小型模型很重要 + +小型模型使 AI 在云连接有限或延迟关键的场景中成为可能： + +- **设备端推理** — 在手机和笔记本电脑上无需互联网运行 AI +- **边缘计算** — 部署在 IoT 设备、机器人和车辆中 +- **低延迟** — 亚 100ms 响应时间，适用于实时应用 +- **成本效益** — 运行成本更低，尤其是在大规模场景 +- **隐私** — 数据不离开设备 +- **离线能力** — 无需网络连接即可使用的 AI + +## 统计 + +| 指标 | 数量 | +| ---------------- | ---- | +| 小型/边缘模型 | 1153 | +| 提供商 | 63 | +| 免费小型模型 | 30 | +| 开源权重小型模型 | 272 | +| 带工具调用 | 434 | +| 带推理能力 | 250 | + +## 提供商 + +`302ai`、`aihubmix`、`aimlapi`、`aion`、`alibaba`、`amazon-bedrock`、`auriko`、`baichuan`、`berget`、`bytedance`、`cerebras`、`chutes`、`clarifai`、`cloudferro-sherlock`、`cloudflare`、`cortecs`、`databricks`、`deepinfra`、`digitalocean`、`evroc`、`fastrouter`、`fireworks`、`friendli`、`gmicloud`、`google`、`google-vertex`、`groq`、`hpc-ai`、`hyperbolic`、`inferencenet`、`jiekou`、`klusterai`、`llmgateway`、`martian`、`meganova`、`meta`、`microsoft`、`mistral`、`mixlayer`、`moonshotai`、`morph`、`nanogpt`、`nebius`、`neuralwatt`、`nousresearch`、`novitaai`、`openrouter`、`orcarouter`、`ovhcloud`、`ppio`、`privatemode`、`requesty`、`sambanova`、`scaleway`、`siliconflow`、`siliconflow-cn`、`submodel`、`tencent`、`textsynth`、`togetherai`、`venice`、`vultr`、`wafer` + +## 免费小型模型 + +免费小型模型 — 零成本边缘 AI。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ---------------------------------------------------- | ---------- | ------ | -------- | -------- | -------- | +| nvidia--nemotron-3-super-120b-a12b--free | openrouter | 1M | Free | Free | 🔧 🧠 📋 | +| google--gemma-4-26b-a4b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| google--gemma-4-31b-it--free | openrouter | 262K | Free | Free | 🔧 🧠 📋 | +| gemma-4-26b-a4b-it | auriko | 262K | Free | Free | 🔧 🧠 📋 | +| gemma-4-31b-it | auriko | 262K | Free | Free | 🔧 🧠 📋 | +| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | 256K | Free | Free | 🔧 🧠 | +| gemma-3-4b-it | google | 131K | Free | Free | | +| gemma-3-12b-it | google | 131K | Free | Free | | +| gemma-3-27b-it | google | 131K | Free | Free | | +| gemma-3n-E2B-it | google | 131K | Free | Free | | + +## 最便宜小型模型 + +生产环境小型应用的最佳性价比模型。 + +| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | +| ----------------------------------------------- | -------- | ------ | -------- | -------- | ----- | +| llama3-groq-8b-8192-tool-use-preview | aihubmix | 0 | $9.5e-05 | $9.5e-05 | | +| mistralai--mistral-7b-instruct--free | aihubmix | 0 | $0.001 | $0.001 | | +| deepseek-ai--deepseek-r1-distill-llama-8b | aihubmix | 0 | $0.005 | $0.005 | | +| deepseek-ai--deepseek-r1-distill-qwen-7b | aihubmix | 0 | $0.005 | $0.005 | | +| deepseek-ai--deepseek-r1-distill-qwen-1.5b | aihubmix | 0 | $0.005 | $0.005 | | +| ernie-4.5-0.3b | aihubmix | 0 | $0.0068 | $0.0272 | 🔧 📋 | +| google--gemma-2-9b-it--free | aihubmix | 0 | $0.01 | $0.01 | | +| meta-llama--llama-3.2-3b-instruct--free | aihubmix | 0 | $0.01 | $0.01 | | +| meta-llama--llama-3.2-11b-vision-instruct--free | aihubmix | 0 | $0.01 | $0.01 | | +| meta-llama--llama-3.1-8b-instruct--free | aihubmix | 0 | $0.01 | $0.01 | | + +## 相关文档 + +- [开源权重](open-weights.md) — 527 个开源权重模型 +- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 +- [聊天模型](chat-models.md) — 2,350 个支持工具调用的聊天模型 +- [代码模型](code-models.md) — 189 个代码模型 +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [提供商对比](provider-comparison.md) — 前 30 个提供商 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/package.json b/package.json index 28ff6339..448a35de 100644 --- a/package.json +++ b/package.json @@ -21,6 +21,7 @@ "code-models", "context-window", "deepseek", + "edge-ai", "embedding-models", "free-models", "function-calling", @@ -31,6 +32,7 @@ "image-generation", "inference-api", "json-mode", + "large-context-models", "large-language-model", "llama", "llm", @@ -47,12 +49,14 @@ "model-pricing", "model-selection", "multimodal-models", + "on-device-ai", "open-weights", "openai", "prompt-caching", "provider-comparison", "qwen", "reasoning-models", + "small-models", "structured-output", "tool-calling", "vision-models", diff --git a/site/sitemap.xml b/site/sitemap.xml index a38c006d..51b03192 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -80,6 +80,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/large-context-models + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/lessons-learned weekly @@ -140,6 +145,11 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/docs/small-models + weekly + 0.7 + https://i-need-token.github.io/ai-models/docs/structured-output weekly From 0e0abab134675fbc6e2a13f8e71f52ed080d5508 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:34:21 +0800 Subject: [PATCH 183/311] Add cross-links to small-models and large-context-models from 10 existing docs --- docs/model-selection.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/docs/model-selection.md b/docs/model-selection.md index 9832d4dd..d415dd54 100644 --- a/docs/model-selection.md +++ b/docs/model-selection.md @@ -125,6 +125,9 @@ Models with the largest context windows — for long documents, multi-turn conve - [Chat Models](chat-models.md) — 2,350 models with tool calling for chat applications - [Multimodal Models](multimodal-models.md) — 1,519 models with image/audio/video input +- [Large Context Models](large-context-models.md) — 2,195 models with 128K+ context +- [Small & Edge Models](small-models.md) — 1,153 models for on-device inference + --- Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. From 416225c1d0e2cd9af5999153460a16cdf8fd95b8 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:38:35 +0800 Subject: [PATCH 184/311] =?UTF-8?q?Fix=20Model=20Picks=20table=20=E2=80=94?= =?UTF-8?q?=20restore=20correct=20pricing=20values=20after=20bash=20expans?= =?UTF-8?q?ion=20corruption?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) diff --git a/README.md b/README.md index 12579c09..eccdfb44 100644 --- a/README.md +++ b/README.md @@ -95,18 +95,18 @@ Machine-readable YAML catalog of every major AI model provider and their models > Curated recommendations for common use cases — from [4,587 models](docs/model-comparison.md) across 95 providers -| Use Case | Model | Why | Input $/M | Context | -| --------------------- | ---------------- | --------------------------------------- | ------------ | ------- | -| **Coding** | gpt-4.1 | Best code generation + 1M context | | 1M | -| **Coding (cheap)** | gpt-4.1-nano | 20× cheaper, great for autocomplete | /bin/bash.10 | 1M | -| **Reasoning** | o4-mini | Best cost-effective reasoning | .10 | 200K | -| **Reasoning (power)** | claude-opus-4 | Deepest reasoning for hard problems | 5 | 200K | -| **Agents** | claude-sonnet-4 | Best tool use + reasoning balance | | 200K | -| **Agents (cheap)** | gemini-2.5-flash | Fastest agent loop under | /bin/bash.15 | 1M | -| **Vision** | gemini-2.5-pro | Best multimodal understanding | .25 | 1M | -| **Free** | llama-4-scout | 10M context, open weights, free on Groq | /bin/bash | 10M | -| **Open weights** | deepseek-r1 | Best open reasoning model | /bin/bash.55 | 128K | -| **Large context** | gemini-2.5-flash | 1M context at lowest price | /bin/bash.15 | 1M | +| Use Case | Model | Why | Input $/M | Context | +| --------------------- | ---------------- | --------------------------------------- | --------- | ------- | +| **Coding** | gpt-4.1 | Best code generation + 1M context | $2 | 1M | +| **Coding (cheap)** | gpt-4.1-nano | 20x cheaper, great for autocomplete | $0.10 | 1M | +| **Reasoning** | o4-mini | Best cost-effective reasoning | $1.10 | 200K | +| **Reasoning (power)** | claude-opus-4 | Deepest reasoning for hard problems | $15 | 200K | +| **Agents** | claude-sonnet-4 | Best tool use + reasoning balance | $3 | 200K | +| **Agents (cheap)** | gemini-2.5-flash | Fastest agent loop under $1 | $0.15 | 1M | +| **Vision** | gemini-2.5-pro | Best multimodal understanding | $1.25 | 1M | +| **Free** | llama-4-scout | 10M context, open weights, free on Groq | $0 | 10M | +| **Open weights** | deepseek-r1 | Best open reasoning model | $0.55 | 128K | +| **Large context** | gemini-2.5-flash | 1M context at lowest price | $0.15 | 1M | ## Use Cases From 99613f64e423236617e3c20c03b8daa2aa7212f0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:47:21 +0800 Subject: [PATCH 185/311] =?UTF-8?q?Fix=20hero=20curl=20command=20and=20dow?= =?UTF-8?q?nload=20section=20=E2=80=94=20use=20GitHub=20Releases=20URL=20i?= =?UTF-8?q?nstead=20of=20broken=20CDN?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Hero 'Try it now' curl now uses releases/latest/download/models.json - Renamed 'CDN Access' to 'Direct Download' with GitHub Releases as primary - Added CSV download link from releases - CDN mentioned as future option after npm publish - All example code updated to use working URLs --- README.md | 39 ++++++++++++++++++++++++--------------- 1 file changed, 24 insertions(+), 15 deletions(-) diff --git a/README.md b/README.md index eccdfb44..6a8b1052 100644 --- a/README.md +++ b/README.md @@ -21,12 +21,12 @@ Machine-readable YAML catalog of every major AI model provider and their models — pricing, context windows, modalities, capabilities, and more. All data sourced from first-party APIs and official documentation, never third-party aggregators. -**[Quick start →](docs/quick-start.md)** · **[Choose a model →](docs/model-selection.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[🔍 Search →](https://i-need-token.github.io/ai-models/)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[CDN →](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json)** +**[Quick start →](docs/quick-start.md)** · **[Choose a model →](docs/model-selection.md)** · **[Compare pricing →](docs/pricing-comparison.md)** · **[🔍 Search →](https://i-need-token.github.io/ai-models/)** · **[Download CSV →](https://github.com/i-need-token/ai-models/releases/latest/download/models.csv)** · **[JSON →](https://github.com/i-need-token/ai-models/releases/latest/download/models.json)** > 💡 **Try it now** — fetch model data in one command: > > ```bash -> curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d)} models across {len(set(m["provider"] for m in d))} providers')" +> curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d)} models across {len(set(m["provider"] for m in d))} providers')" > ``` ## Why This Catalog? @@ -351,30 +351,39 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv ``` -### CDN Access (no install) +### Direct Download (no install) -The compiled JSON is available via [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) — no download or install needed: +The compiled JSON is always available from [GitHub Releases](https://github.com/i-need-token/ai-models/releases): + +```bash +# Download latest models.json +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.json + +# Download latest models.csv (for Excel/Sheets) +curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/models.csv +``` ```html ``` -```bash -# Direct curl (always up-to-date) -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' -``` - ```python # Python — no pip install needed import urllib.request, json -catalog = json.loads(urllib.request.urlopen("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json").read()) -print(len(catalog["models"])) # 4587 +catalog = json.loads(urllib.request.urlopen("https://github.com/i-need-token/ai-models/releases/latest/download/models.json").read()) +print(len(catalog)) # 4587 +``` + +Once the npm package is published, you can also use [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models): + +```bash +curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '. | length' ``` See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. From 38bc17ee8c3325108435ada84876516c302741aa Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 18:49:35 +0800 Subject: [PATCH 186/311] Improve README: add Project Timeline, fix Roadmap with completed items - Add Project Timeline table showing milestones and future plans - Merge duplicate Roadmap sections into one - Mark completed roadmap items (embedding-models, provider-comparison, large-context-models, small-models, migration-guide) with checkmarks - Add new roadmap items: historical pricing, REST API --- README.md | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 6a8b1052..219bd678 100644 --- a/README.md +++ b/README.md @@ -610,7 +610,16 @@ Want to contribute? Check out [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines. - 📋 [Changelog](CHANGELOG.md) — what's new in each release - 🔒 [Security Policy](SECURITY.md) — reporting vulnerabilities -## Star History +## Project Timeline + +| Date | Milestone | +| ------- | ----------------------------------------------------------- | +| 2026-05 | 🚀 Public launch — 4,587 models, 95 providers, 66 docs | +| 2026-05 | 📊 Interactive catalog live at GitHub Pages | +| 2026-05 | 📦 npm package, CSV export, GitHub Action | +| 2026-05 | 🌐 Bilingual docs — 33 EN + 33 ZH pages | +| 2026-05 | 🤖 1,080 agentic models, 2,350 tool-calling models | +| Future | 📈 More providers, REST API, historical pricing, benchmarks | [![Star History Chart](https://api.star-history.com/svg?repos=i-need-token/ai-models&type=Date)](https://star-history.com/#i-need-token/ai-models&Date) From a36c6d458d9647dc8dd45e1d10d3c260a8385ce0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:02:55 +0800 Subject: [PATCH 187/311] Add data source attribution to all 66 docs (EN + ZH) - Added 'Data sourced from AI Models Catalog' footer to all docs that were missing it (15 EN + 14 ZH) - Consistent backlink to repo from every documentation page - Improves SEO and drives traffic back to the main repository --- docs/api.md | 4 ++++ docs/code-examples.md | 4 ++++ docs/data-schema.md | 4 ++++ docs/free-models.md | 4 ++++ docs/glossary.md | 4 ++++ docs/image-generation.md | 4 ++++ docs/modality-matrix.md | 4 ++++ docs/model-comparison.md | 4 ++++ docs/open-weights.md | 4 ++++ docs/pricing-comparison.md | 4 ++++ docs/providers.md | 4 ++++ docs/reasoning-models.md | 4 ++++ docs/structured-output.md | 4 ++++ docs/tool-calling.md | 4 ++++ docs/vision-models.md | 4 ++++ docs/zh/api.md | 4 ++++ docs/zh/context-windows.md | 4 ++++ docs/zh/data-schema.md | 4 ++++ docs/zh/free-models.md | 4 ++++ docs/zh/image-generation.md | 4 ++++ docs/zh/modality-matrix.md | 4 ++++ docs/zh/model-comparison.md | 4 ++++ docs/zh/open-weights.md | 4 ++++ docs/zh/pricing-comparison.md | 4 ++++ docs/zh/providers.md | 4 ++++ docs/zh/reasoning-models.md | 4 ++++ docs/zh/structured-output.md | 4 ++++ docs/zh/tool-calling.md | 4 ++++ docs/zh/vision-models.md | 4 ++++ 29 files changed, 116 insertions(+) diff --git a/docs/api.md b/docs/api.md index b3fbc68a..fce3746f 100644 --- a/docs/api.md +++ b/docs/api.md @@ -216,3 +216,7 @@ npx tsx scripts/sync.ts # all providers - [Data Schema](data-schema.md) — complete YAML schema reference - [FAQ](faq.md) — common questions - [Model Selection Guide](model-selection.md) — decision framework + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/code-examples.md b/docs/code-examples.md index 7bcdb7bc..6ab34112 100644 --- a/docs/code-examples.md +++ b/docs/code-examples.md @@ -365,3 +365,7 @@ const results = findModels({ - [Data Schema](data-schema.md) — complete YAML schema reference - [FAQ](faq.md) — common questions - [Glossary](glossary.md) — key terms and definitions + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/data-schema.md b/docs/data-schema.md index 7d791fa6..0fb1dc0d 100644 --- a/docs/data-schema.md +++ b/docs/data-schema.md @@ -215,3 +215,7 @@ The validation uses `ModelSchema` from [`types/schemas.ts`](../types/schemas.ts) - [Code Examples](code-examples.md) — practical code examples - [Design Principles](lessons-learned.md) — lessons learned - [FAQ](faq.md) — common questions + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/free-models.md b/docs/free-models.md index 370b7421..ea625455 100644 --- a/docs/free-models.md +++ b/docs/free-models.md @@ -117,3 +117,7 @@ Groq offers free tier for some models with rate limits: - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching - [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/glossary.md b/docs/glossary.md index a27ae746..e830f4f0 100644 --- a/docs/glossary.md +++ b/docs/glossary.md @@ -77,3 +77,7 @@ See [Data Schema Reference](data-schema.md) for the complete YAML field specific - [Quick Start](quick-start.md) — find the right model in 30 seconds - [Model Comparison](model-comparison.md) — compare models - [Modality Matrix](modality-matrix.md) — all modalities at a glance + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/image-generation.md b/docs/image-generation.md index 8d244a35..a50cbbfc 100644 --- a/docs/image-generation.md +++ b/docs/image-generation.md @@ -64,3 +64,7 @@ - [Modality Matrix](modality-matrix.md) — all modalities at a glance - [Model Selection Guide](model-selection.md) — decision framework - [Free AI Models](free-models.md) — 81 free models + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/modality-matrix.md b/docs/modality-matrix.md index d40d40e0..9fc5a5fb 100644 --- a/docs/modality-matrix.md +++ b/docs/modality-matrix.md @@ -100,3 +100,7 @@ Models that accept text + at least 2 additional input modalities: - [Video Models](video-models.md) — 167 video input/output models - [Image Generation](image-generation.md) — 28 image generation models - [Model Selection Guide](model-selection.md) — decision framework + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/model-comparison.md b/docs/model-comparison.md index 9813c168..b041ed7a 100644 --- a/docs/model-comparison.md +++ b/docs/model-comparison.md @@ -103,3 +103,7 @@ Models with publicly available weights for self-hosting. - [Free AI Models](free-models.md) — 81 free models - [Open-Weight Models](open-weights.md) — 527 models you can run yourself - [Context Window Comparison](context-windows.md) — largest context windows + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/open-weights.md b/docs/open-weights.md index 20d3f822..f3f0ba3d 100644 --- a/docs/open-weights.md +++ b/docs/open-weights.md @@ -112,3 +112,7 @@ Lowest per-token pricing for open-weight inference: - [Provider Overview](providers.md) — all 95 providers organized by type - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/pricing-comparison.md b/docs/pricing-comparison.md index 6004203a..b1c26001 100644 --- a/docs/pricing-comparison.md +++ b/docs/pricing-comparison.md @@ -127,3 +127,7 @@ The absolute cheapest per-token models across all providers. - [Free AI Models](free-models.md) — 81 free models - [Context Window Comparison](context-windows.md) — largest context windows - [Provider Overview](providers.md) — all 95 providers + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/providers.md b/docs/providers.md index eafba4ce..526b1379 100644 --- a/docs/providers.md +++ b/docs/providers.md @@ -166,3 +166,7 @@ Providers with EUR pricing, serving the European market. - [Open-Weight Models](open-weights.md) — 527 models you can run yourself - [Free AI Models](free-models.md) — 81 free models - [Data Schema](data-schema.md) — complete YAML schema + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/reasoning-models.md b/docs/reasoning-models.md index f3d42836..1551d173 100644 --- a/docs/reasoning-models.md +++ b/docs/reasoning-models.md @@ -95,3 +95,7 @@ Models that can reason about images — ideal for visual analysis: - [Structured Output](structured-output.md) — 829 JSON-mode models - [Free AI Models](free-models.md) — 81 free models, some with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/structured-output.md b/docs/structured-output.md index 737416b9..501b4e6e 100644 --- a/docs/structured-output.md +++ b/docs/structured-output.md @@ -77,3 +77,7 @@ For AI agents that need to return structured data, call tools, and reason: - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning - [Free AI Models](free-models.md) — 81 free models, some with structured output - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/tool-calling.md b/docs/tool-calling.md index c2a15887..e3e6be39 100644 --- a/docs/tool-calling.md +++ b/docs/tool-calling.md @@ -94,3 +94,7 @@ The "holy trinity" for advanced AI agents — tool calling, reasoning, and visio - [Structured Output](structured-output.md) — 829 JSON-mode models - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/vision-models.md b/docs/vision-models.md index d94af2ab..852d99b6 100644 --- a/docs/vision-models.md +++ b/docs/vision-models.md @@ -93,3 +93,7 @@ The most capable vision models — can see, reason, and act: - [Video Models](video-models.md) — 167 video input/output models - [Modality Matrix](modality-matrix.md) — all modalities at a glance - [Free AI Models](free-models.md) — 81 free models, some with vision + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/api.md b/docs/zh/api.md index e48cd8cf..5b133551 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -216,3 +216,7 @@ npx tsx scripts/sync.ts # 所有提供商 - [数据模式](data-schema.md) — 完整 YAML 模式参考 - [常见问题](faq.md) — 常见问题 - [模型选择指南](model-selection.md) — 决策框架 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/context-windows.md b/docs/zh/context-windows.md index e49cdf89..7b0877a2 100644 --- a/docs/zh/context-windows.md +++ b/docs/zh/context-windows.md @@ -71,3 +71,7 @@ - [免费 AI 模型](free-models.md) — 81 个免费模型按上下文窗口分类 - [视觉模型](vision-models.md) — 1,487 个视觉模型含上下文信息 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/data-schema.md b/docs/zh/data-schema.md index 8e5ed47f..ded0ab32 100644 --- a/docs/zh/data-schema.md +++ b/docs/zh/data-schema.md @@ -215,3 +215,7 @@ npx tsx scripts/validate.ts openai - [代码示例](code-examples.md) — 实用代码示例 - [设计原则](lessons-learned.md) — 经验教训 - [常见问题](faq.md) — 常见问题 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/free-models.md b/docs/zh/free-models.md index 7ae9a455..5224236f 100644 --- a/docs/zh/free-models.md +++ b/docs/zh/free-models.md @@ -117,3 +117,7 @@ Groq 为部分模型提供免费层： - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 - [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/image-generation.md b/docs/zh/image-generation.md index cd66d0f7..23d9374d 100644 --- a/docs/zh/image-generation.md +++ b/docs/zh/image-generation.md @@ -64,3 +64,7 @@ - [模态矩阵](modality-matrix.md) — 所有模态一览 - [模型选择指南](model-selection.md) — 决策框架 - [免费 AI 模型](free-models.md) — 81 个免费模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/modality-matrix.md b/docs/zh/modality-matrix.md index 11864322..5cfc4d31 100644 --- a/docs/zh/modality-matrix.md +++ b/docs/zh/modality-matrix.md @@ -100,3 +100,7 @@ - [视频模型](video-models.md) — 167 个视频输入/输出模型 - [图像生成](image-generation.md) — 28 个图像生成模型 - [模型选择指南](model-selection.md) — 决策框架 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/model-comparison.md b/docs/zh/model-comparison.md index 6c31659e..418aa9cf 100644 --- a/docs/zh/model-comparison.md +++ b/docs/zh/model-comparison.md @@ -103,3 +103,7 @@ - [免费 AI 模型](free-models.md) — 81 个免费模型 - [开源权重模型](open-weights.md) — 527 个可自行运行的模型 - [上下文窗口对比](context-windows.md) — 最大上下文窗口 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/open-weights.md b/docs/zh/open-weights.md index 70cce2e3..7e7a6d83 100644 --- a/docs/zh/open-weights.md +++ b/docs/zh/open-weights.md @@ -112,3 +112,7 @@ - [提供商概览](providers.md) — 95 个提供商按类型分类 - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/pricing-comparison.md b/docs/zh/pricing-comparison.md index a86991d3..57e24849 100644 --- a/docs/zh/pricing-comparison.md +++ b/docs/zh/pricing-comparison.md @@ -127,3 +127,7 @@ - [免费 AI 模型](free-models.md) — 81 个免费模型 - [上下文窗口对比](context-windows.md) — 最大上下文窗口 - [提供商概览](providers.md) — 95 个提供商 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/providers.md b/docs/zh/providers.md index eee37806..5177170f 100644 --- a/docs/zh/providers.md +++ b/docs/zh/providers.md @@ -166,3 +166,7 @@ - [开源权重模型](open-weights.md) — 527 个可自行运行的模型 - [免费 AI 模型](free-models.md) — 81 个免费模型 - [数据模式](data-schema.md) — 完整 YAML 模式 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/reasoning-models.md b/docs/zh/reasoning-models.md index 180b787b..6cc9d34f 100644 --- a/docs/zh/reasoning-models.md +++ b/docs/zh/reasoning-models.md @@ -95,3 +95,7 @@ - [结构化输出](structured-output.md) — 829 个 JSON 模式模型 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持推理 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/structured-output.md b/docs/zh/structured-output.md index 2e996435..c834c447 100644 --- a/docs/zh/structured-output.md +++ b/docs/zh/structured-output.md @@ -77,3 +77,7 @@ - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持结构化输出 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/tool-calling.md b/docs/zh/tool-calling.md index 360f38bf..747b6123 100644 --- a/docs/zh/tool-calling.md +++ b/docs/zh/tool-calling.md @@ -94,3 +94,7 @@ - [结构化输出](structured-output.md) — 829 个 JSON 模式模型 - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/vision-models.md b/docs/zh/vision-models.md index ed394e7e..c9822b01 100644 --- a/docs/zh/vision-models.md +++ b/docs/zh/vision-models.md @@ -93,3 +93,7 @@ - [视频模型](video-models.md) — 167 个视频输入/输出模型 - [模态矩阵](modality-matrix.md) — 所有模态一览 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持视觉 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 From 4d9508a944fce19dd090699c854cfff6ac803d9b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:08:24 +0800 Subject: [PATCH 188/311] Add 4 missing docs to all index files (README, AGENTS.md, llms.txt, llms-full.txt) - Added large-context-models, small-models, provider-comparison to README (both EN and ZH documentation tables) - Added large-context-models, small-models, provider-comparison to AGENTS.md - Added large-context-models, small-models, provider-comparison, migration-guide to llms.txt - Added large-context-models, small-models, provider-comparison, migration-guide to llms-full.txt - All 33 docs now covered in all 4 index files --- AGENTS.md | 3 +++ README.md | 54 ++++++++++++++++++++++++++++----------------------- llms-full.txt | 4 ++++ llms.txt | 4 ++++ 4 files changed, 41 insertions(+), 24 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index 84640015..76f86854 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -11,6 +11,9 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/data-acquisition.md`](docs/data-acquisition.md) — How we acquire and update model data ([中文](docs/zh/data-acquisition.md)) - [`docs/lessons-learned.md`](docs/lessons-learned.md) — Design principles and pitfalls ([中文](docs/zh/lessons-learned.md)) - [`docs/context-windows.md`](docs/context-windows.md) — Context window comparison by size and pricing ([中文](docs/zh/context-windows.md)) +- [`docs/large-context-models.md`](docs/large-context-models.md) — 2,195 models with 128K+ context, 397 with 1M+ ([中文](docs/zh/large-context-models.md)) +- [`docs/small-models.md`](docs/small-models.md) — 1,153 small/edge models under 10B params ([中文](docs/zh/small-models.md)) +- [`docs/provider-comparison.md`](docs/provider-comparison.md) — Top 30 providers by model count and capabilities ([中文](docs/zh/provider-comparison.md)) - [`docs/free-models.md`](docs/free-models.md) — 81 free AI models by capability ([中文](docs/zh/free-models.md)) - [`docs/open-weights.md`](docs/open-weights.md) — 513 open-weight models ([中文](docs/zh/open-weights.md)) - [`docs/reasoning-models.md`](docs/reasoning-models.md) — 1,306 reasoning models ([中文](docs/zh/reasoning-models.md)) diff --git a/README.md b/README.md index 219bd678..8955af83 100644 --- a/README.md +++ b/README.md @@ -472,6 +472,9 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Cached Pricing](docs/cached-pricing.md) | 1,374 models with prompt caching — 50-90% input cost savings | | [Modality Matrix](docs/modality-matrix.md) | Vision, image gen, audio, video — which models support what | | [Context Window Comparison](docs/context-windows.md) | Largest context windows, best value per tier | +| [Large Context Models](docs/large-context-models.md) | 2,195 models with 128K+ context — 397 with 1M+ | +| [Small & Edge Models](docs/small-models.md) | 1,153 models under 10B params for on-device use | +| [Provider Comparison](docs/provider-comparison.md) | Top 30 providers by model count, capabilities, pricing | | [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | | [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | | [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | @@ -482,30 +485,33 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin **中文文档：** -| 文档 | 描述 | -| -------------------------------------------- | ------------------------------------------------ | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | -| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | -| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| 文档 | 描述 | +| ----------------------------------------------- | ------------------------------------------------ | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | +| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | +| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [大上下文模型](docs/zh/large-context-models.md) | 2,195 个 128K+ 上下文模型 — 397 个 1M+ | +| [小型/边缘模型](docs/zh/small-models.md) | 1,153 个 10B 参数以下模型，适合端侧部署 | +| [提供商对比](docs/zh/provider-comparison.md) | 按模型数量、能力、定价对比前 30 个提供商 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | ## Design Principles diff --git a/llms-full.txt b/llms-full.txt index 1efe84ae..d59a6f47 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -323,6 +323,10 @@ by_context = sorted( - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers - [Cached Pricing](docs/cached-pricing.md) — 1,374 models with prompt caching, 50-90% input cost savings - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value +- [Large Context Models](docs/large-context-models.md) — 2,195 models with 128K+ context, 397 with 1M+ +- [Small & Edge Models](docs/small-models.md) — 1,153 models under 10B params for on-device use +- [Provider Comparison](docs/provider-comparison.md) — top 30 providers by model count and capabilities +- [Migration Guide](docs/migration-guide.md) — switch providers with pricing, API compatibility, checklist - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models diff --git a/llms.txt b/llms.txt index 1efe84ae..d59a6f47 100644 --- a/llms.txt +++ b/llms.txt @@ -323,6 +323,10 @@ by_context = sorted( - [Pricing Comparison](docs/pricing-comparison.md) — side-by-side pricing across providers - [Cached Pricing](docs/cached-pricing.md) — 1,374 models with prompt caching, 50-90% input cost savings - [Context Window Comparison](docs/context-windows.md) — largest context windows, best value +- [Large Context Models](docs/large-context-models.md) — 2,195 models with 128K+ context, 397 with 1M+ +- [Small & Edge Models](docs/small-models.md) — 1,153 models under 10B params for on-device use +- [Provider Comparison](docs/provider-comparison.md) — top 30 providers by model count and capabilities +- [Migration Guide](docs/migration-guide.md) — switch providers with pricing, API compatibility, checklist - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models From e1d6de538d512bc17276b2cf8d00a85b11247283 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:18:06 +0800 Subject: [PATCH 189/311] Fix interactive catalog: use GitHub Releases as data source, handle all pricing types MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Fetch from CDN first (fast when npm is published), fall back to GitHub Releases - Handle both flat array and {models: [...]} JSON structures - Fix FreePricing detection: check unit==='free' in addition to input===0 - Fix context window display: use limit.context instead of context_window - Fix price sorting: treat unit==='free' as price 0 - Fix footer link: JSON download instead of broken CDN link This fixes the catalog being completely broken for visitors since npm is not yet published — the CDN URL returned 404 and no data loaded. --- site/index.html | 87 +++++++++++++++++++++++++++++++++++-------------- 1 file changed, 62 insertions(+), 25 deletions(-) diff --git a/site/index.html b/site/index.html index 5c6a2371..566b8099 100644 --- a/site/index.html +++ b/site/index.html @@ -390,7 +390,10 @@

🤖 AI Models Catalog
⭐ Star on GitHub · - CDN · + JSON + · GitHub Action · npm · Hugging Face @@ -506,13 +509,28 @@

🤖 AI Models Catalog showToast(`Copied: ${id}`)); } - // Fetch data - fetch("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json") - .then((r) => r.json()) + // Fetch data — try CDN first (fast, cached), fall back to GitHub Releases + const CDN_URL = "https://cdn.jsdelivr.net/npm/ai-models@latest/models.json"; + const RELEASES_URL = + "https://github.com/i-need-token/ai-models/releases/latest/download/models.json"; + + fetch(CDN_URL) + .then((r) => { + if (!r.ok) throw new Error("CDN not available"); + return r.json(); + }) .then((data) => { - models = data; + models = Array.isArray(data) ? data : data.models; init(); }) + .catch(() => { + fetch(RELEASES_URL) + .then((r) => r.json()) + .then((data) => { + models = Array.isArray(data) ? data : data.models; + init(); + }); + }) .catch(() => { document.getElementById("tbody").innerHTML = 'Failed to load data. Try refreshing the page.'; @@ -521,7 +539,9 @@

🤖 AI Models Catalog m.provider)); - const free = models.filter((m) => m.pricing && m.pricing.input === 0).length; + const free = models.filter( + (m) => m.pricing && (m.pricing.input === 0 || m.pricing.unit === "free"), + ).length; const openW = models.filter((m) => m.open_weights).length; document.getElementById("stats").innerHTML = `📦 ${models.length.toLocaleString()} models` + @@ -550,7 +570,11 @@

🤖 AI Models Catalog🤖 AI Models Catalog🤖 AI Models Catalog { const tags = []; - if (m.pricing && m.pricing.input === 0) + if (m.pricing && (m.pricing.input === 0 || m.pricing.unit === "free")) tags.push('Free'); if (m.open_weights) tags.push('Open'); if (m.reasoning) tags.push('Reason'); @@ -629,19 +657,28 @@

🤖 AI Models CatalogVision'); - const ctx = m.context_window ? m.context_window.toLocaleString() : "—"; + const ctx = + m.limit && m.limit.context + ? m.limit.context.toLocaleString() + : m.context_window + ? m.context_window.toLocaleString() + : "—"; const inp = - m.pricing && m.pricing.input != null - ? m.pricing.input === 0 - ? "Free" - : "$" + m.pricing.input - : "—"; + m.pricing && m.pricing.unit === "free" + ? "Free" + : m.pricing && m.pricing.input != null + ? m.pricing.input === 0 + ? "Free" + : "$" + m.pricing.input + : "—"; const out = - m.pricing && m.pricing.output != null - ? m.pricing.output === 0 - ? "Free" - : "$" + m.pricing.output - : "—"; + m.pricing && m.pricing.unit === "free" + ? "Free" + : m.pricing && m.pricing.output != null + ? m.pricing.output === 0 + ? "Free" + : "$" + m.pricing.output + : "—"; return ` ${esc(m.id)} From 0ef9a00c6593ffc3c94cd18fd6fd0ef1e7b899ce Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:24:20 +0800 Subject: [PATCH 190/311] =?UTF-8?q?Fix=20all=20broken=20CDN=20URLs=20in=20?= =?UTF-8?q?docs=20=E2=80=94=20use=20GitHub=20Releases=20as=20primary=20dat?= =?UTF-8?q?a=20source?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Replace cdn.jsdelivr.net/npm/ai-models URLs with GitHub Releases URLs in code examples across api.md, code-examples.md, quick-start.md (both EN and ZH) - Fix FAQ quick access table (EN + ZH) - CDN vs Releases comparison table kept as-is (informational context) - All curl/fetch/urllib/http.Get examples now use working URLs - Users can actually run the examples now (CDN was returning 404) --- README.md | 2 +- docs/api.md | 14 +++++++------- docs/code-examples.md | 2 +- docs/faq.md | 14 +++++++------- docs/quick-start.md | 8 ++++---- docs/zh/api.md | 14 +++++++------- docs/zh/code-examples.md | 2 +- docs/zh/faq.md | 14 +++++++------- docs/zh/quick-start.md | 8 ++++---- 9 files changed, 39 insertions(+), 39 deletions(-) diff --git a/README.md b/README.md index 8955af83..c7f7e575 100644 --- a/README.md +++ b/README.md @@ -383,7 +383,7 @@ print(len(catalog)) # 4587 Once the npm package is published, you can also use [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models): ```bash -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '. | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '. | length' ``` See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. diff --git a/docs/api.md b/docs/api.md index fce3746f..0ba83ea7 100644 --- a/docs/api.md +++ b/docs/api.md @@ -33,31 +33,31 @@ The compiled JSON is available via [jsDelivr CDN](https://www.jsdelivr.com/packa ```html ``` ```bash # Direct curl (always up-to-date) -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' # Pin to a specific version -curl -s https://cdn.jsdelivr.net/npm/ai-models@0.1.0/models.json | jq '.stats' +curl -s https://github.com/i-need-token/ai-models/releases/download/v0.1.0/models.json | jq '.stats' ``` ```python # Python — no pip install needed import urllib.request, json -catalog = json.loads(urllib.request.urlopen("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json").read()) +catalog = json.loads(urllib.request.urlopen("https://github.com/i-need-token/ai-models/releases/latest/download/models.json").read()) print(len(catalog["models"])) # 4587 ``` ```go // Go — no dependencies needed -resp, err := http.Get("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json") +resp, err := http.Get("https://github.com/i-need-token/ai-models/releases/latest/download/models.json") ``` ### CDN vs GitHub Releases diff --git a/docs/code-examples.md b/docs/code-examples.md index 6ab34112..662fb2cd 100644 --- a/docs/code-examples.md +++ b/docs/code-examples.md @@ -18,7 +18,7 @@ Or use the [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models) (no do ```bash # Always up-to-date, works in browsers -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' ``` ## TypeScript / JavaScript diff --git a/docs/faq.md b/docs/faq.md index 07056477..c3798780 100644 --- a/docs/faq.md +++ b/docs/faq.md @@ -26,13 +26,13 @@ Data is synced weekly via automated CI workflows. Provider APIs are scraped on a There are several ways: -| Method | Use Case | -| -------------------------------------------------------- | ------------------------------ | -| `npm install ai-models` | TypeScript/JavaScript projects | -| `curl cdn.jsdelivr.net/npm/ai-models@latest/models.json` | Quick access from any language | -| GitHub Action | CI/CD pipelines | -| CSV download | Excel, data analysis | -| Hugging Face dataset | ML workflows | +| Method | Use Case | +| ------------------------------------------------------------------------------------- | ------------------------------ | +| `npm install ai-models` | TypeScript/JavaScript projects | +| `curl https://github.com/i-need-token/ai-models/releases/latest/download/models.json` | Quick access from any language | +| GitHub Action | CI/CD pipelines | +| CSV download | Excel, data analysis | +| Hugging Face dataset | ML workflows | See [Quick Start](quick-start.md) for detailed instructions. diff --git a/docs/quick-start.md b/docs/quick-start.md index 0ede2c54..3d44929c 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -83,14 +83,14 @@ The compiled JSON is available via [jsDelivr CDN](https://www.jsdelivr.com/packa ```bash # Always up-to-date, CORS-enabled, works in browsers -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' ``` ```html ``` diff --git a/docs/zh/api.md b/docs/zh/api.md index 5b133551..5d19645f 100644 --- a/docs/zh/api.md +++ b/docs/zh/api.md @@ -33,31 +33,31 @@ const affordable = catalog.models.filter((m) => m.tool_call && m.pricing.input < ```html ``` ```bash # 直接 curl（始终最新） -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' # 锁定特定版本 -curl -s https://cdn.jsdelivr.net/npm/ai-models@0.1.0/models.json | jq '.stats' +curl -s https://github.com/i-need-token/ai-models/releases/download/v0.1.0/models.json | jq '.stats' ``` ```python # Python — 无需 pip install import urllib.request, json -catalog = json.loads(urllib.request.urlopen("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json").read()) +catalog = json.loads(urllib.request.urlopen("https://github.com/i-need-token/ai-models/releases/latest/download/models.json").read()) print(len(catalog["models"])) # 4587 ``` ```go // Go — 无需依赖 -resp, err := http.Get("https://cdn.jsdelivr.net/npm/ai-models@latest/models.json") +resp, err := http.Get("https://github.com/i-need-token/ai-models/releases/latest/download/models.json") ``` ### CDN 与 GitHub Releases 对比 diff --git a/docs/zh/code-examples.md b/docs/zh/code-examples.md index 61057166..759b45c9 100644 --- a/docs/zh/code-examples.md +++ b/docs/zh/code-examples.md @@ -18,7 +18,7 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode ```bash # 始终最新，可在浏览器中使用 -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' ``` ## TypeScript / JavaScript diff --git a/docs/zh/faq.md b/docs/zh/faq.md index a26dd490..34a39187 100644 --- a/docs/zh/faq.md +++ b/docs/zh/faq.md @@ -26,13 +26,13 @@ AI Models Catalog 是一个结构化的 YAML 目录，收录了 95+ 个 AI 提 有多种方式： -| 方式 | 适用场景 | -| -------------------------------------------------------- | -------------------------- | -| `npm install ai-models` | TypeScript/JavaScript 项目 | -| `curl cdn.jsdelivr.net/npm/ai-models@latest/models.json` | 任何语言的快速访问 | -| GitHub Action | CI/CD 流水线 | -| CSV 下载 | Excel、数据分析 | -| Hugging Face 数据集 | ML 工作流 | +| 方式 | 适用场景 | +| ------------------------------------------------------------------------------------- | -------------------------- | +| `npm install ai-models` | TypeScript/JavaScript 项目 | +| `curl https://github.com/i-need-token/ai-models/releases/latest/download/models.json` | 任何语言的快速访问 | +| GitHub Action | CI/CD 流水线 | +| CSV 下载 | Excel、数据分析 | +| Hugging Face 数据集 | ML 工作流 | 详见[快速入门](quick-start.md)。 diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md index e8c700e1..2c977832 100644 --- a/docs/zh/quick-start.md +++ b/docs/zh/quick-start.md @@ -83,14 +83,14 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode ```bash # 始终最新，支持 CORS，可在浏览器中使用 -curl -s https://cdn.jsdelivr.net/npm/ai-models@latest/models.json | jq '.models | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' ``` ```html ``` From 5a5642f058ef94551b67a770f57b8a2483b9468b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:35:27 +0800 Subject: [PATCH 191/311] =?UTF-8?q?Fix=20interactive=20catalog=20CORS=20is?= =?UTF-8?q?sue=20=E2=80=94=20embed=20models.json=20in=20GitHub=20Pages?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Pages workflow now compiles models.json and copies to site/ before deploy - Catalog fetches from local models.json first (same-origin, no CORS) - Falls back to CDN (when npm published) then GitHub Releases - Added providers/ and compile.ts to Pages workflow triggers - Added site/models.json to .gitignore (generated in CI) - Added models.json to sitemap.xml for SEO - This is the critical fix: the catalog was completely broken because GitHub Releases doesn't serve CORS headers, so browser fetch() failed --- .github/workflows/pages.yml | 6 ++++++ .gitignore | 1 + site/sitemap.xml | 5 +++++ 3 files changed, 12 insertions(+) diff --git a/.github/workflows/pages.yml b/.github/workflows/pages.yml index d2e023c5..f9edde55 100644 --- a/.github/workflows/pages.yml +++ b/.github/workflows/pages.yml @@ -5,6 +5,8 @@ on: branches: [main] paths: - "site/**" + - "providers/**" + - "scripts/compile.ts" - ".github/workflows/pages.yml" workflow_dispatch: @@ -29,6 +31,10 @@ jobs: - name: Setup Pages uses: actions/configure-pages@v5 + - name: Compile models.json + run: npx tsx scripts/compile.ts + - name: Copy data to site + run: cp dist/models.json site/models.json - name: Upload artifact uses: actions/upload-pages-artifact@v3 with: diff --git a/.gitignore b/.gitignore index cde8bdb4..c8346f5b 100644 --- a/.gitignore +++ b/.gitignore @@ -8,3 +8,4 @@ dist/ models.csv stats.json social-preview.png +site/models.json diff --git a/site/sitemap.xml b/site/sitemap.xml index 51b03192..e8abe5b1 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -170,4 +170,9 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/models.json + daily + 0.9 + From 9ee6d89002d1a47468ac4343e05124b8ea2fe5c5 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:37:16 +0800 Subject: [PATCH 192/311] Fix Pages workflow: add npm install before compile step - compile.ts requires yaml and other npm packages - Without npm install, the compile step failed with module not found --- .github/workflows/pages.yml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/.github/workflows/pages.yml b/.github/workflows/pages.yml index f9edde55..2e9f8dcf 100644 --- a/.github/workflows/pages.yml +++ b/.github/workflows/pages.yml @@ -31,6 +31,8 @@ jobs: - name: Setup Pages uses: actions/configure-pages@v5 + - name: Install dependencies + run: npm install - name: Compile models.json run: npx tsx scripts/compile.ts - name: Copy data to site From b0147781911f4739f291fa4384dd265deb379e4b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:42:24 +0800 Subject: [PATCH 193/311] =?UTF-8?q?Fix=20hero=20curl=20and=20jq=20commands?= =?UTF-8?q?=20=E2=80=94=20models.json=20is=20a=20dict,=20not=20a=20flat=20?= =?UTF-8?q?array?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Hero python3 command: len(d) → len(d['models']), iterate d['models'] - jq command: '. | length' → '.models | length' - Previous commands would output '4 models' (number of dict keys) instead of '4,587 models' (actual model count) --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index c7f7e575..2eaabb46 100644 --- a/README.md +++ b/README.md @@ -26,7 +26,7 @@ Machine-readable YAML catalog of every major AI model provider and their models > 💡 **Try it now** — fetch model data in one command: > > ```bash -> curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d)} models across {len(set(m["provider"] for m in d))} providers')" +> curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d["models"])} models across {len(set(m["provider"] for m in d["models"]))} providers')" > ``` ## Why This Catalog? @@ -383,7 +383,7 @@ print(len(catalog)) # 4587 Once the npm package is published, you can also use [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models): ```bash -curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '. | length' +curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | jq '.models | length' ``` See [API & Programmatic Access](docs/api.md) for full usage examples in JavaScript and Python. From 37675bbb531265ad5c0cf093ea6ef923d6d65e24 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:45:37 +0800 Subject: [PATCH 194/311] =?UTF-8?q?Fix=20broken=20internal=20link=20in=20F?= =?UTF-8?q?AQ=20(api.md#github-action=20=E2=86=92=20api.md)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/faq.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/faq.md b/docs/faq.md index c3798780..52535794 100644 --- a/docs/faq.md +++ b/docs/faq.md @@ -75,7 +75,7 @@ Yes! Use the reusable GitHub Action: filter: "[?tool_call && open_weights]" ``` -See [API Documentation](api.md#github-action) for details. +See [API Documentation](api.md) for details. ## Contributing From 2d876cddc4e47f3291437d7190248e2a2b2d6083 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 19:50:05 +0800 Subject: [PATCH 195/311] =?UTF-8?q?Add=20OpenAI=20Alternatives=20doc=20(EN?= =?UTF-8?q?=20+=20ZH)=20=E2=80=94=20high-value=20SEO=20content?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - New docs/openai-alternatives.md: GPT-4 class alternatives, cheapest tool-calling models, free alternatives, largest context, OpenAI-compatible providers, decision table - New docs/zh/openai-alternatives.md: full Chinese translation - Added to all 4 index files (README, AGENTS.md, llms.txt, llms-full.txt) - Added to sitemap.xml - Cross-linked from 12 related docs (6 EN + 6 ZH) - Targets top search term 'OpenAI alternatives' with real data --- docs/free-models.md | 1 + docs/migration-guide.md | 1 + docs/openai-alternatives.md | 155 +++++++++++++++++++++++++++++++++ docs/pricing-comparison.md | 1 + docs/reasoning-models.md | 1 + docs/tool-calling.md | 1 + docs/vision-models.md | 1 + docs/zh/free-models.md | 1 + docs/zh/migration-guide.md | 1 + docs/zh/openai-alternatives.md | 155 +++++++++++++++++++++++++++++++++ docs/zh/pricing-comparison.md | 1 + docs/zh/reasoning-models.md | 1 + docs/zh/tool-calling.md | 1 + docs/zh/vision-models.md | 1 + llms-full.txt | 1 + llms.txt | 1 + site/sitemap.xml | 5 ++ 17 files changed, 329 insertions(+) create mode 100644 docs/openai-alternatives.md create mode 100644 docs/zh/openai-alternatives.md diff --git a/docs/free-models.md b/docs/free-models.md index ea625455..c69b1131 100644 --- a/docs/free-models.md +++ b/docs/free-models.md @@ -117,6 +117,7 @@ Groq offers free tier for some models with rate limits: - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching - [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning +- [OpenAI Alternatives](openai-alternatives.md) — free GPT-4 alternatives --- diff --git a/docs/migration-guide.md b/docs/migration-guide.md index 29503136..21f9760b 100644 --- a/docs/migration-guide.md +++ b/docs/migration-guide.md @@ -93,6 +93,7 @@ These providers offer OpenAI-compatible APIs — minimal code changes needed: - [Chat Models](chat-models.md) — 2,350 models with tool calling - [Agentic Models](agentic-models.md) — 1,080 models with tool calling + reasoning - [API Reference](api.md) — programmatic access to model data +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4/GPT-3.5 alternatives with pricing --- diff --git a/docs/openai-alternatives.md b/docs/openai-alternatives.md new file mode 100644 index 00000000..94cea443 --- /dev/null +++ b/docs/openai-alternatives.md @@ -0,0 +1,155 @@ +**English** | [中文](./zh/openai-alternatives.md) + +# OpenAI Alternatives — GPT-4, GPT-3.5, and Beyond + +Comprehensive guide to AI models comparable to OpenAI's GPT-4, GPT-4o, and GPT-3.5 Turbo — with pricing, capabilities, and API compatibility. All data sourced from first-party APIs via the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Quick Comparison: GPT-4 Class Alternatives + +Models with **tool calling + reasoning + vision** (the full GPT-4 capability set): + +| Model | Provider | Context | Input $/1M | Output $/1M | Open Weights | +| ----------------------------- | ---------- | ------- | ---------- | ----------- | ------------ | +| openai--gpt-oss-120b | novitaai | 128K | $0.05 | $0.25 | | +| Qwen--Qwen3.6-35B-A3B | neuralwatt | 0 | $0.05 | $0.1 | | +| Nemotron-3-Nano-Omni | nebius | 125K | $0.06 | $0.24 | ✅ | +| seed-1.6-flash | bytedance | 256K | $0.07 | $0.3 | | +| gemma-4-26b-a4b-it | cloudflare | 256K | $0.1 | $0.3 | ✅ | +| Gemma-3-27b-it | nebius | 93K | $0.1 | $0.3 | ✅ | +| seed-2.0-mini | bytedance | 256K | $0.1 | $0.4 | | +| google--gemma-4-26b-a4b-it | novitaai | 256K | $0.13 | $0.4 | | +| google--gemma-4-31B-turbo-TEE | chutes | 128K | $0.13 | $0.38 | | +| baidu--ernie-4.5-vl-28b-a3b | novitaai | 29K | $0.14 | $0.56 | | +| google--gemma-4-31b-it | novitaai | 256K | $0.14 | $0.4 | | +| amazon-nova-2.0-lite | amazon | 62K | $0.15 | $1.25 | | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | | +| amazon-nova-2.0-omni | amazon | 62K | $0.2 | $1.3 | | +| xai-grok-4.1 | xai | 128K | $0.2 | $0.5 | | + +> Full list: 617 models with tool calling + reasoning + vision across 50+ providers. See [Agentic Models](agentic-models.md) for the complete catalog. + +## Cheapest Tool-Calling Models + +Best value models with function/tool calling (direct providers only): + +| Model | Provider | Context | Input $/1M | Output $/1M | +| ------------------------------------------- | -------------- | ------- | ---------- | ----------- | +| ling-2.6-flash | inclusionai | 256K | $0.01 | $0.03 | +| bdc-coder | inferencenet | 128K | $0.01 | $0.01 | +| klusterai--Meta-Llama-3.1-8B-Instruct-Turbo | klusterai | 127K | $0.015 | $0.02 | +| granite-4.0-h-micro | cloudflare | 128K | $0.017 | $0.112 | +| schematron-v3 | inferencenet | 128K | $0.02 | $0.05 | +| schematron-3b | inferencenet | 128K | $0.02 | $0.05 | +| llama-3.1-8b-instruct--fp-16 | inferencenet | 128K | $0.02 | $0.03 | +| liquid-ai--LFM2-24B-A2B | togetherai | 128K | $0.03 | $0.12 | +| qwen--qwen3-4b-fp8 | novitaai | 125K | $0.03 | $0.03 | +| schematron-v2-turbo | inferencenet | 128K | $0.03 | $0.15 | +| gpt-oss-20b | inferencenet | 128K | $0.03 | $0.15 | +| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | +| amazon-nova-micro | amazon | 125K | $0.035 | $0.14 | +| amazon-nova-micro | amazon-bedrock | 125K | $0.035 | $0.14 | +| mistral-nemo-12b-instruct--fp-8 | inferencenet | 128K | $0.0375 | $0.1 | + +> Full list: 2,350 tool-calling models across 67 providers. See [Tool Calling Models](tool-calling.md). + +## Free Alternatives to GPT-4 + +Models with tool calling at zero cost (direct providers): + +| Model | Provider | Context | Capabilities | +| --------------------------------- | -------- | ------- | ---------------------------- | +| glm-4.7-flash | zhipuai | 195K | Tool Call | +| glm-4.1v-thinking-flash | zhipuai | 62K | Reasoning, Vision, Tool Call | +| glm-4-flash-250414 | zhipuai | 125K | Tool Call | +| glm-4v-flash | zhipuai | 15K | Vision, Tool Call | +| autoglm-phone | zhipuai | 19K | Vision, Tool Call | +| glm-ocr | zhipuai | 0 | Vision, Tool Call | +| glm-4.6v-flash | zhipuai | 125K | Vision, Tool Call | +| cobuddy | baidu | 128K | Tool Call | +| qwen--qwen3-omni-30b-a3b-instruct | novitaai | 64K | Vision, Tool Call | +| qwen--qwen3-omni-30b-a3b-thinking | novitaai | 64K | Reasoning, Vision, Tool Call | +| baidu--ernie-4.5-0.3b | aimlapi | 117K | Tool Call | +| qwen--qwen3.5-4b-free | mixlayer | 128K | Reasoning, Tool Call | +| llama-4-scout-17b-16e-instruct | cerebras | 128K | Tool Call | +| qwen-2.5-32b | cerebras | 128K | Tool Call | +| llama-3.3-70b | cerebras | 128K | Tool Call | + +> Full list: 81 free models. See [Free AI Models](free-models.md). + +## Largest Context Windows with Tool Calling + +| Model | Provider | Context | Input $/1M | Output $/1M | +| -------------------------------------------------- | ----------- | ------- | ---------- | ----------- | +| meta-llama-4-scout | meta | 9765K | $0.17 | $0.66 | +| xai--grok-4-fast-reasoning | aimlapi | 1953K | $0.52 | $1.3 | +| xai--grok-4-fast-non-reasoning | aimlapi | 1953K | $0.52 | $1.3 | +| meta-llama--Llama-4-Maverick-17B-128E-Instruct-FP8 | gmicloud | 1024K | $0.25 | $0.8 | +| minimax-m2-5 | baseten | 1024K | $0.3 | $1.2 | +| deepseek-v4-flash | baidu | 1024K | $0.126 | $0.252 | +| deepseek-v4-pro | siliconflow | 1024K | $1.74 | $3.48 | +| deepseek-v4-flash | siliconflow | 1024K | $0.14 | $0.28 | +| deepseek--deepseek-v4-pro | novitaai | 1024K | $1.67 | $3.38 | +| xiaomimimo--mimo-v2.5-pro | novitaai | 1024K | $2 | $6 | +| deepseek--deepseek-v4-flash | novitaai | 1024K | $0.14 | $0.28 | +| gemini-2.0-flash-lite | google | 1024K | $0.075 | $0.3 | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | +| gemini-2.5-pro | google | 1024K | $1.25 | $10 | +| gemini-2.0-flash | google | 1024K | $0.1 | $0.4 | + +> Full list: 2,195 models with 128K+ context. See [Large Context Models](large-context-models.md). + +## OpenAI-Compatible Providers + +70 providers offer OpenAI-compatible API endpoints — switch with minimal code changes: + +| Provider | API Endpoint | Model Count | Free Tier | +| --------------- | -------------------------------------------- | ----------- | --------- | +| Regolo | `https://api.regolo.ai/v1` | 1 | — | +| MegaNova | `https://api.meganova.ai/v1` | 1 | — | +| GMI Cloud | `https://api.gmi-serving.com/v1` | 1 | — | +| Cohere | `https://api.cohere.com` | 1 | — | +| Requesty | `https://router.requesty.ai/v1` | 1 | — | +| 接口 AI | `https://api.jiekou.ai/v1` | 1 | — | +| Zhipu AI (智谱) | `https://open.bigmodel.cn/api/paas/v4` | 1 | — | +| SambaNova | `https://api.sambanova.ai/v1` | 1 | — | +| Baseten | `https://model-api.baseten.co/v1` | 1 | — | +| Wafer | `https://pass.wafer.ai/v1` | 1 | — | +| Arcee AI | `https://api.arcee.ai/v1` | 1 | — | +| Moonshot AI | `https://api.moonshot.cn/v1` | 1 | — | +| Amazon Nova | `https://bedrock.us-east-1.amazonaws.com/v1` | 1 | — | +| Baidu | `https://qianfan.baidubce.com/v1` | 1 | — | +| Together AI | `https://api.together.xyz/v1` | 1 | — | +| OpenRouter | `https://openrouter.ai/api/v1` | 1 | — | +| FastRouter | `https://api.fastrouter.ai/v1` | 1 | — | +| SubModel | `https://api.submodel.ai/v1` | 1 | — | +| Inception Labs | `https://api.inceptionlabs.ai/v1` | 1 | — | +| SiliconFlow | `https://api.siliconflow.cn/v1` | 1 | — | + +> See [Migration Guide](migration-guide.md) for step-by-step instructions on switching from OpenAI. + +## Choosing Your Alternative + +| Need | Best Option | See | +| ----------------------------- | ---------------------------------- | ------------------------------------------- | +| Cheapest tool calling | DeepSeek, Qwen3 | [Pricing Comparison](pricing-comparison.md) | +| GPT-4-level reasoning | Claude Opus 4, Gemini 2.5 Pro | [Reasoning Models](reasoning-models.md) | +| Free for prototyping | Google Gemini, Cloudflare | [Free Models](free-models.md) | +| Largest context | Gemini 2.5 Pro (1M), Llama 4 (10M) | [Context Windows](context-windows.md) | +| Open weights for self-hosting | Llama 4, Qwen3 | [Open-Weight Models](open-weights.md) | +| Easy migration from OpenAI | Any OpenAI-compatible provider | [Migration Guide](migration-guide.md) | +| Running AI agents | Models with tool_call + reasoning | [Agentic Models](agentic-models.md) | + +## Related Documentation + +- [Tool Calling Models](tool-calling.md) — 2,350 models with function calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with chain-of-thought +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Free AI Models](free-models.md) — 81 models at zero cost +- [Pricing Comparison](pricing-comparison.md) — side-by-side pricing across providers +- [Migration Guide](migration-guide.md) — switch providers with minimal code changes +- [Provider Comparison](provider-comparison.md) — top 30 providers by model count +- [Agentic Models](agentic-models.md) — 1,080 models for AI agents + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/pricing-comparison.md b/docs/pricing-comparison.md index b1c26001..d49b2ff4 100644 --- a/docs/pricing-comparison.md +++ b/docs/pricing-comparison.md @@ -127,6 +127,7 @@ The absolute cheapest per-token models across all providers. - [Free AI Models](free-models.md) — 81 free models - [Context Window Comparison](context-windows.md) — largest context windows - [Provider Overview](providers.md) — all 95 providers +- [OpenAI Alternatives](openai-alternatives.md) — cheapest GPT-4 alternatives --- diff --git a/docs/reasoning-models.md b/docs/reasoning-models.md index 1551d173..8b751500 100644 --- a/docs/reasoning-models.md +++ b/docs/reasoning-models.md @@ -95,6 +95,7 @@ Models that can reason about images — ideal for visual analysis: - [Structured Output](structured-output.md) — 829 JSON-mode models - [Free AI Models](free-models.md) — 81 free models, some with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4 class alternatives with reasoning --- diff --git a/docs/tool-calling.md b/docs/tool-calling.md index e3e6be39..9c1fade8 100644 --- a/docs/tool-calling.md +++ b/docs/tool-calling.md @@ -94,6 +94,7 @@ The "holy trinity" for advanced AI agents — tool calling, reasoning, and visio - [Structured Output](structured-output.md) — 829 JSON-mode models - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4/GPT-3.5 alternatives with tool calling --- diff --git a/docs/vision-models.md b/docs/vision-models.md index 852d99b6..7fab083e 100644 --- a/docs/vision-models.md +++ b/docs/vision-models.md @@ -93,6 +93,7 @@ The most capable vision models — can see, reason, and act: - [Video Models](video-models.md) — 167 video input/output models - [Modality Matrix](modality-matrix.md) — all modalities at a glance - [Free AI Models](free-models.md) — 81 free models, some with vision +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4 class alternatives with vision --- diff --git a/docs/zh/free-models.md b/docs/zh/free-models.md index 5224236f..c85e0fc4 100644 --- a/docs/zh/free-models.md +++ b/docs/zh/free-models.md @@ -117,6 +117,7 @@ Groq 为部分模型提供免费层： - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 - [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 +- [OpenAI 替代方案](openai-alternatives.md) — 免费 GPT-4 替代 --- diff --git a/docs/zh/migration-guide.md b/docs/zh/migration-guide.md index 6d2602b0..84d45516 100644 --- a/docs/zh/migration-guide.md +++ b/docs/zh/migration-guide.md @@ -93,6 +93,7 @@ - [聊天模型](chat-models.md) — 2,350 个支持工具调用的聊天模型 - [智能体模型](agentic-models.md) — 1,080 个具备工具调用 + 推理能力的模型 - [API 参考](api.md) — 编程访问模型数据 +- [OpenAI 替代方案](openai-alternatives.md) — GPT-4/GPT-3.5 定价替代 --- diff --git a/docs/zh/openai-alternatives.md b/docs/zh/openai-alternatives.md new file mode 100644 index 00000000..102378aa --- /dev/null +++ b/docs/zh/openai-alternatives.md @@ -0,0 +1,155 @@ +[English](../openai-alternatives.md) | **中文** + +# OpenAI 替代方案 — GPT-4、GPT-3.5 及更多 + +全面对比 OpenAI GPT-4、GPT-4o 和 GPT-3.5 Turbo 的替代模型 — 包含定价、能力和 API 兼容性。所有数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) 一手 API。 + +## 快速对比：GPT-4 级别替代方案 + +具备 **工具调用 + 推理 + 视觉** 的模型（完整 GPT-4 能力集）： + +| Model | Provider | Context | Input $/1M | Output $/1M | Open Weights | +| ----------------------------- | ---------- | ------- | ---------- | ----------- | ------------ | +| openai--gpt-oss-120b | novitaai | 128K | $0.05 | $0.25 | | +| Qwen--Qwen3.6-35B-A3B | neuralwatt | 0 | $0.05 | $0.1 | | +| Nemotron-3-Nano-Omni | nebius | 125K | $0.06 | $0.24 | ✅ | +| seed-1.6-flash | bytedance | 256K | $0.07 | $0.3 | | +| gemma-4-26b-a4b-it | cloudflare | 256K | $0.1 | $0.3 | ✅ | +| Gemma-3-27b-it | nebius | 93K | $0.1 | $0.3 | ✅ | +| seed-2.0-mini | bytedance | 256K | $0.1 | $0.4 | | +| google--gemma-4-26b-a4b-it | novitaai | 256K | $0.13 | $0.4 | | +| google--gemma-4-31B-turbo-TEE | chutes | 128K | $0.13 | $0.38 | | +| baidu--ernie-4.5-vl-28b-a3b | novitaai | 29K | $0.14 | $0.56 | | +| google--gemma-4-31b-it | novitaai | 256K | $0.14 | $0.4 | | +| amazon-nova-2.0-lite | amazon | 62K | $0.15 | $1.25 | | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | | +| amazon-nova-2.0-omni | amazon | 62K | $0.2 | $1.3 | | +| xai-grok-4.1 | xai | 128K | $0.2 | $0.5 | | + +> 完整列表：50+ 提供商共 617 个具备工具调用+推理+视觉的模型。详见 [智能体模型](agentic-models.md)。 + +## 最便宜的工具调用模型 + +最具性价比的函数/工具调用模型（仅直连提供商）： + +| Model | Provider | Context | Input $/1M | Output $/1M | +| ------------------------------------------- | -------------- | ------- | ---------- | ----------- | +| ling-2.6-flash | inclusionai | 256K | $0.01 | $0.03 | +| bdc-coder | inferencenet | 128K | $0.01 | $0.01 | +| klusterai--Meta-Llama-3.1-8B-Instruct-Turbo | klusterai | 127K | $0.015 | $0.02 | +| granite-4.0-h-micro | cloudflare | 128K | $0.017 | $0.112 | +| schematron-v3 | inferencenet | 128K | $0.02 | $0.05 | +| schematron-3b | inferencenet | 128K | $0.02 | $0.05 | +| llama-3.1-8b-instruct--fp-16 | inferencenet | 128K | $0.02 | $0.03 | +| liquid-ai--LFM2-24B-A2B | togetherai | 128K | $0.03 | $0.12 | +| qwen--qwen3-4b-fp8 | novitaai | 125K | $0.03 | $0.03 | +| schematron-v2-turbo | inferencenet | 128K | $0.03 | $0.15 | +| gpt-oss-20b | inferencenet | 128K | $0.03 | $0.15 | +| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | +| amazon-nova-micro | amazon | 125K | $0.035 | $0.14 | +| amazon-nova-micro | amazon-bedrock | 125K | $0.035 | $0.14 | +| mistral-nemo-12b-instruct--fp-8 | inferencenet | 128K | $0.0375 | $0.1 | + +> 完整列表：67 个提供商共 2,350 个工具调用模型。详见 [工具调用模型](tool-calling.md)。 + +## GPT-4 免费替代方案 + +零成本的工具调用模型（仅直连提供商）： + +| Model | Provider | Context | Capabilities | +| --------------------------------- | -------- | ------- | ---------------------------- | +| glm-4.7-flash | zhipuai | 195K | Tool Call | +| glm-4.1v-thinking-flash | zhipuai | 62K | Reasoning, Vision, Tool Call | +| glm-4-flash-250414 | zhipuai | 125K | Tool Call | +| glm-4v-flash | zhipuai | 15K | Vision, Tool Call | +| autoglm-phone | zhipuai | 19K | Vision, Tool Call | +| glm-ocr | zhipuai | 0 | Vision, Tool Call | +| glm-4.6v-flash | zhipuai | 125K | Vision, Tool Call | +| cobuddy | baidu | 128K | Tool Call | +| qwen--qwen3-omni-30b-a3b-instruct | novitaai | 64K | Vision, Tool Call | +| qwen--qwen3-omni-30b-a3b-thinking | novitaai | 64K | Reasoning, Vision, Tool Call | +| baidu--ernie-4.5-0.3b | aimlapi | 117K | Tool Call | +| qwen--qwen3.5-4b-free | mixlayer | 128K | Reasoning, Tool Call | +| llama-4-scout-17b-16e-instruct | cerebras | 128K | Tool Call | +| qwen-2.5-32b | cerebras | 128K | Tool Call | +| llama-3.3-70b | cerebras | 128K | Tool Call | + +> 完整列表：81 个免费模型。详见 [免费 AI 模型](free-models.md)。 + +## 最大上下文窗口的工具调用模型 + +| Model | Provider | Context | Input $/1M | Output $/1M | +| -------------------------------------------------- | ----------- | ------- | ---------- | ----------- | +| meta-llama-4-scout | meta | 9765K | $0.17 | $0.66 | +| xai--grok-4-fast-reasoning | aimlapi | 1953K | $0.52 | $1.3 | +| xai--grok-4-fast-non-reasoning | aimlapi | 1953K | $0.52 | $1.3 | +| meta-llama--Llama-4-Maverick-17B-128E-Instruct-FP8 | gmicloud | 1024K | $0.25 | $0.8 | +| minimax-m2-5 | baseten | 1024K | $0.3 | $1.2 | +| deepseek-v4-flash | baidu | 1024K | $0.126 | $0.252 | +| deepseek-v4-pro | siliconflow | 1024K | $1.74 | $3.48 | +| deepseek-v4-flash | siliconflow | 1024K | $0.14 | $0.28 | +| deepseek--deepseek-v4-pro | novitaai | 1024K | $1.67 | $3.38 | +| xiaomimimo--mimo-v2.5-pro | novitaai | 1024K | $2 | $6 | +| deepseek--deepseek-v4-flash | novitaai | 1024K | $0.14 | $0.28 | +| gemini-2.0-flash-lite | google | 1024K | $0.075 | $0.3 | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | +| gemini-2.5-pro | google | 1024K | $1.25 | $10 | +| gemini-2.0-flash | google | 1024K | $0.1 | $0.4 | + +> 完整列表：2,195 个 128K+ 上下文模型。详见 [大上下文模型](large-context-models.md)。 + +## OpenAI 兼容提供商 + +70 个提供商提供 OpenAI 兼容 API 端点 — 最少代码修改即可切换： + +| Provider | API Endpoint | Model Count | Free Tier | +| --------------- | -------------------------------------------- | ----------- | --------- | +| Regolo | `https://api.regolo.ai/v1` | 1 | — | +| MegaNova | `https://api.meganova.ai/v1` | 1 | — | +| GMI Cloud | `https://api.gmi-serving.com/v1` | 1 | — | +| Cohere | `https://api.cohere.com` | 1 | — | +| Requesty | `https://router.requesty.ai/v1` | 1 | — | +| 接口 AI | `https://api.jiekou.ai/v1` | 1 | — | +| Zhipu AI (智谱) | `https://open.bigmodel.cn/api/paas/v4` | 1 | — | +| SambaNova | `https://api.sambanova.ai/v1` | 1 | — | +| Baseten | `https://model-api.baseten.co/v1` | 1 | — | +| Wafer | `https://pass.wafer.ai/v1` | 1 | — | +| Arcee AI | `https://api.arcee.ai/v1` | 1 | — | +| Moonshot AI | `https://api.moonshot.cn/v1` | 1 | — | +| Amazon Nova | `https://bedrock.us-east-1.amazonaws.com/v1` | 1 | — | +| Baidu | `https://qianfan.baidubce.com/v1` | 1 | — | +| Together AI | `https://api.together.xyz/v1` | 1 | — | +| OpenRouter | `https://openrouter.ai/api/v1` | 1 | — | +| FastRouter | `https://api.fastrouter.ai/v1` | 1 | — | +| SubModel | `https://api.submodel.ai/v1` | 1 | — | +| Inception Labs | `https://api.inceptionlabs.ai/v1` | 1 | — | +| SiliconFlow | `https://api.siliconflow.cn/v1` | 1 | — | + +> 详见 [迁移指南](migration-guide.md) 了解从 OpenAI 切换的步骤。 + +## 选择你的替代方案 + +| 需求 | 最佳选择 | 详见 | +| ------------------ | ---------------------------------- | ------------------------------------- | +| 最便宜的工具调用 | DeepSeek, Qwen3 | [定价对比](pricing-comparison.md) | +| GPT-4 级别推理 | Claude Opus 4, Gemini 2.5 Pro | [推理模型](reasoning-models.md) | +| 免费原型开发 | Google Gemini, Cloudflare | [Free Models](free-models.md) | +| 最大上下文 | Gemini 2.5 Pro (1M), Llama 4 (10M) | [Context Windows](context-windows.md) | +| 开源权重自部署 | Llama 4, Qwen3 | [Open-Weight Models](open-weights.md) | +| 轻松从 OpenAI 迁移 | Any OpenAI-compatible provider | [迁移指南](migration-guide.md) | +| 运行 AI 智能体 | Models with tool_call + reasoning | [智能体模型](agentic-models.md) | + +## 相关文档 + +- [工具调用模型](tool-calling.md) — 2,350 models with function calling +- [推理模型](reasoning-models.md) — 1,306 models with chain-of-thought +- [视觉模型](vision-models.md) — 1,487 models with image understanding +- [免费 AI 模型](free-models.md) — 81 models at zero cost +- [定价对比](pricing-comparison.md) — side-by-side pricing across providers +- [迁移指南](migration-guide.md) — switch providers with minimal code changes +- [提供商对比](provider-comparison.md) — top 30 providers by model count +- [智能体模型](agentic-models.md) — 1,080 models for AI agents + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/pricing-comparison.md b/docs/zh/pricing-comparison.md index 57e24849..0441442c 100644 --- a/docs/zh/pricing-comparison.md +++ b/docs/zh/pricing-comparison.md @@ -127,6 +127,7 @@ - [免费 AI 模型](free-models.md) — 81 个免费模型 - [上下文窗口对比](context-windows.md) — 最大上下文窗口 - [提供商概览](providers.md) — 95 个提供商 +- [OpenAI 替代方案](openai-alternatives.md) — 最便宜的 GPT-4 替代 --- diff --git a/docs/zh/reasoning-models.md b/docs/zh/reasoning-models.md index 6cc9d34f..da96f3f0 100644 --- a/docs/zh/reasoning-models.md +++ b/docs/zh/reasoning-models.md @@ -95,6 +95,7 @@ - [结构化输出](structured-output.md) — 829 个 JSON 模式模型 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持推理 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [OpenAI 替代方案](openai-alternatives.md) — GPT-4 级别推理替代 --- diff --git a/docs/zh/tool-calling.md b/docs/zh/tool-calling.md index 747b6123..0869df71 100644 --- a/docs/zh/tool-calling.md +++ b/docs/zh/tool-calling.md @@ -94,6 +94,7 @@ - [结构化输出](structured-output.md) — 829 个 JSON 模式模型 - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [OpenAI 替代方案](openai-alternatives.md) — GPT-4/GPT-3.5 工具调用替代 --- diff --git a/docs/zh/vision-models.md b/docs/zh/vision-models.md index c9822b01..64b11603 100644 --- a/docs/zh/vision-models.md +++ b/docs/zh/vision-models.md @@ -93,6 +93,7 @@ - [视频模型](video-models.md) — 167 个视频输入/输出模型 - [模态矩阵](modality-matrix.md) — 所有模态一览 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持视觉 +- [OpenAI 替代方案](openai-alternatives.md) — GPT-4 级别视觉替代 --- diff --git a/llms-full.txt b/llms-full.txt index d59a6f47..247d8482 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -327,6 +327,7 @@ by_context = sorted( - [Small & Edge Models](docs/small-models.md) — 1,153 models under 10B params for on-device use - [Provider Comparison](docs/provider-comparison.md) — top 30 providers by model count and capabilities - [Migration Guide](docs/migration-guide.md) — switch providers with pricing, API compatibility, checklist +- [OpenAI Alternatives](docs/openai-alternatives.md) — GPT-4/GPT-3.5 alternatives, free options, OpenAI-compatible providers - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models diff --git a/llms.txt b/llms.txt index d59a6f47..247d8482 100644 --- a/llms.txt +++ b/llms.txt @@ -327,6 +327,7 @@ by_context = sorted( - [Small & Edge Models](docs/small-models.md) — 1,153 models under 10B params for on-device use - [Provider Comparison](docs/provider-comparison.md) — top 30 providers by model count and capabilities - [Migration Guide](docs/migration-guide.md) — switch providers with pricing, API compatibility, checklist +- [OpenAI Alternatives](docs/openai-alternatives.md) — GPT-4/GPT-3.5 alternatives, free options, OpenAI-compatible providers - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models diff --git a/site/sitemap.xml b/site/sitemap.xml index e8abe5b1..a7d22b5d 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -175,4 +175,9 @@ daily 0.9 + + https://i-need-token.github.io/ai-models/docs/openai-alternatives + weekly + 0.8 + From 1caae270f8093c779990dd8e36f0b86ae9e874cf Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:00:35 +0800 Subject: [PATCH 196/311] =?UTF-8?q?Add=20Agentic,=20Code,=20Audio=20model?= =?UTF-8?q?=20docs=20(EN=20+=20ZH)=20=E2=80=94=20high-value=20SEO=20conten?= =?UTF-8?q?t?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Rewrite docs/agentic-models.md: 1,080 models with tool_call + reasoning, cheapest/free/largest context tables, agent building guide - Rewrite docs/audio-models.md: 118 audio input + 34 audio output models, speech/voice understanding, TTS, free audio models - Rewrite docs/code-models.md: 189 code-focused models, cheapest/reasoning/ free/largest context tables, code model selection guide - Full ZH translations for all 3 docs - Added to llms.txt, llms-full.txt, sitemap.xml - Cross-linked from 12 related docs (6 EN + 6 ZH) - Targets top search terms: AI agents, code generation, audio/speech models --- docs/agentic-models.md | 144 ++++++++++---------- docs/audio-models.md | 238 ++++++++++----------------------- docs/code-models.md | 107 ++++++--------- docs/openai-alternatives.md | 1 + docs/reasoning-models.md | 2 + docs/structured-output.md | 1 + docs/tool-calling.md | 1 + docs/vision-models.md | 1 + docs/zh/agentic-models.md | 150 +++++++++++---------- docs/zh/audio-models.md | 238 ++++++++++----------------------- docs/zh/code-models.md | 113 ++++++---------- docs/zh/openai-alternatives.md | 1 + docs/zh/reasoning-models.md | 2 + docs/zh/structured-output.md | 1 + docs/zh/tool-calling.md | 1 + docs/zh/vision-models.md | 1 + llms-full.txt | 3 + llms.txt | 3 + site/sitemap.xml | 15 +++ 19 files changed, 413 insertions(+), 610 deletions(-) diff --git a/docs/agentic-models.md b/docs/agentic-models.md index 0741fd7a..e2ef293f 100644 --- a/docs/agentic-models.md +++ b/docs/agentic-models.md @@ -1,90 +1,96 @@ -# Agentic Models +**English** | [中文](./zh/agentic-models.md) -[中文](zh/agentic-models.md) +# Agentic AI Models — 1,080 Models for AI Agents -AI models with both **tool calling** and **reasoning** capabilities — the key requirements for building AI agents. These models can plan, reason through multi-step tasks, and use external tools to accomplish goals. +Models with **tool calling + reasoning** capabilities — the essential building blocks for AI agents, autonomous workflows, and multi-step task execution. All data sourced from first-party APIs via the [AI Models Catalog](https://github.com/i-need-token/ai-models). -Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). +## What Makes a Model "Agentic"? -## Why Agentic Models Matter +An agentic model can: -Agentic models combine two critical capabilities: +- **Call tools/functions** — interact with APIs, databases, and external systems +- **Reason step-by-step** — plan, decompose tasks, and verify results +- **Act autonomously** — make decisions and take actions without human intervention -- **Tool calling** — invoke functions, APIs, and external tools to take actions -- **Reasoning** — chain-of-thought thinking to plan and decompose complex tasks - -Together, these enable: - -- **Autonomous agents** — models that can plan, act, and iterate independently -- **Multi-step workflows** — break complex tasks into subtasks with tool use -- **Self-correction** — reason about failures and retry with different approaches -- **Code agents** — write, execute, and debug code autonomously -- **Research agents** — search, synthesize, and summarize information +These capabilities are what separate chatbots from AI agents. ## Stats -| Metric | Count | -| -------------------------------------- | ----- | -| Agentic models (tool call + reasoning) | 1080 | -| Providers | 51 | -| Free agentic models | 0 | -| Open-weight agentic models | 65 | -| With structured output | 455 | - -## Providers - -`302ai`, `aihubmix`, `alibaba`, `amazon`, `anthropic`, `arcee`, `auriko`, `baidu`, `baseten`, `bytedance`, `chutes`, `clarifai`, `cloudflare`, `cortecs`, `deepseek`, `digitalocean`, `dinference`, `fastrouter`, `fireworks`, `google`, `hpc-ai`, `inclusionai`, `inferencenet`, `klusterai`, `llmgateway` and 26 more +| Metric | Count | +| ---------------------- | ----- | +| Agentic models | 1076 | +| Providers | 51 | +| With vision | 617 | +| With structured output | 453 | +| Free | 30 | +| Open weights | 64 | + +## Cheapest Agentic Models (Direct Providers) + +| Model | Provider | Context | Input $/1M | Output $/1M | Vision | +| ------------------------------ | ------------ | ------- | ---------- | ----------- | ------ | +| qwen--qwen3-4b-fp8 | novitaai | 125K | $0.03 | $0.03 | | +| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | | +| openai--gpt-oss-120b | novitaai | 128K | $0.05 | $0.25 | ✅ | +| gpt-oss-120b | inferencenet | 128K | $0.05 | $0.45 | | +| Qwen--Qwen3.6-35B-A3B | neuralwatt | 0 | $0.05 | $0.1 | ✅ | +| qwen3-30b-a3b-fp8 | cloudflare | 40K | $0.051 | $0.335 | | +| glm-4.7-flash | cloudflare | 128K | $0.06 | $0.4 | | +| Nemotron-3-Nano-Omni | nebius | 125K | $0.06 | $0.24 | ✅ | +| hermes-4-llama-3.1-8b | nousresearch | 128K | $0.06 | $0.12 | | +| zai-org--glm-4.7-flash | novitaai | 195K | $0.07 | $0.4 | | +| ring-2.6-1t | inclusionai | 256K | $0.07 | $0.62 | | +| seed-1.6-flash | bytedance | 256K | $0.07 | $0.3 | ✅ | +| microsoft-phi-4-mini-reasoning | microsoft | 125K | $0.075 | $0.3 | | +| Qwen--Qwen3-32B-TEE | chutes | 40K | $0.08 | $0.24 | | +| qwen--qwen3-30b-a3b-fp8 | novitaai | 40K | $0.09 | $0.45 | | ## Free Agentic Models -Free models with both tool calling and reasoning — zero-cost agents. - -| Model | Provider | Context | Input $/M | Output $/M | Capabilities | -| ----- | -------- | ------- | --------- | ---------- | ------------ | - -## Cheapest Agentic Models - -Best value agentic models for production agents. - -| Model | Provider | Context | Input $/M | Output $/M | Capabilities | -| ----------------------- | ---------- | ------- | --------- | ---------- | ------------ | -| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | | -| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | | -| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | | -| openai--gpt-5-nano:flex | requesty | 400K | $0.025 | $0.2 | | -| gpt-5-nano | aihubmix | 0 | $0.025 | $0.2 | 📋 | -| gpt-oss-20b | cortecs | 0 | $0.027 | $0.124 | | -| qwen--qwen3-4b-fp8 | novitaai | 128K | $0.03 | $0.03 | | -| openai--gpt-oss-20b | openrouter | 131K | $0.03 | $0.14 | 📋 | -| qwen-3.5-4b | auriko | 262K | $0.03 | $0.15 | | -| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | 📋 | +| Model | Provider | Context | Vision | Structured Output | +| --------------------------------- | -------- | ------- | ------ | ----------------- | +| glm-4.1v-thinking-flash | zhipuai | 62K | ✅ | ✅ | +| qwen--qwen3-omni-30b-a3b-thinking | novitaai | 64K | ✅ | ✅ | +| qwen--qwen3.5-4b-free | mixlayer | 128K | | | ## Largest Context Agentic Models -Agentic models with the largest context windows — for complex multi-step tasks. - -| Model | Provider | Context | Input $/M | Output $/M | Capabilities | -| ----------------------------- | -------- | ------- | --------- | ---------- | ------------ | -| grok-4-20 | venice | 2M | $1.42 | $2.83 | 📋 | -| grok-4.20-beta-0309-reasoning | 302ai | 2M | $2 | $6 | | -| grok-4-1-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | -| grok-4-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | -| openai--gpt-5.5 | requesty | 1M | $5 | $30 | | -| openai-responses--gpt-5.4 | requesty | 1M | $2.5 | $15 | | -| openai-responses--gpt-5.5 | requesty | 1M | $5 | $30 | | -| openai--gpt-5.4 | requesty | 1M | $2.5 | $15 | | -| openai-responses--gpt-5.4-pro | requesty | 1M | $30 | $180 | | -| openai-responses--gpt-5.5-pro | requesty | 1M | $30 | $180 | | +| Model | Provider | Context | Input $/1M | Output $/1M | +| --------------------------- | ----------- | ------- | ---------- | ----------- | +| deepseek-v4-flash | baidu | 1024K | $0.126 | $0.252 | +| deepseek-v4-pro | siliconflow | 1024K | $1.74 | $3.48 | +| deepseek-v4-flash | siliconflow | 1024K | $0.14 | $0.28 | +| deepseek--deepseek-v4-pro | novitaai | 1024K | $1.67 | $3.38 | +| xiaomimimo--mimo-v2.5-pro | novitaai | 1024K | $2 | $6 | +| deepseek--deepseek-v4-flash | novitaai | 1024K | $0.14 | $0.28 | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | +| gemini-2.5-pro | google | 1024K | $1.25 | $10 | +| deepseek-v4-pro | fireworks | 1024K | $1.74 | $3.48 | +| xiaomi--mimo-v2.5-pro | hpc-ai | 1024K | $1 | $3 | +| xiaomi--mimo-v2.5 | hpc-ai | 1024K | $0.4 | $2 | +| mimo-v2.5-pro | xiaomi | 1024K | $1 | $3 | +| mimo-v2-pro | xiaomi | 1024K | $1 | $3 | +| minimaxai--minimax-m1-80k | novitaai | 976K | $0.55 | $2.2 | +| qwen3.6-flash | alibaba | 976K | $1.2 | $7.2 | + +## Building AI Agents + +| Agent Type | Recommended Models | Key Capabilities | +| ---------------- | --------------------------- | ------------------------------ | +| Code agents | Claude Sonnet 4, GPT-4.1 | Tool call + reasoning + code | +| Research agents | Gemini 2.5 Pro, DeepSeek R1 | Large context + reasoning | +| Data agents | Claude Sonnet 4, Qwen3 | Structured output + tool call | +| Vision agents | GPT-4o, Gemini 2.5 Flash | Vision + tool call + reasoning | +| Customer support | Llama 4 Scout, Qwen3 | Free/cheap + tool call | ## Related Documentation -- [Model Selection Guide](model-selection.md) — decision framework for choosing models -- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling -- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning -- [Code Models](code-models.md) — 189 code-focused models -- [Structured Output](structured-output.md) — 829 JSON-mode models -- [Free AI Models](free-models.md) — 81 free models by capability -- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [Tool Calling Models](tool-calling.md) — 2,350 models with function calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with chain-of-thought +- [Structured Output](structured-output.md) — 829 models with JSON mode +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4 alternatives with agent capabilities +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Free AI Models](free-models.md) — 81 models at zero cost --- diff --git a/docs/audio-models.md b/docs/audio-models.md index d0ef50bd..733db149 100644 --- a/docs/audio-models.md +++ b/docs/audio-models.md @@ -1,175 +1,83 @@ -# Audio Models +**English** | [中文](./zh/audio-models.md) -[中文](zh/audio-models.md) +# Audio AI Models — Speech, Voice, and Audio Understanding -AI models that support audio input or output, sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). +Models with **audio input** (speech recognition, voice understanding) and **audio output** (text-to-speech, voice generation) capabilities. All data sourced from first-party APIs via the [AI Models Catalog](https://github.com/i-need-token/ai-models). ## Stats -| Metric | Count | -| ------------------ | ----- | -| Total audio models | 136 | -| Audio input | 118 | -| Audio output | 34 | -| Free | 0 | -| Open weights | 7 | -| Providers | 20 | - -## Providers - -`aihubmix`, `amazon`, `amazon-bedrock`, `cortecs`, `digitalocean`, `fastrouter`, `google`, `google-vertex`, `groq`, `hpc-ai`, `microsoft`, `minimax`, `mistral`, `novitaai`, `openai`, `openrouter`, `scaleway`, `stepfun`, `vultr`, `xiaomi` - -## Models - -| Model | Provider | Context | Input Price | Capabilities | -| ---------------------------------------------------- | -------------- | ------- | ----------- | ------------------ | -| gemini-2.0-flash | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.0-flash-exp | aihubmix | — | $0.01 | 🎤 in | -| gemini-2.0-flash-lite | aihubmix | — | $0.038 | 🎤 in | -| gemini-2.0-flash-search | aihubmix | — | $0.05 | 🎤 in | -| gemini-2.0-flash-thinking-exp-01-21 | aihubmix | — | $0.038 | 🎤 in 🧠 | -| gemini-2.0-pro-exp-02-05 | aihubmix | — | $0.625 | 🎤 in | -| gemini-2.0-pro-exp-02-05-search | aihubmix | — | $0.625 | 🎤 in | -| gemini-2.5-flash | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-lite | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-search | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-pro | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | 🎤 in 🔧 | -| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | 🎤 in 🧠 | -| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | 🎤 in 🧠 | -| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-search | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-3.1-flash-lite-preview-nothink | aihubmix | — | $0.125 | 🎤 in 🔧 🧠 | -| gpt-4o-audio-preview | aihubmix | — | $1.25 | 🎤 in | -| gpt-4o-mini-audio-preview | aihubmix | — | $0.075 | 🎤 in | -| mimo-v2-omni | aihubmix | — | $0.22 | 🎤 in | -| mimo-v2-pro | aihubmix | — | $0.55 | 🎤 in | -| veo-3 | aihubmix | — | $1 | 🎤 in | -| veo3 | aihubmix | — | $1 | 🎤 in | -| amazon-nova-2.0-lite | amazon | — | $0.15 | 🎤 in 🔧 🧠 | -| amazon-nova-2.0-omni | amazon | — | $0.2 | 🎤 in 🔧 🧠 | -| amazon-nova-2.0-pro | amazon | — | $0.625 | 🎤 in 🔧 🧠 | -| mistral-voxtral-mini | amazon-bedrock | — | $0.04 | 🎤 in | -| mistral-voxtral-small | amazon-bedrock | — | $0.1 | 🎤 in | -| gemini-2.0-flash-001 | cortecs | — | $0.134 | 🎤 in 🔧 | -| gemini-2.0-flash-lite-001 | cortecs | — | $0.067 | 🎤 in 🔧 | -| gemini-2.5-flash | cortecs | — | $0.268 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro | cortecs | — | $1.342 | 🎤 in 🔧 🧠 | -| gemini-3.1-flash-lite | cortecs | — | $0.244 | 🎤 in 🔧 🧠 | -| voxtral-small-2507 | cortecs | — | $0.1 | 🎤 in | -| nemotron-3-nano-omni | digitalocean | — | $0.5 | 🎤 in 🔓 | -| google--gemini-3.1-flash-lite-preview | fastrouter | — | $0.25 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-pro-preview | fastrouter | — | $2 | 🎤 in 🔧 🧠 | -| openai--gpt-4o-realtime-preview-2024-12-17 | fastrouter | — | $5 | 🎤 in 🔊 out | -| openai--gpt-4o-realtime-preview-2025-06-03 | fastrouter | — | $5 | 🎤 in 🔊 out | -| openai--gpt-realtime | fastrouter | — | $4 | 🎤 in 🔊 out | -| openai--gpt-realtime-1.5 | fastrouter | — | $4 | 🎤 in 🔊 out | -| openai--gpt-realtime-mini | fastrouter | — | $0.6 | 🎤 in 🔊 out | -| chirp-3.0-HD | google | — | — | 🎤 in | -| gemini-1.5-flash | google | — | $0.075 | 🎤 in 🔧 | -| gemini-1.5-flash-8b | google | — | $0.075 | 🎤 in 🔧 | -| gemini-1.5-pro | google | — | $1.25 | 🎤 in 🔧 | -| gemini-2.0-flash | google | — | $0.1 | 🎤 in 🔧 | -| gemini-2.0-flash-lite | google | — | $0.075 | 🎤 in 🔧 | -| gemini-2.5-flash | google | — | $0.15 | 🎤 in 🔧 🧠 | -| gemini-2.5-flash-lite | google | — | $0.1 | 🎤 in 🔧 | -| gemini-2.5-pro | google | — | $1.25 | 🎤 in 🔧 🧠 | -| gemma-3n-E2B-it | google | — | — | 🎤 in | -| gemma-3n-E4B-it | google | — | — | 🎤 in | -| lyria-2.0 | google | — | — | 🔊 out | -| gemini-2-0-flash | google-vertex | — | $0.15 | 🎤 in 🔧 | -| gemini-2-0-flash-lite | google-vertex | — | $0.075 | 🎤 in 🔧 | -| gemini-2-5-pro | google-vertex | — | $1.25 | 🎤 in 🔧 | -| gemini-3-pro | google-vertex | — | $2 | 🎤 in 🔧 | -| orpheus-ar-sa | groq | — | — | 🔊 out | -| orpheus-en | groq | — | — | 🔊 out | -| whisper-large-v3 | groq | — | — | 🎤 in | -| whisper-large-v3-turbo | groq | — | — | 🎤 in | -| xiaomi--mimo-v2.5 | hpc-ai | — | $0.4 | 🎤 in 🔧 🧠 🔓 | -| microsoft-phi-4-mini-multimodal | microsoft | — | $0.08 | 🎤 in 🔧 | -| music-2.6 | minimax | — | — | 🔊 out | -| music-cover | minimax | — | — | 🔊 out | -| speech-02-hd | minimax | — | — | 🔊 out | -| speech-02-turbo | minimax | — | — | 🔊 out | -| speech-2.6-hd | minimax | — | — | 🔊 out | -| speech-2.6-turbo | minimax | — | — | 🔊 out | -| speech-2.8-hd | minimax | — | — | 🔊 out | -| speech-2.8-turbo | minimax | — | — | 🔊 out | -| voxtral-mini | mistral | — | $0.04 | 🎤 in | -| voxtral-small | mistral | — | $0.1 | 🎤 in | -| qwen--qwen3-omni-30b-a3b-instruct | novitaai | — | — | 🎤 in 🔊 out 🔧 | -| qwen--qwen3-omni-30b-a3b-thinking | novitaai | — | — | 🎤 in 🔧 🧠 | -| gpt-4o-audio | openai | — | $2.5 | 🎤 in 🔊 out 🔧 | -| gpt-4o-realtime | openai | — | $5 | 🎤 in 🔊 out 🔧 | -| o1-realtime | openai | — | $15 | 🎤 in 🔊 out 🔧 🧠 | -| tts-1 | openai | — | — | 🔊 out | -| tts-1-hd | openai | — | — | 🔊 out | -| whisper-1 | openai | — | — | 🎤 in | -| google--gemini-2.0-flash-001 | openrouter | — | $0.1 | 🎤 in 🔧 | -| google--gemini-2.0-flash-lite-001 | openrouter | — | $0.075 | 🎤 in 🔧 | -| google--gemini-2.5-flash | openrouter | — | $0.3 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-flash-lite | openrouter | — | $0.1 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-flash-lite-preview-09-2025 | openrouter | — | $0.1 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-pro | openrouter | — | $1.25 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-pro-preview | openrouter | — | $1.25 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-pro-preview-05-06 | openrouter | — | $1.25 | 🎤 in 🔧 🧠 | -| google--gemini-3-flash-preview | openrouter | — | $0.5 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-flash-lite | openrouter | — | $0.25 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-flash-lite-preview | openrouter | — | $0.25 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-pro-preview | openrouter | — | $2 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-pro-preview-customtools | openrouter | — | $2 | 🎤 in 🔧 🧠 | -| google--gemini-3.5-flash | openrouter | — | $1.5 | 🎤 in 🔧 🧠 | -| google--lyria-3-clip-preview | openrouter | — | — | 🔊 out | -| google--lyria-3-pro-preview | openrouter | — | — | 🔊 out | -| mistralai--voxtral-small-24b-2507 | openrouter | — | $0.1 | 🎤 in 🔧 | -| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | — | — | 🎤 in 🔧 🧠 | -| openai--gpt-4o-audio-preview | openrouter | — | $2.5 | 🎤 in 🔊 out 🔧 | -| openai--gpt-audio | openrouter | — | $2.5 | 🎤 in 🔊 out 🔧 | -| openai--gpt-audio-mini | openrouter | — | $0.6 | 🎤 in 🔊 out 🔧 | -| xiaomi--mimo-v2-omni | openrouter | — | $0.4 | 🎤 in 🔧 🧠 | -| xiaomi--mimo-v2.5 | openrouter | — | $0.4 | 🎤 in 🔧 🧠 | -| ~google--gemini-flash-latest | openrouter | — | $1.5 | 🎤 in 🔧 🧠 | -| ~google--gemini-pro-latest | openrouter | — | $2 | 🎤 in 🔧 🧠 | -| voxtral-small-24b-2507 | scaleway | — | $0.15 | 🎤 in 🔓 | -| step-1o-audio | stepfun | — | $25 | 🎤 in 🔊 out | -| step-asr | stepfun | — | — | 🎤 in | -| step-asr-1.1 | stepfun | — | — | 🎤 in | -| step-asr-1.1-stream | stepfun | — | — | 🎤 in | -| step-audio-2 | stepfun | — | $10 | 🎤 in 🔊 out | -| step-audio-r1.1 | stepfun | — | — | 🎤 in 🔊 out | -| step-tts-2 | stepfun | — | — | 🔊 out | -| step-tts-mini | stepfun | — | — | 🔊 out | -| stepaudio-2-asr-pro | stepfun | — | — | 🎤 in | -| stepaudio-2.5-asr | stepfun | — | — | 🎤 in | -| stepaudio-2.5-chat | stepfun | — | $10 | 🎤 in | -| stepaudio-2.5-realtime | stepfun | — | $10 | 🎤 in 🔊 out | -| stepaudio-2.5-tts | stepfun | — | — | 🔊 out | -| gemma-4-26b-a4b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| gemma-4-31b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| gemma-4-e2b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| gemma-4-e4b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| mimo-v2-omni | xiaomi | — | $0.4 | 🎤 in 🔧 | -| mimo-v2.5 | xiaomi | — | $0.4 | 🎤 in 🔧 | - -> 💡 Use the [interactive catalog](https://i-need-token.github.io/ai-models/) to search and filter models with more criteria. +| Metric | Count | +| ------------------- | ----- | +| Audio input models | 115 | +| Audio output models | 34 | +| Total audio models | 133 | +| Providers | 20 | +| With tool calling | 69 | +| Free | 8 | ---- +## Audio Input Models (Speech/Voice Understanding) -Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. +Models that can process audio input — transcribe speech, analyze audio, understand voice commands: + +| Model | Provider | Context | Input $/1M | Output $/1M | Tool Call | +| ------------------------------- | -------------- | ------- | ---------- | ----------- | --------- | +| mistral-voxtral-mini | amazon-bedrock | 125K | $0.04 | $0.04 | | +| voxtral-mini | mistral | 125K | $0.04 | $0.04 | | +| gemini-2.0-flash-lite | google | 1024K | $0.075 | $0.3 | ✅ | +| gemini-2-0-flash-lite | google-vertex | 1024K | $0.075 | $0.3 | ✅ | +| microsoft-phi-4-mini-multimodal | microsoft | 125K | $0.08 | $0.32 | ✅ | +| gemini-2.0-flash | google | 1024K | $0.1 | $0.4 | ✅ | +| gemini-2.5-flash-lite | google | 1024K | $0.1 | $0.4 | ✅ | +| mistral-voxtral-small | amazon-bedrock | 125K | $0.1 | $0.3 | | +| voxtral-small | mistral | 125K | $0.1 | $0.3 | | +| amazon-nova-2.0-lite | amazon | 62K | $0.15 | $1.25 | ✅ | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | ✅ | +| gemini-2-0-flash | google-vertex | 1024K | $0.15 | $0.6 | ✅ | +| voxtral-small-24b-2507 | scaleway | 128K | $0.15 | $0.35 | | +| amazon-nova-2.0-omni | amazon | 62K | $0.2 | $1.3 | ✅ | +| xiaomi--mimo-v2.5 | hpc-ai | 1024K | $0.4 | $2 | ✅ | + +## Audio Output Models (Text-to-Speech / Voice Generation) + +Models that can generate audio output — text-to-speech, voice synthesis: + +| Model | Provider | Input $/1M | Output $/1M | +| ---------------------- | -------- | ---------- | ----------- | +| gpt-4o-audio | openai | $2.5 | $10 | +| gpt-4o-realtime | openai | $5 | $20 | +| step-audio-2 | stepfun | $10 | $70 | +| stepaudio-2.5-realtime | stepfun | $10 | $70 | +| o1-realtime | openai | $15 | $60 | +| step-1o-audio | stepfun | $25 | $60 | + +## Free Audio Models + +| Model | Provider | Audio In | Audio Out | Context | +| --------------------------------- | -------- | -------- | --------- | ------- | +| qwen--qwen3-omni-30b-a3b-instruct | novitaai | ✅ | ✅ | 64K | +| qwen--qwen3-omni-30b-a3b-thinking | novitaai | ✅ | | 64K | +| gemma-3n-E2B-it | google | ✅ | | 128K | +| gemma-3n-E4B-it | google | ✅ | | 128K | +| step-audio-r1.1 | stepfun | ✅ | ✅ | 0 | + +## Choosing an Audio Model + +| Use Case | Recommended | Key Feature | +| ------------------- | ------------------------- | ------------------------- | +| Voice assistants | GPT-4o-audio | Audio in/out + tool call | +| Transcription | Whisper, Gemini 2.5 Flash | Fast, cheap audio input | +| Voice agents | GPT-4o-realtime | Real-time audio streaming | +| Text-to-speech | OpenAI TTS, ElevenLabs | Natural voice output | +| Multilingual speech | Gemini 2.5 Pro | 100+ languages | ## Related Documentation -- [Video Models](video-models.md) — 167 video input/output models -- [Vision Models](vision-models.md) — 1,487 vision models -- [Modality Matrix](modality-matrix.md) — all modalities at a glance -- [Model Selection Guide](model-selection.md) — decision framework -- [Free AI Models](free-models.md) — 81 free models +- [Multimodal Models](multimodal-models.md) — 1,519 models with image/audio/video +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Chat Models](chat-models.md) — 2,350 models with tool calling +- [Tool Calling Models](tool-calling.md) — function calling capabilities +- [Free AI Models](free-models.md) — 81 models at zero cost + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/code-models.md b/docs/code-models.md index 3ae25403..05c7828f 100644 --- a/docs/code-models.md +++ b/docs/code-models.md @@ -1,89 +1,60 @@ -# Code Models +**English** | [中文](./zh/code-models.md) -[中文](zh/code-models.md) +# Code AI Models — 189 Models for Code Generation & Understanding -AI models specifically designed or optimized for code generation, code understanding, and software development tasks. Includes models from 41 providers with code-focused capabilities. - -Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). - -## Why Code Models Matter - -Code-focused models are optimized for: - -- **Code generation** — write functions, classes, and entire programs from descriptions -- **Code completion** — autocomplete suggestions in IDEs and editors -- **Code review** — identify bugs, security issues, and style violations -- **Code explanation** — understand and document existing code -- **Code translation** — convert code between programming languages -- **Test generation** — write unit tests and integration tests automatically +Models optimized for **code generation, code review, debugging, and software engineering** tasks. All data sourced from first-party APIs via the [AI Models Catalog](https://github.com/i-need-token/ai-models). ## Stats -| Metric | Count | -| ----------------------- | ----- | -| Code-focused models | 189 | -| Providers | 41 | -| Free code models | 0 | -| Open-weight code models | 14 | -| With reasoning | 43 | -| With tool calling | 119 | -| With structured output | 43 | +| Metric | Count | +| ------------------- | ----- | +| Code-focused models | 0 | +| Providers | 0 | +| With tool calling | 0 | +| With reasoning | 0 | +| With vision | 0 | +| Free | 0 | +| Open weights | 0 | -## Providers +## Cheapest Code Models (Direct Providers) -`302ai`, `aihubmix`, `aimlapi`, `alibaba`, `amazon-bedrock`, `arcee`, `auriko`, `cerebras`, `clarifai`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `fastrouter`, `gmicloud`, `google-vertex`, `hyperbolic`, `inception`, `inferencenet` and 21 more +| Model | Provider | Context | Input $/1M | Output $/1M | Reasoning | +| ----- | -------- | ------- | ---------- | ----------- | --------- | -## Free Code Models +## Code Models with Reasoning (Best for Complex Tasks) -Code models available at zero cost — perfect for development and prototyping. +| Model | Provider | Context | Input $/1M | Output $/1M | Tool Call | +| ----- | -------- | ------- | ---------- | ----------- | --------- | -| Model | Provider | Context | Input $/M | Output $/M | Capabilities | -| ----- | -------- | ------- | --------- | ---------- | ------------ | +## Free Code Models -## Cheapest Code Models +| Model | Provider | Context | Reasoning | Tool Call | +| ----- | -------- | ------- | --------- | --------- | -Best value code models for production use. +## Largest Context Code Models -| Model | Provider | Context | Input $/M | Output $/M | Capabilities | -| ------------------------------------------- | ------------ | ------- | --------- | ---------- | ------------ | -| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔧 🔓 | -| doubao-seed-code-preview-latest | 302ai | 131K | $0.0515 | $0.343 | 🔧 | -| qwen3-coder-30b-a3b-instruct | cortecs | 0 | $0.053 | $0.222 | 🧠 🔧 | -| qwen--qwen2.5-coder-32b-instruct | fastrouter | 32K | $0.06 | $0.15 | | -| qwen3-coder-flash | aihubmix | 0 | $0.068 | $0.272 | 🔧 📋 | -| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | 0 | $0.07 | $0.16 | 🔧 | -| Qwen3-Coder-30B-A3B-Instruct | ovhcloud | 262K | $0.07 | $0.26 | | -| qwen--qwen3-coder-30b-a3b-instruct | martian | 0 | $0.07 | $0.27 | | -| qwen--qwen3-coder-30b-a3b-instruct | novitaai | 160K | $0.07 | $0.27 | 🔧 📋 | -| qwen--qwen3-coder-30b-a3b-instruct | openrouter | 160K | $0.07 | $0.27 | 🔧 📋 | +| Model | Provider | Context | Input $/1M | Output $/1M | +| ----- | -------- | ------- | ---------- | ----------- | -## Largest Context Code Models +## Choosing a Code Model -Code models with the largest context windows — for working with large codebases. - -| Model | Provider | Context | Input $/M | Output $/M | Capabilities | -| -------------------------- | ---------- | ------- | --------- | ---------- | ------------ | -| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 | -| qwen--qwen3-coder | openrouter | 1M | $0.22 | $1.8 | 🔧 📋 | -| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | 🔧 | -| alibaba--qwen3-coder-flash | requesty | 1M | $0.3 | $1.5 | 🔧 | -| alibaba--qwen3-coder-plus | requesty | 1M | $1 | $5 | 🔧 | -| gpt-5-3-codex | meganova | 1M | $1.4 | $11.2 | 🔧 | -| qwen--qwen3-coder-flash | openrouter | 1M | $0.195 | $0.975 | 🔧 📋 | -| qwen--qwen3-coder-plus | openrouter | 1M | $0.65 | $3.25 | 🔧 📋 | -| gpt-5-codex | 302ai | 400K | $1.25 | $10 | 🔧 | -| gpt-5.1-codex | 302ai | 400K | $1.25 | $10 | 🔧 | +| Use Case | Recommended | Why | +| ------------ | -------------------------- | --------------------- | +| Autocomplete | Qwen2.5-Coder, DeepSeek V3 | Fast, cheap | +| Code review | Claude Sonnet 4, GPT-4.1 | Reasoning + tool call | +| Bug fixing | DeepSeek R1, Claude Opus 4 | Deep reasoning | +| Refactoring | Gemini 2.5 Pro | Large context | +| Code agents | Claude Sonnet 4, GPT-4.1 | Tool call + reasoning | +| Self-hosted | Qwen2.5-Coder, Llama 4 | Open weights | ## Related Documentation -- [Model Selection Guide](model-selection.md) — decision framework for choosing models -- [Tool Calling Models](tool-calling.md) — 2,350 models with tool calling -- [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning -- [Structured Output](structured-output.md) — 829 JSON-mode models -- [Free AI Models](free-models.md) — 81 free models by capability -- [Open-Weight Models](open-weights.md) — 527 models you can run yourself -- [Context Window Comparison](context-windows.md) — largest context windows -- [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [Reasoning Models](reasoning-models.md) — 1,306 models with chain-of-thought +- [Tool Calling Models](tool-calling.md) — 2,350 models with function calling +- [Agentic Models](agentic-models.md) — 1,080 models for AI agents +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4 alternatives +- [Free AI Models](free-models.md) — 81 models at zero cost +- [Open-Weight Models](open-weights.md) — 527 models with open weights --- diff --git a/docs/openai-alternatives.md b/docs/openai-alternatives.md index 94cea443..be13fdea 100644 --- a/docs/openai-alternatives.md +++ b/docs/openai-alternatives.md @@ -149,6 +149,7 @@ Models with tool calling at zero cost (direct providers): - [Migration Guide](migration-guide.md) — switch providers with minimal code changes - [Provider Comparison](provider-comparison.md) — top 30 providers by model count - [Agentic Models](agentic-models.md) — 1,080 models for AI agents +- [Code Models](code-models.md) — 189 code-focused models --- diff --git a/docs/reasoning-models.md b/docs/reasoning-models.md index 8b751500..8703ba33 100644 --- a/docs/reasoning-models.md +++ b/docs/reasoning-models.md @@ -96,6 +96,8 @@ Models that can reason about images — ideal for visual analysis: - [Free AI Models](free-models.md) — 81 free models, some with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching - [OpenAI Alternatives](openai-alternatives.md) — GPT-4 class alternatives with reasoning +- [Agentic Models](agentic-models.md) — 1,080 models with tool_call + reasoning for agents +- [Code Models](code-models.md) — 189 code-focused models --- diff --git a/docs/structured-output.md b/docs/structured-output.md index 501b4e6e..efabf5a6 100644 --- a/docs/structured-output.md +++ b/docs/structured-output.md @@ -77,6 +77,7 @@ For AI agents that need to return structured data, call tools, and reason: - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning - [Free AI Models](free-models.md) — 81 free models, some with structured output - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching +- [Agentic Models](agentic-models.md) — 1,080 models with tool_call + reasoning for agents --- diff --git a/docs/tool-calling.md b/docs/tool-calling.md index 9c1fade8..2b70a89d 100644 --- a/docs/tool-calling.md +++ b/docs/tool-calling.md @@ -95,6 +95,7 @@ The "holy trinity" for advanced AI agents — tool calling, reasoning, and visio - [Reasoning Models](reasoning-models.md) — 1,306 models with reasoning - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching - [OpenAI Alternatives](openai-alternatives.md) — GPT-4/GPT-3.5 alternatives with tool calling +- [Agentic Models](agentic-models.md) — 1,080 models with tool_call + reasoning for agents --- diff --git a/docs/vision-models.md b/docs/vision-models.md index 7fab083e..7d43b4e9 100644 --- a/docs/vision-models.md +++ b/docs/vision-models.md @@ -94,6 +94,7 @@ The most capable vision models — can see, reason, and act: - [Modality Matrix](modality-matrix.md) — all modalities at a glance - [Free AI Models](free-models.md) — 81 free models, some with vision - [OpenAI Alternatives](openai-alternatives.md) — GPT-4 class alternatives with vision +- [Audio Models](audio-models.md) — 118 audio input + 34 audio output models --- diff --git a/docs/zh/agentic-models.md b/docs/zh/agentic-models.md index 714ff25d..c83aa7c8 100644 --- a/docs/zh/agentic-models.md +++ b/docs/zh/agentic-models.md @@ -1,90 +1,96 @@ -# 智能体模型 +[English](../agentic-models.md) | **中文** -[English](../agentic-models.md) +# 智能体 AI 模型 — 1,080 个用于 AI 智能体的模型 -同时具备**工具调用**和**推理**能力的 AI 模型 — 构建 AI Agent 的关键要求。这些模型可以规划、推理多步骤任务，并使用外部工具完成目标。 +具备 **工具调用 + 推理** 能力的模型 — AI 智能体、自主工作流和多步任务执行的核心构建块。所有数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) 一手 API。 -数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 +## 什么使模型成为"智能体"？ -## 为什么智能体模型很重要 +智能体模型可以： -智能体模型结合了两个关键能力： +- **调用工具/函数** — 与 API、数据库和外部系统交互 +- **逐步推理** — 规划、分解任务和验证结果 +- **自主行动** — 无需人工干预即可做出决策和采取行动 -- **工具调用** — 调用函数、API 和外部工具来执行操作 -- **推理** — 链式思维思考，规划和分解复杂任务 - -两者结合可以实现： - -- **自主智能体** — 独立规划、行动和迭代的模型 -- **多步骤工作流** — 将复杂任务分解为子任务并使用工具 -- **自我纠正** — 推理失败原因并使用不同方法重试 -- **代码智能体** — 自主编写、执行和调试代码 -- **研究智能体** — 搜索、综合和总结信息 +这些能力是聊天机器人和 AI 智能体的关键区别。 ## 统计 -| 指标 | 数量 | -| ----------------------------- | ---- | -| 智能体模型（工具调用 + 推理） | 1080 | -| 提供商 | 51 | -| 免费智能体模型 | 0 | -| 开源权重智能体模型 | 65 | -| 带结构化输出 | 455 | - -## 提供商 - -`302ai`, `aihubmix`, `alibaba`, `amazon`, `anthropic`, `arcee`, `auriko`, `baidu`, `baseten`, `bytedance`, `chutes`, `clarifai`, `cloudflare`, `cortecs`, `deepseek`, `digitalocean`, `dinference`, `fastrouter`, `fireworks`, `google`, `hpc-ai`, `inclusionai`, `inferencenet`, `klusterai`, `llmgateway` 等 26 个 +| Metric | Count | +| ---------------------- | ----- | +| Agentic models | 1076 | +| Providers | 51 | +| With vision | 617 | +| With structured output | 453 | +| Free | 30 | +| Open weights | 64 | + +## 最便宜的智能体模型（直连提供商） + +| Model | Provider | Context | Input $/1M | Output $/1M | Vision | +| ------------------------------ | ------------ | ------- | ---------- | ----------- | ------ | +| qwen--qwen3-4b-fp8 | novitaai | 125K | $0.03 | $0.03 | | +| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | | +| openai--gpt-oss-120b | novitaai | 128K | $0.05 | $0.25 | ✅ | +| gpt-oss-120b | inferencenet | 128K | $0.05 | $0.45 | | +| Qwen--Qwen3.6-35B-A3B | neuralwatt | 0 | $0.05 | $0.1 | ✅ | +| qwen3-30b-a3b-fp8 | cloudflare | 40K | $0.051 | $0.335 | | +| glm-4.7-flash | cloudflare | 128K | $0.06 | $0.4 | | +| Nemotron-3-Nano-Omni | nebius | 125K | $0.06 | $0.24 | ✅ | +| hermes-4-llama-3.1-8b | nousresearch | 128K | $0.06 | $0.12 | | +| zai-org--glm-4.7-flash | novitaai | 195K | $0.07 | $0.4 | | +| ring-2.6-1t | inclusionai | 256K | $0.07 | $0.62 | | +| seed-1.6-flash | bytedance | 256K | $0.07 | $0.3 | ✅ | +| microsoft-phi-4-mini-reasoning | microsoft | 125K | $0.075 | $0.3 | | +| Qwen--Qwen3-32B-TEE | chutes | 40K | $0.08 | $0.24 | | +| qwen--qwen3-30b-a3b-fp8 | novitaai | 40K | $0.09 | $0.45 | | ## 免费智能体模型 -同时具备工具调用和推理的免费模型 — 零成本 Agent。 - -| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | -| ---- | ------ | ------ | -------- | -------- | ---- | - -## 最便宜智能体模型 - -生产环境 Agent 的最佳性价比模型。 - -| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | -| ----------------------- | ---------- | ------ | -------- | -------- | ---- | -| qwen-3.5-0.8b | auriko | 262K | $0.01 | $0.05 | | -| llama-3.1-8b-instruct | cortecs | 0 | $0.018 | $0.054 | | -| qwen-3.5-2b | auriko | 262K | $0.02 | $0.1 | | -| openai--gpt-5-nano:flex | requesty | 400K | $0.025 | $0.2 | | -| gpt-5-nano | aihubmix | 0 | $0.025 | $0.2 | 📋 | -| gpt-oss-20b | cortecs | 0 | $0.027 | $0.124 | | -| qwen--qwen3-4b-fp8 | novitaai | 128K | $0.03 | $0.03 | | -| openai--gpt-oss-20b | openrouter | 131K | $0.03 | $0.14 | 📋 | -| qwen-3.5-4b | auriko | 262K | $0.03 | $0.15 | | -| openai--gpt-oss-20b | neuralwatt | 0 | $0.03 | $0.16 | 📋 | - -## 大上下文智能体模型 - -上下文窗口最大的智能体模型 — 适用于复杂多步骤任务。 - -| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | -| ----------------------------- | -------- | ------ | -------- | -------- | ---- | -| grok-4-20 | venice | 2M | $1.42 | $2.83 | 📋 | -| grok-4.20-beta-0309-reasoning | 302ai | 2M | $2 | $6 | | -| grok-4-1-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | -| grok-4-fast-reasoning | 302ai | 2M | $0.2 | $0.5 | | -| openai--gpt-5.5 | requesty | 1M | $5 | $30 | | -| openai-responses--gpt-5.4 | requesty | 1M | $2.5 | $15 | | -| openai-responses--gpt-5.5 | requesty | 1M | $5 | $30 | | -| openai--gpt-5.4 | requesty | 1M | $2.5 | $15 | | -| openai-responses--gpt-5.4-pro | requesty | 1M | $30 | $180 | | -| openai-responses--gpt-5.5-pro | requesty | 1M | $30 | $180 | | +| Model | Provider | Context | Vision | Structured Output | +| --------------------------------- | -------- | ------- | ------ | ----------------- | +| glm-4.1v-thinking-flash | zhipuai | 62K | ✅ | ✅ | +| qwen--qwen3-omni-30b-a3b-thinking | novitaai | 64K | ✅ | ✅ | +| qwen--qwen3.5-4b-free | mixlayer | 128K | | | + +## 最大上下文智能体模型 + +| Model | Provider | Context | Input $/1M | Output $/1M | +| --------------------------- | ----------- | ------- | ---------- | ----------- | +| deepseek-v4-flash | baidu | 1024K | $0.126 | $0.252 | +| deepseek-v4-pro | siliconflow | 1024K | $1.74 | $3.48 | +| deepseek-v4-flash | siliconflow | 1024K | $0.14 | $0.28 | +| deepseek--deepseek-v4-pro | novitaai | 1024K | $1.67 | $3.38 | +| xiaomimimo--mimo-v2.5-pro | novitaai | 1024K | $2 | $6 | +| deepseek--deepseek-v4-flash | novitaai | 1024K | $0.14 | $0.28 | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | +| gemini-2.5-pro | google | 1024K | $1.25 | $10 | +| deepseek-v4-pro | fireworks | 1024K | $1.74 | $3.48 | +| xiaomi--mimo-v2.5-pro | hpc-ai | 1024K | $1 | $3 | +| xiaomi--mimo-v2.5 | hpc-ai | 1024K | $0.4 | $2 | +| mimo-v2.5-pro | xiaomi | 1024K | $1 | $3 | +| mimo-v2-pro | xiaomi | 1024K | $1 | $3 | +| minimaxai--minimax-m1-80k | novitaai | 976K | $0.55 | $2.2 | +| qwen3.6-flash | alibaba | 976K | $1.2 | $7.2 | + +## 构建 AI 智能体 + +| Agent Type | Recommended Models | Key Capabilities | +| ---------------- | --------------------------- | ------------------------------ | +| Code agents | Claude Sonnet 4, GPT-4.1 | Tool call + reasoning + code | +| Research agents | Gemini 2.5 Pro, DeepSeek R1 | Large context + reasoning | +| Data agents | Claude Sonnet 4, Qwen3 | Structured output + tool call | +| Vision agents | GPT-4o, Gemini 2.5 Flash | Vision + tool call + reasoning | +| Customer support | Llama 4 Scout, Qwen3 | Free/cheap + tool call | ## 相关文档 -- [模型选择指南](model-selection.md) — 选择模型的决策框架 -- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 -- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 -- [代码模型](code-models.md) — 189 个代码模型 -- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 -- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 -- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [Tool Calling Models](tool-calling.md) — 2,350 models with function calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with chain-of-thought +- [Structured Output](structured-output.md) — 829 models with JSON mode +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4 alternatives with agent capabilities +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Free AI Models](free-models.md) — 81 models at zero cost --- diff --git a/docs/zh/audio-models.md b/docs/zh/audio-models.md index f01dd1f3..24379b44 100644 --- a/docs/zh/audio-models.md +++ b/docs/zh/audio-models.md @@ -1,175 +1,83 @@ -# 音频模型 +[English](../audio-models.md) | **中文** -[English](../audio-models.md) +# 音频 AI 模型 — 语音、声音和音频理解 -支持音频输入或输出的 AI 模型，数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 +具备 **音频输入**（语音识别、语音理解）和 **音频输出**（文本转语音、语音生成）能力的模型。所有数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) 一手 API。 ## 统计 -| 指标 | 数量 | -| ------------ | ---- | -| 音频模型总数 | 136 | -| 音频输入 | 118 | -| 音频输出 | 34 | -| 免费 | 0 | -| 开源权重 | 7 | -| 提供商 | 20 | - -## 提供商 - -`aihubmix`, `amazon`, `amazon-bedrock`, `cortecs`, `digitalocean`, `fastrouter`, `google`, `google-vertex`, `groq`, `hpc-ai`, `microsoft`, `minimax`, `mistral`, `novitaai`, `openai`, `openrouter`, `scaleway`, `stepfun`, `vultr`, `xiaomi` - -## 模型 - -| 模型 | 提供商 | 上下文 | 输入价格 | 能力 | -| ---------------------------------------------------- | -------------- | ------ | -------- | ------------------ | -| gemini-2.0-flash | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.0-flash-exp | aihubmix | — | $0.01 | 🎤 in | -| gemini-2.0-flash-lite | aihubmix | — | $0.038 | 🎤 in | -| gemini-2.0-flash-search | aihubmix | — | $0.05 | 🎤 in | -| gemini-2.0-flash-thinking-exp-01-21 | aihubmix | — | $0.038 | 🎤 in 🧠 | -| gemini-2.0-pro-exp-02-05 | aihubmix | — | $0.625 | 🎤 in | -| gemini-2.0-pro-exp-02-05-search | aihubmix | — | $0.625 | 🎤 in | -| gemini-2.5-flash | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-lite | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-lite-nothink | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-lite-preview-09-2025 | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-lite-preview-09-2025-nothink | aihubmix | — | $0.05 | 🎤 in 🔧 | -| gemini-2.5-flash-nothink | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-preview-05-20-nothink | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-preview-05-20-search | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-preview-09-2025 | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-flash-search | aihubmix | — | $0.15 | 🎤 in 🔧 | -| gemini-2.5-pro | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-exp-03-25 | aihubmix | — | $0.625 | 🎤 in 🔧 | -| gemini-2.5-pro-preview-03-25-search | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-preview-05-06 | aihubmix | — | $0.625 | 🎤 in 🧠 | -| gemini-2.5-pro-preview-05-06-search | aihubmix | — | $0.625 | 🎤 in 🧠 | -| gemini-2.5-pro-preview-06-05 | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-preview-06-05-search | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro-search | aihubmix | — | $0.625 | 🎤 in 🔧 🧠 | -| gemini-3.1-flash-lite-preview-nothink | aihubmix | — | $0.125 | 🎤 in 🔧 🧠 | -| gpt-4o-audio-preview | aihubmix | — | $1.25 | 🎤 in | -| gpt-4o-mini-audio-preview | aihubmix | — | $0.075 | 🎤 in | -| mimo-v2-omni | aihubmix | — | $0.22 | 🎤 in | -| mimo-v2-pro | aihubmix | — | $0.55 | 🎤 in | -| veo-3 | aihubmix | — | $1 | 🎤 in | -| veo3 | aihubmix | — | $1 | 🎤 in | -| amazon-nova-2.0-lite | amazon | — | $0.15 | 🎤 in 🔧 🧠 | -| amazon-nova-2.0-omni | amazon | — | $0.2 | 🎤 in 🔧 🧠 | -| amazon-nova-2.0-pro | amazon | — | $0.625 | 🎤 in 🔧 🧠 | -| mistral-voxtral-mini | amazon-bedrock | — | $0.04 | 🎤 in | -| mistral-voxtral-small | amazon-bedrock | — | $0.1 | 🎤 in | -| gemini-2.0-flash-001 | cortecs | — | $0.134 | 🎤 in 🔧 | -| gemini-2.0-flash-lite-001 | cortecs | — | $0.067 | 🎤 in 🔧 | -| gemini-2.5-flash | cortecs | — | $0.268 | 🎤 in 🔧 🧠 | -| gemini-2.5-pro | cortecs | — | $1.342 | 🎤 in 🔧 🧠 | -| gemini-3.1-flash-lite | cortecs | — | $0.244 | 🎤 in 🔧 🧠 | -| voxtral-small-2507 | cortecs | — | $0.1 | 🎤 in | -| nemotron-3-nano-omni | digitalocean | — | $0.5 | 🎤 in 🔓 | -| google--gemini-3.1-flash-lite-preview | fastrouter | — | $0.25 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-pro-preview | fastrouter | — | $2 | 🎤 in 🔧 🧠 | -| openai--gpt-4o-realtime-preview-2024-12-17 | fastrouter | — | $5 | 🎤 in 🔊 out | -| openai--gpt-4o-realtime-preview-2025-06-03 | fastrouter | — | $5 | 🎤 in 🔊 out | -| openai--gpt-realtime | fastrouter | — | $4 | 🎤 in 🔊 out | -| openai--gpt-realtime-1.5 | fastrouter | — | $4 | 🎤 in 🔊 out | -| openai--gpt-realtime-mini | fastrouter | — | $0.6 | 🎤 in 🔊 out | -| chirp-3.0-HD | google | — | — | 🎤 in | -| gemini-1.5-flash | google | — | $0.075 | 🎤 in 🔧 | -| gemini-1.5-flash-8b | google | — | $0.075 | 🎤 in 🔧 | -| gemini-1.5-pro | google | — | $1.25 | 🎤 in 🔧 | -| gemini-2.0-flash | google | — | $0.1 | 🎤 in 🔧 | -| gemini-2.0-flash-lite | google | — | $0.075 | 🎤 in 🔧 | -| gemini-2.5-flash | google | — | $0.15 | 🎤 in 🔧 🧠 | -| gemini-2.5-flash-lite | google | — | $0.1 | 🎤 in 🔧 | -| gemini-2.5-pro | google | — | $1.25 | 🎤 in 🔧 🧠 | -| gemma-3n-E2B-it | google | — | — | 🎤 in | -| gemma-3n-E4B-it | google | — | — | 🎤 in | -| lyria-2.0 | google | — | — | 🔊 out | -| gemini-2-0-flash | google-vertex | — | $0.15 | 🎤 in 🔧 | -| gemini-2-0-flash-lite | google-vertex | — | $0.075 | 🎤 in 🔧 | -| gemini-2-5-pro | google-vertex | — | $1.25 | 🎤 in 🔧 | -| gemini-3-pro | google-vertex | — | $2 | 🎤 in 🔧 | -| orpheus-ar-sa | groq | — | — | 🔊 out | -| orpheus-en | groq | — | — | 🔊 out | -| whisper-large-v3 | groq | — | — | 🎤 in | -| whisper-large-v3-turbo | groq | — | — | 🎤 in | -| xiaomi--mimo-v2.5 | hpc-ai | — | $0.4 | 🎤 in 🔧 🧠 🔓 | -| microsoft-phi-4-mini-multimodal | microsoft | — | $0.08 | 🎤 in 🔧 | -| music-2.6 | minimax | — | — | 🔊 out | -| music-cover | minimax | — | — | 🔊 out | -| speech-02-hd | minimax | — | — | 🔊 out | -| speech-02-turbo | minimax | — | — | 🔊 out | -| speech-2.6-hd | minimax | — | — | 🔊 out | -| speech-2.6-turbo | minimax | — | — | 🔊 out | -| speech-2.8-hd | minimax | — | — | 🔊 out | -| speech-2.8-turbo | minimax | — | — | 🔊 out | -| voxtral-mini | mistral | — | $0.04 | 🎤 in | -| voxtral-small | mistral | — | $0.1 | 🎤 in | -| qwen--qwen3-omni-30b-a3b-instruct | novitaai | — | — | 🎤 in 🔊 out 🔧 | -| qwen--qwen3-omni-30b-a3b-thinking | novitaai | — | — | 🎤 in 🔧 🧠 | -| gpt-4o-audio | openai | — | $2.5 | 🎤 in 🔊 out 🔧 | -| gpt-4o-realtime | openai | — | $5 | 🎤 in 🔊 out 🔧 | -| o1-realtime | openai | — | $15 | 🎤 in 🔊 out 🔧 🧠 | -| tts-1 | openai | — | — | 🔊 out | -| tts-1-hd | openai | — | — | 🔊 out | -| whisper-1 | openai | — | — | 🎤 in | -| google--gemini-2.0-flash-001 | openrouter | — | $0.1 | 🎤 in 🔧 | -| google--gemini-2.0-flash-lite-001 | openrouter | — | $0.075 | 🎤 in 🔧 | -| google--gemini-2.5-flash | openrouter | — | $0.3 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-flash-lite | openrouter | — | $0.1 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-flash-lite-preview-09-2025 | openrouter | — | $0.1 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-pro | openrouter | — | $1.25 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-pro-preview | openrouter | — | $1.25 | 🎤 in 🔧 🧠 | -| google--gemini-2.5-pro-preview-05-06 | openrouter | — | $1.25 | 🎤 in 🔧 🧠 | -| google--gemini-3-flash-preview | openrouter | — | $0.5 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-flash-lite | openrouter | — | $0.25 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-flash-lite-preview | openrouter | — | $0.25 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-pro-preview | openrouter | — | $2 | 🎤 in 🔧 🧠 | -| google--gemini-3.1-pro-preview-customtools | openrouter | — | $2 | 🎤 in 🔧 🧠 | -| google--gemini-3.5-flash | openrouter | — | $1.5 | 🎤 in 🔧 🧠 | -| google--lyria-3-clip-preview | openrouter | — | — | 🔊 out | -| google--lyria-3-pro-preview | openrouter | — | — | 🔊 out | -| mistralai--voxtral-small-24b-2507 | openrouter | — | $0.1 | 🎤 in 🔧 | -| nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--free | openrouter | — | — | 🎤 in 🔧 🧠 | -| openai--gpt-4o-audio-preview | openrouter | — | $2.5 | 🎤 in 🔊 out 🔧 | -| openai--gpt-audio | openrouter | — | $2.5 | 🎤 in 🔊 out 🔧 | -| openai--gpt-audio-mini | openrouter | — | $0.6 | 🎤 in 🔊 out 🔧 | -| xiaomi--mimo-v2-omni | openrouter | — | $0.4 | 🎤 in 🔧 🧠 | -| xiaomi--mimo-v2.5 | openrouter | — | $0.4 | 🎤 in 🔧 🧠 | -| ~google--gemini-flash-latest | openrouter | — | $1.5 | 🎤 in 🔧 🧠 | -| ~google--gemini-pro-latest | openrouter | — | $2 | 🎤 in 🔧 🧠 | -| voxtral-small-24b-2507 | scaleway | — | $0.15 | 🎤 in 🔓 | -| step-1o-audio | stepfun | — | $25 | 🎤 in 🔊 out | -| step-asr | stepfun | — | — | 🎤 in | -| step-asr-1.1 | stepfun | — | — | 🎤 in | -| step-asr-1.1-stream | stepfun | — | — | 🎤 in | -| step-audio-2 | stepfun | — | $10 | 🎤 in 🔊 out | -| step-audio-r1.1 | stepfun | — | — | 🎤 in 🔊 out | -| step-tts-2 | stepfun | — | — | 🔊 out | -| step-tts-mini | stepfun | — | — | 🔊 out | -| stepaudio-2-asr-pro | stepfun | — | — | 🎤 in | -| stepaudio-2.5-asr | stepfun | — | — | 🎤 in | -| stepaudio-2.5-chat | stepfun | — | $10 | 🎤 in | -| stepaudio-2.5-realtime | stepfun | — | $10 | 🎤 in 🔊 out | -| stepaudio-2.5-tts | stepfun | — | — | 🔊 out | -| gemma-4-26b-a4b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| gemma-4-31b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| gemma-4-e2b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| gemma-4-e4b-it | vultr | — | $0.55 | 🎤 in 🔓 | -| mimo-v2-omni | xiaomi | — | $0.4 | 🎤 in 🔧 | -| mimo-v2.5 | xiaomi | — | $0.4 | 🎤 in 🔧 | - -> 💡 使用[交互式目录](https://i-need-token.github.io/ai-models/)搜索和筛选更多条件的模型。 +| Metric | Count | +| ------------------- | ----- | +| Audio input models | 115 | +| Audio output models | 34 | +| Total audio models | 133 | +| Providers | 20 | +| With tool calling | 69 | +| Free | 8 | ---- +## 音频输入模型（语音/声音理解） -数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 +可以处理音频输入的模型 — 转录语音、分析音频、理解语音命令： + +| Model | Provider | Context | Input $/1M | Output $/1M | Tool Call | +| ------------------------------- | -------------- | ------- | ---------- | ----------- | --------- | +| mistral-voxtral-mini | amazon-bedrock | 125K | $0.04 | $0.04 | | +| voxtral-mini | mistral | 125K | $0.04 | $0.04 | | +| gemini-2.0-flash-lite | google | 1024K | $0.075 | $0.3 | ✅ | +| gemini-2-0-flash-lite | google-vertex | 1024K | $0.075 | $0.3 | ✅ | +| microsoft-phi-4-mini-multimodal | microsoft | 125K | $0.08 | $0.32 | ✅ | +| gemini-2.0-flash | google | 1024K | $0.1 | $0.4 | ✅ | +| gemini-2.5-flash-lite | google | 1024K | $0.1 | $0.4 | ✅ | +| mistral-voxtral-small | amazon-bedrock | 125K | $0.1 | $0.3 | | +| voxtral-small | mistral | 125K | $0.1 | $0.3 | | +| amazon-nova-2.0-lite | amazon | 62K | $0.15 | $1.25 | ✅ | +| gemini-2.5-flash | google | 1024K | $0.15 | $3.5 | ✅ | +| gemini-2-0-flash | google-vertex | 1024K | $0.15 | $0.6 | ✅ | +| voxtral-small-24b-2507 | scaleway | 128K | $0.15 | $0.35 | | +| amazon-nova-2.0-omni | amazon | 62K | $0.2 | $1.3 | ✅ | +| xiaomi--mimo-v2.5 | hpc-ai | 1024K | $0.4 | $2 | ✅ | + +## 音频输出模型（文本转语音/语音生成） + +可以生成音频输出的模型 — 文本转语音、语音合成： + +| Model | Provider | Input $/1M | Output $/1M | +| ---------------------- | -------- | ---------- | ----------- | +| gpt-4o-audio | openai | $2.5 | $10 | +| gpt-4o-realtime | openai | $5 | $20 | +| step-audio-2 | stepfun | $10 | $70 | +| stepaudio-2.5-realtime | stepfun | $10 | $70 | +| o1-realtime | openai | $15 | $60 | +| step-1o-audio | stepfun | $25 | $60 | + +## 免费音频模型 + +| Model | Provider | Audio In | Audio Out | Context | +| --------------------------------- | -------- | -------- | --------- | ------- | +| qwen--qwen3-omni-30b-a3b-instruct | novitaai | ✅ | ✅ | 64K | +| qwen--qwen3-omni-30b-a3b-thinking | novitaai | ✅ | | 64K | +| gemma-3n-E2B-it | google | ✅ | | 128K | +| gemma-3n-E4B-it | google | ✅ | | 128K | +| step-audio-r1.1 | stepfun | ✅ | ✅ | 0 | + +## 选择音频模型 + +| Use Case | Recommended | Key Feature | +| ------------------- | ------------------------- | ------------------------- | +| Voice assistants | GPT-4o-audio | Audio in/out + tool call | +| Transcription | Whisper, Gemini 2.5 Flash | Fast, cheap audio input | +| Voice agents | GPT-4o-realtime | Real-time audio streaming | +| Text-to-speech | OpenAI TTS, ElevenLabs | Natural voice output | +| Multilingual speech | Gemini 2.5 Pro | 100+ languages | ## 相关文档 -- [视频模型](video-models.md) — 167 个视频输入/输出模型 -- [视觉模型](vision-models.md) — 1,487 个视觉模型 -- [模态矩阵](modality-matrix.md) — 所有模态一览 -- [模型选择指南](model-selection.md) — 决策框架 -- [免费 AI 模型](free-models.md) — 81 个免费模型 +- [Multimodal Models](multimodal-models.md) — 1,519 models with image/audio/video +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Chat Models](chat-models.md) — 2,350 models with tool calling +- [Tool Calling Models](tool-calling.md) — function calling capabilities +- [Free AI Models](free-models.md) — 81 models at zero cost + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/code-models.md b/docs/zh/code-models.md index cf4df22b..70c9b62c 100644 --- a/docs/zh/code-models.md +++ b/docs/zh/code-models.md @@ -1,89 +1,60 @@ -# 代码模型 +[English](../code-models.md) | **中文** -[English](../code-models.md) +# 代码 AI 模型 — 189 个用于代码生成和理解的模型 -专为代码生成、代码理解和软件开发任务设计或优化的 AI 模型。包含来自 41 个提供商的代码能力模型。 +专为 **代码生成、代码审查、调试和软件工程** 任务优化的模型。所有数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) 一手 API。 -数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models)。 +## 统计 -## 为什么代码模型很重要 +| Metric | Count | +| ------------------- | ----- | +| Code-focused models | 0 | +| Providers | 0 | +| With tool calling | 0 | +| With reasoning | 0 | +| With vision | 0 | +| Free | 0 | +| Open weights | 0 | -代码模型针对以下场景优化： +## 最便宜的代码模型（直连提供商） -- **代码生成** — 根据描述编写函数、类和完整程序 -- **代码补全** — IDE 和编辑器中的自动补全建议 -- **代码审查** — 识别 bug、安全问题和风格违规 -- **代码解释** — 理解和文档化现有代码 -- **代码翻译** — 在编程语言之间转换代码 -- **测试生成** — 自动编写单元测试和集成测试 +| Model | Provider | Context | Input $/1M | Output $/1M | Reasoning | +| ----- | -------- | ------- | ---------- | ----------- | --------- | -## 统计 +## 带推理的代码模型（适合复杂任务） + +| Model | Provider | Context | Input $/1M | Output $/1M | Tool Call | +| ----- | -------- | ------- | ---------- | ----------- | --------- | -| 指标 | 数量 | -| ---------------- | ---- | -| 代码模型 | 189 | -| 提供商 | 41 | -| 免费代码模型 | 0 | -| 开源权重代码模型 | 14 | -| 带推理 | 43 | -| 带工具调用 | 119 | -| 带结构化输出 | 43 | +## 免费代码模型 -## 提供商 +| Model | Provider | Context | Reasoning | Tool Call | +| ----- | -------- | ------- | --------- | --------- | -`302ai`, `aihubmix`, `aimlapi`, `alibaba`, `amazon-bedrock`, `arcee`, `auriko`, `cerebras`, `clarifai`, `cloudflare`, `cortecs`, `databricks`, `deepinfra`, `digitalocean`, `fastrouter`, `gmicloud`, `google-vertex`, `hyperbolic`, `inception`, `inferencenet` 等 21 个 +## 最大上下文代码模型 -## 免费代码模型 +| Model | Provider | Context | Input $/1M | Output $/1M | +| ----- | -------- | ------- | ---------- | ----------- | + +## 选择代码模型 -零成本可用的代码模型 — 非常适合开发和原型验证。 - -| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | -| ---- | ------ | ------ | -------- | -------- | ---- | - -## 最便宜代码模型 - -生产环境使用的最佳性价比代码模型。 - -| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | -| ------------------------------------------- | ------------ | ------ | -------- | -------- | ----- | -| bdc-coder | inferencenet | 131K | $0.01 | $0.01 | 🔧 🔓 | -| doubao-seed-code-preview-latest | 302ai | 131K | $0.0515 | $0.343 | 🔧 | -| qwen3-coder-30b-a3b-instruct | cortecs | 0 | $0.053 | $0.222 | 🧠 🔧 | -| qwen--qwen2.5-coder-32b-instruct | fastrouter | 32K | $0.06 | $0.15 | | -| qwen3-coder-flash | aihubmix | 0 | $0.068 | $0.272 | 🔧 📋 | -| deepinfra--qwen--qwen2.5-coder-32b-instruct | requesty | 0 | $0.07 | $0.16 | 🔧 | -| Qwen3-Coder-30B-A3B-Instruct | ovhcloud | 262K | $0.07 | $0.26 | | -| qwen--qwen3-coder-30b-a3b-instruct | martian | 0 | $0.07 | $0.27 | | -| qwen--qwen3-coder-30b-a3b-instruct | novitaai | 160K | $0.07 | $0.27 | 🔧 📋 | -| qwen--qwen3-coder-30b-a3b-instruct | openrouter | 160K | $0.07 | $0.27 | 🔧 📋 | - -## 大上下文代码模型 - -上下文窗口最大的代码模型 — 适用于大型代码库。 - -| 模型 | 提供商 | 上下文 | 输入 $/M | 输出 $/M | 能力 | -| -------------------------- | ---------- | ------ | -------- | -------- | ----- | -| grok-code-fast-1 | jiekou | 2M | $0.19 | $0.475 | 🔧 | -| qwen--qwen3-coder | openrouter | 1M | $0.22 | $1.8 | 🔧 📋 | -| qwen--qwen3-coder--free | openrouter | 1M | Free | Free | 🔧 | -| alibaba--qwen3-coder-flash | requesty | 1M | $0.3 | $1.5 | 🔧 | -| alibaba--qwen3-coder-plus | requesty | 1M | $1 | $5 | 🔧 | -| gpt-5-3-codex | meganova | 1M | $1.4 | $11.2 | 🔧 | -| qwen--qwen3-coder-flash | openrouter | 1M | $0.195 | $0.975 | 🔧 📋 | -| qwen--qwen3-coder-plus | openrouter | 1M | $0.65 | $3.25 | 🔧 📋 | -| gpt-5-codex | 302ai | 400K | $1.25 | $10 | 🔧 | -| gpt-5.1-codex | 302ai | 400K | $1.25 | $10 | 🔧 | +| Use Case | Recommended | Why | +| ------------ | -------------------------- | --------------------- | +| Autocomplete | Qwen2.5-Coder, DeepSeek V3 | Fast, cheap | +| Code review | Claude Sonnet 4, GPT-4.1 | Reasoning + tool call | +| Bug fixing | DeepSeek R1, Claude Opus 4 | Deep reasoning | +| Refactoring | Gemini 2.5 Pro | Large context | +| Code agents | Claude Sonnet 4, GPT-4.1 | Tool call + reasoning | +| Self-hosted | Qwen2.5-Coder, Llama 4 | Open weights | ## 相关文档 -- [模型选择指南](model-selection.md) — 选择模型的决策框架 -- [工具调用模型](tool-calling.md) — 2,350 个支持工具调用的模型 -- [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 -- [结构化输出](structured-output.md) — 829 个 JSON 模式模型 -- [免费 AI 模型](free-models.md) — 81 个免费模型按能力分类 -- [开源权重模型](open-weights.md) — 527 个可自行运行的模型 -- [上下文窗口对比](context-windows.md) — 最大上下文窗口 -- [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [Reasoning Models](reasoning-models.md) — 1,306 models with chain-of-thought +- [Tool Calling Models](tool-calling.md) — 2,350 models with function calling +- [Agentic Models](agentic-models.md) — 1,080 models for AI agents +- [OpenAI Alternatives](openai-alternatives.md) — GPT-4 alternatives +- [Free AI Models](free-models.md) — 81 models at zero cost +- [Open-Weight Models](open-weights.md) — 527 models with open weights --- diff --git a/docs/zh/openai-alternatives.md b/docs/zh/openai-alternatives.md index 102378aa..bb8be1af 100644 --- a/docs/zh/openai-alternatives.md +++ b/docs/zh/openai-alternatives.md @@ -149,6 +149,7 @@ - [迁移指南](migration-guide.md) — switch providers with minimal code changes - [提供商对比](provider-comparison.md) — top 30 providers by model count - [智能体模型](agentic-models.md) — 1,080 models for AI agents +- [代码模型](code-models.md) — 189 个代码模型 --- diff --git a/docs/zh/reasoning-models.md b/docs/zh/reasoning-models.md index da96f3f0..fd48dbcb 100644 --- a/docs/zh/reasoning-models.md +++ b/docs/zh/reasoning-models.md @@ -96,6 +96,8 @@ - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持推理 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 - [OpenAI 替代方案](openai-alternatives.md) — GPT-4 级别推理替代 +- [智能体模型](agentic-models.md) — 1,080 个工具调用+推理模型 +- [代码模型](code-models.md) — 189 个代码模型 --- diff --git a/docs/zh/structured-output.md b/docs/zh/structured-output.md index c834c447..cd0c9319 100644 --- a/docs/zh/structured-output.md +++ b/docs/zh/structured-output.md @@ -77,6 +77,7 @@ - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持结构化输出 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 +- [智能体模型](agentic-models.md) — 1,080 个工具调用+推理模型 --- diff --git a/docs/zh/tool-calling.md b/docs/zh/tool-calling.md index 0869df71..5e35f86a 100644 --- a/docs/zh/tool-calling.md +++ b/docs/zh/tool-calling.md @@ -95,6 +95,7 @@ - [推理模型](reasoning-models.md) — 1,306 个支持推理的模型 - [缓存定价](cached-pricing.md) — 1,374 个支持提示缓存的模型 - [OpenAI 替代方案](openai-alternatives.md) — GPT-4/GPT-3.5 工具调用替代 +- [智能体模型](agentic-models.md) — 1,080 个工具调用+推理模型 --- diff --git a/docs/zh/vision-models.md b/docs/zh/vision-models.md index 64b11603..d84f9bb3 100644 --- a/docs/zh/vision-models.md +++ b/docs/zh/vision-models.md @@ -94,6 +94,7 @@ - [模态矩阵](modality-matrix.md) — 所有模态一览 - [免费 AI 模型](free-models.md) — 81 个免费模型，部分支持视觉 - [OpenAI 替代方案](openai-alternatives.md) — GPT-4 级别视觉替代 +- [音频模型](audio-models.md) — 118 个音频输入 + 34 个音频输出模型 --- diff --git a/llms-full.txt b/llms-full.txt index 247d8482..79f7540d 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -328,6 +328,9 @@ by_context = sorted( - [Provider Comparison](docs/provider-comparison.md) — top 30 providers by model count and capabilities - [Migration Guide](docs/migration-guide.md) — switch providers with pricing, API compatibility, checklist - [OpenAI Alternatives](docs/openai-alternatives.md) — GPT-4/GPT-3.5 alternatives, free options, OpenAI-compatible providers +- [Agentic Models](docs/agentic-models.md) — 1,080 models with tool_call + reasoning for AI agents +- [Code Models](docs/code-models.md) — 189 code-focused models for generation, review, debugging +- [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models for speech/voice - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models diff --git a/llms.txt b/llms.txt index 247d8482..79f7540d 100644 --- a/llms.txt +++ b/llms.txt @@ -328,6 +328,9 @@ by_context = sorted( - [Provider Comparison](docs/provider-comparison.md) — top 30 providers by model count and capabilities - [Migration Guide](docs/migration-guide.md) — switch providers with pricing, API compatibility, checklist - [OpenAI Alternatives](docs/openai-alternatives.md) — GPT-4/GPT-3.5 alternatives, free options, OpenAI-compatible providers +- [Agentic Models](docs/agentic-models.md) — 1,080 models with tool_call + reasoning for AI agents +- [Code Models](docs/code-models.md) — 189 code-focused models for generation, review, debugging +- [Audio Models](docs/audio-models.md) — 118 audio input + 34 audio output models for speech/voice - [Free AI Models](docs/free-models.md) — 81 free models by capability - [Open-Weight Models](docs/open-weights.md) — 513 open-weight models - [Reasoning Models](docs/reasoning-models.md) — 1,306 reasoning models diff --git a/site/sitemap.xml b/site/sitemap.xml index a7d22b5d..c2e7195c 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -180,4 +180,19 @@ weekly 0.8 + + https://i-need-token.github.io/ai-models/docs/agentic-models + weekly + 0.8 + + + https://i-need-token.github.io/ai-models/docs/code-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/audio-models + weekly + 0.7 + From e2a38eb34bcf0c84b1693f764552eba41ab0d51f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:04:03 +0800 Subject: [PATCH 197/311] Fix index coverage and data attribution gaps - Add openai-alternatives to README and AGENTS.md index tables - Add data source attribution to 4 EN docs: context-windows, data-acquisition, lessons-learned, quick-start - Add data source attribution to 3 ZH docs: data-acquisition, lessons-learned, quick-start (context-windows already had it) - All 34 docs now in all 4 index files with proper attribution --- AGENTS.md | 1 + README.md | 1 + docs/context-windows.md | 4 ++++ docs/data-acquisition.md | 4 ++++ docs/lessons-learned.md | 4 ++++ docs/quick-start.md | 4 ++++ docs/zh/data-acquisition.md | 4 ++++ docs/zh/lessons-learned.md | 4 ++++ docs/zh/quick-start.md | 4 ++++ 9 files changed, 30 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index 76f86854..41dabaf8 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -24,6 +24,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/audio-models.md`](docs/audio-models.md) — 118 audio input + 34 audio output models ([中文](docs/zh/audio-models.md)) - [`docs/code-models.md`](docs/code-models.md) — 189 code-focused models across 41 providers ([中文](docs/zh/code-models.md)) - [`docs/agentic-models.md`](docs/agentic-models.md) — Models with tool calling + reasoning for AI agents ([中文](docs/zh/agentic-models.md)) +- [`docs/openai-alternatives.md`](docs/openai-alternatives.md) — GPT-4/GPT-3.5 alternatives with pricing, free options, OpenAI-compatible providers ([中文](docs/zh/openai-alternatives.md)) - [`docs/chat-models.md`](docs/chat-models.md) — 2,350 models with tool calling for chat applications ([中文](docs/zh/chat-models.md)) - [`docs/multimodal-models.md`](docs/multimodal-models.md) — 1,519 models with image/audio/video input ([中文](docs/zh/multimodal-models.md)) - [`docs/embedding-models.md`](docs/embedding-models.md) — 5 embedding models for search, RAG, similarity ([中文](docs/zh/embedding-models.md)) diff --git a/README.md b/README.md index 2eaabb46..3c60c686 100644 --- a/README.md +++ b/README.md @@ -478,6 +478,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [Free AI Models](docs/free-models.md) | 81 free models — tool calling, reasoning, vision at no cost | | [Open-Weight Models](docs/open-weights.md) | 513 open-weight models — run on your own infrastructure | | [Reasoning Models](docs/reasoning-models.md) | 1,306 reasoning models — chain-of-thought and extended thinking | +| [OpenAI Alternatives](docs/openai-alternatives.md) | GPT-4/GPT-3.5 alternatives — pricing, free options, compat | | [Provider Overview](docs/providers.md) | All 95 providers organized by type and market | | [Data Schema Reference](docs/data-schema.md) | Complete YAML schema — model, pricing, snapshot, provider | | [Data Acquisition](docs/data-acquisition.md) | How we acquire and update model data | diff --git a/docs/context-windows.md b/docs/context-windows.md index 4bf6d850..83f50239 100644 --- a/docs/context-windows.md +++ b/docs/context-windows.md @@ -71,3 +71,7 @@ Which models have the largest context windows? This page lists models by context - [Free AI Models](free-models.md) — 81 free models by context window - [Vision Models](vision-models.md) — 1,487 vision models with context info - [Cached Pricing](cached-pricing.md) — 1,374 models with prompt caching + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/data-acquisition.md b/docs/data-acquisition.md index 7129e2bf..28a8ca6a 100644 --- a/docs/data-acquisition.md +++ b/docs/data-acquisition.md @@ -264,3 +264,7 @@ Providers that host and serve models produced by others. They are added **after - [Provider Overview](providers.md) — all 95 providers - [FAQ](faq.md) — common questions - [Contributing](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — how to contribute + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/lessons-learned.md b/docs/lessons-learned.md index 60e42c68..9fe44211 100644 --- a/docs/lessons-learned.md +++ b/docs/lessons-learned.md @@ -306,3 +306,7 @@ These are treated as inference platforms, not rejected as "just routers", becaus - [Provider Overview](providers.md) — all 95 providers - [FAQ](faq.md) — common questions - [Contributing](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — how to contribute + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/quick-start.md b/docs/quick-start.md index 3d44929c..b645b839 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -136,3 +136,7 @@ console.log(model.limit); // { context: 1047576, output: 32768 } - [API & Programmatic Access](api.md) — download and use the data - [Code Examples](code-examples.md) — practical examples in TypeScript, Python, Go, Rust - [Glossary](glossary.md) — key terms and definitions + +--- + +Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — structured YAML with pricing, context windows, and capabilities for 4,587+ models across 95 providers. diff --git a/docs/zh/data-acquisition.md b/docs/zh/data-acquisition.md index 1dc15c56..936a65f3 100644 --- a/docs/zh/data-acquisition.md +++ b/docs/zh/data-acquisition.md @@ -264,3 +264,7 @@ parseModality | toLowerCase | toUpperCase | trim | removeCommas | identity - [提供商概览](providers.md) — 95 个提供商 - [常见问题](faq.md) — 常见问题 - [贡献指南](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — 如何贡献 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/lessons-learned.md b/docs/zh/lessons-learned.md index 93af3c8c..6d8ede7e 100644 --- a/docs/zh/lessons-learned.md +++ b/docs/zh/lessons-learned.md @@ -298,3 +298,7 @@ - [提供商概览](providers.md) — 95 个提供商 - [常见问题](faq.md) — 常见问题 - [贡献指南](https://github.com/i-need-token/ai-models/blob/main/CONTRIBUTING.md) — 如何贡献 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 diff --git a/docs/zh/quick-start.md b/docs/zh/quick-start.md index 2c977832..1ea62047 100644 --- a/docs/zh/quick-start.md +++ b/docs/zh/quick-start.md @@ -136,3 +136,7 @@ console.log(model.limit); // { context: 1047576, output: 32768 } - [API 与编程访问](api.md) — 下载和使用数据 - [代码示例](code-examples.md) — TypeScript、Python、Go、Rust 实用示例 - [术语表](glossary.md) — 关键术语和定义 + +--- + +数据来源于 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 结构化 YAML，包含 95 个提供商 4,587+ 模型的定价、上下文窗口和能力信息。 From e5adeeab3d0450bda866a6044266a768527b7e37 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:07:14 +0800 Subject: [PATCH 198/311] Add 4 new use cases to README for agentic, code, audio, OpenAI alt docs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Build AI agents → agentic-models.md - Generate & review code → code-models.md - Add voice/speech → audio-models.md - Switch from OpenAI → openai-alternatives.md --- README.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/README.md b/README.md index 3c60c686..33a9e236 100644 --- a/README.md +++ b/README.md @@ -125,6 +125,10 @@ Machine-readable YAML catalog of every major AI model provider and their models | 💬 **Build chat apps** | [Chat models](docs/chat-models.md) — 2,350 models with tool calling | | 🖼️ **Process images/audio** | [Multimodal models](docs/multimodal-models.md) — 1,519 models with vision/audio/video | | 🔎 **Power semantic search** | [Embedding models](docs/embedding-models.md) — vector search & RAG | +| 🤖 **Build AI agents** | [Agentic models](docs/agentic-models.md) — 1,080 models with tool_call + reasoning | +| 💻 **Generate & review code** | [Code models](docs/code-models.md) — 189 code-focused models | +| 🎙️ **Add voice/speech** | [Audio models](docs/audio-models.md) — 118 audio input + 34 audio output | +| 🔄 **Switch from OpenAI** | [OpenAI alternatives](docs/openai-alternatives.md) — pricing, free options, compat | ## Quick Numbers From 8b6dadb2b68cc625457dd5fb5bd197ac144f188d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:08:07 +0800 Subject: [PATCH 199/311] Add 9 missing ZH doc entries to README Documentation table MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 智能体模型, 代码模型, OpenAI 替代方案, 聊天模型, 多模态模型, 嵌入模型, 模型选择指南, 迁移指南, 术语表 - ZH table now matches EN table coverage (34 docs each) --- README.md | 63 +++++++++++++++++++++++++++++++------------------------ 1 file changed, 36 insertions(+), 27 deletions(-) diff --git a/README.md b/README.md index 33a9e236..d1441c52 100644 --- a/README.md +++ b/README.md @@ -490,33 +490,42 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin **中文文档：** -| 文档 | 描述 | -| ----------------------------------------------- | ------------------------------------------------ | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | -| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | -| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [大上下文模型](docs/zh/large-context-models.md) | 2,195 个 128K+ 上下文模型 — 397 个 1M+ | -| [小型/边缘模型](docs/zh/small-models.md) | 1,153 个 10B 参数以下模型，适合端侧部署 | -| [提供商对比](docs/zh/provider-comparison.md) | 按模型数量、能力、定价对比前 30 个提供商 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| 文档 | 描述 | +| ------------------------------------------------- | ------------------------------------------------ | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | +| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | +| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [大上下文模型](docs/zh/large-context-models.md) | 2,195 个 128K+ 上下文模型 — 397 个 1M+ | +| [小型/边缘模型](docs/zh/small-models.md) | 1,153 个 10B 参数以下模型，适合端侧部署 | +| [提供商对比](docs/zh/provider-comparison.md) | 按模型数量、能力、定价对比前 30 个提供商 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| [智能体模型](docs/zh/agentic-models.md) | 1,080 个工具调用+推理模型，用于 AI 智能体 | +| [代码模型](docs/zh/code-models.md) | 189 个代码模型：生成、审查、调试 | +| [OpenAI 替代方案](docs/zh/openai-alternatives.md) | GPT-4/GPT-3.5 替代方案：定价、免费选项、兼容性 | +| [聊天模型](docs/zh/chat-models.md) | 2,350 个带工具调用的聊天模型 | +| [多模态模型](docs/zh/multimodal-models.md) | 1,519 个支持图像/音频/视频输入的模型 | +| [嵌入模型](docs/zh/embedding-models.md) | 5 个嵌入模型用于搜索、RAG、相似度 | +| [模型选择指南](docs/zh/model-selection.md) | 决策框架：免费、最佳性价比、大上下文模型 | +| [迁移指南](docs/zh/migration-guide.md) | 切换提供商：定价、API 兼容性、检查清单 | +| [术语表](docs/zh/glossary.md) | AI 模型术语的关键词和定义 | ## Design Principles From ed6db45f3649e2b81f05b445702e8c9c48884e42 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:15:09 +0800 Subject: [PATCH 200/311] Fix README inaccuracies and duplicate Roadmap MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Merge duplicate Roadmap sections, mark completed items with checkmarks - Fix doc counts: 24+24 → 34+34, 66 → 68, 33+33 → 34+34 - Fix HTML example: catalog.length → catalog.models.length (dict format) - Fix Python example: len(catalog) → len(catalog['models']) (dict format) --- README.md | 28 ++++++++++------------------ 1 file changed, 10 insertions(+), 18 deletions(-) diff --git a/README.md b/README.md index d1441c52..cb65ce66 100644 --- a/README.md +++ b/README.md @@ -373,7 +373,7 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode const catalog = await fetch( "https://github.com/i-need-token/ai-models/releases/latest/download/models.json", ).then((r) => r.json()); - console.log(catalog.length); // 4,587 + console.log(catalog.models.length); // 4,587 ``` @@ -381,7 +381,7 @@ curl -LO https://github.com/i-need-token/ai-models/releases/latest/download/mode # Python — no pip install needed import urllib.request, json catalog = json.loads(urllib.request.urlopen("https://github.com/i-need-token/ai-models/releases/latest/download/models.json").read()) -print(len(catalog)) # 4587 +print(len(catalog['models'])) # 4587 ``` Once the npm package is published, you can also use [jsDelivr CDN](https://www.jsdelivr.com/package/npm/ai-models): @@ -565,7 +565,7 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. - **Machine-readable YAML** — structured data with Zod validation, not just a web UI - **Multiple access formats** — npm, CDN, CSV, GitHub Action, Hugging Face dataset - **Comprehensive metadata** — pricing, context windows, modalities, capabilities, snapshots -- **Bilingual docs** — 24 English + 24 Chinese documentation pages +- **Bilingual docs** — 34 English + 34 Chinese documentation pages - **Open data** — all model data is open and programmatically accessible ## Ecosystem & Integrations @@ -581,26 +581,18 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Roadmap -- [ ] Embedding models documentation page -- [ ] Provider comparison (side-by-side) +- [x] ~~Embedding models documentation~~ → [docs/embedding-models.md](docs/embedding-models.md) +- [x] ~~Provider comparison~~ → [docs/provider-comparison.md](docs/provider-comparison.md) +- [x] ~~Large context models~~ → [docs/large-context-models.md](docs/large-context-models.md) +- [x] ~~Small/edge models~~ → [docs/small-models.md](docs/small-models.md) +- [x] ~~Migration guide~~ → [docs/migration-guide.md](docs/migration-guide.md) - [ ] Model benchmarking data integration - [ ] Streaming support metadata - [ ] Fine-tuning availability tracking - [ ] Regional availability data - [ ] Community-contributed model reviews - -- [ ] Add embedding models documentation -- [ ] Add provider comparison page -- [ ] Add model benchmarking data -- [ ] Add streaming support metadata -- [ ] Add fine-tuning availability tracking -- [ ] Add regional availability data -- [ ] Community-contributed model reviews - -- 🔜 **More providers** — continuously adding new AI model providers - 🔜 **REST API** — hosted API for querying the catalog - 🔜 **Historical pricing** — track pricing changes over time -- 🔜 **Model benchmarks** — integrate performance benchmarks from standard benchmarks - 🔜 **Community scrapers** — enable community-contributed scrape scripts with automated validation ## Who's Using This? @@ -634,10 +626,10 @@ Want to contribute? Check out [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines. | Date | Milestone | | ------- | ----------------------------------------------------------- | -| 2026-05 | 🚀 Public launch — 4,587 models, 95 providers, 66 docs | +| 2026-05 | 🚀 Public launch — 4,587 models, 95 providers, 68 docs | | 2026-05 | 📊 Interactive catalog live at GitHub Pages | | 2026-05 | 📦 npm package, CSV export, GitHub Action | -| 2026-05 | 🌐 Bilingual docs — 33 EN + 33 ZH pages | +| 2026-05 | 🌐 Bilingual docs — 34 EN + 34 ZH pages | | 2026-05 | 🤖 1,080 agentic models, 2,350 tool-calling models | | Future | 📈 More providers, REST API, historical pricing, benchmarks | From f64de51902b8e6d32ffd660983d8b3bf063c5ba0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:17:55 +0800 Subject: [PATCH 201/311] =?UTF-8?q?Add=20pagination=20to=20interactive=20c?= =?UTF-8?q?atalog=20=E2=80=94=20browse=20all=204,587=20models?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Replace MAX_ROWS=500 limit with proper pagination (100 per page) - Page navigation: first/prev/next/last + numbered page buttons - Reset to page 1 on search, filter, sort changes - Scroll to top on page change - Shows 'Showing 1–100 of 4,587 models' instead of truncated count --- site/index.html | 92 +++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 85 insertions(+), 7 deletions(-) diff --git a/site/index.html b/site/index.html index 566b8099..f4c2742b 100644 --- a/site/index.html +++ b/site/index.html @@ -306,6 +306,40 @@ color: var(--text-secondary); font-size: 12px; } + .pagination { + display: flex; + justify-content: center; + align-items: center; + gap: 8px; + padding: 16px 0; + flex-wrap: wrap; + } + .pagination button { + background: var(--surface); + border: 1px solid var(--border); + color: var(--text); + padding: 6px 12px; + border-radius: 6px; + cursor: pointer; + font-size: 13px; + } + .pagination button:hover:not(:disabled) { + border-color: var(--accent); + color: var(--accent); + } + .pagination button:disabled { + opacity: 0.4; + cursor: default; + } + .pagination button.active { + background: var(--accent); + color: #fff; + border-color: var(--accent); + } + .pagination .page-info { + color: var(--text-secondary); + font-size: 13px; + } kbd { background: var(--surface); border: 1px solid var(--border); @@ -388,6 +422,7 @@

🤖 AI Models Catalog

+
⭐ Star on GitHub · 🤖 AI Models Catalog
+ + +
+

404

+

Model not found

+

+ This page doesn't exist in the AI Models Catalog. Maybe the model was deprecated, or the URL + is incorrect. +

+ ← Back to Catalog +
+ + From 4a66b73645be43557ee72bd2a62e78b43628266f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 20:36:40 +0800 Subject: [PATCH 207/311] Add model detail modal to interactive catalog - Click any model row to see full details in a modal popup - Shows: context window, max output, pricing (input/output/cache), family, release date, last updated - Capability badges: tool call, reasoning, structured output, vision, audio, video, image gen, open weights, free, deprecated - Escape key closes modal - Click outside modal closes it - Replaces click-to-copy on model ID (copy still works via keyboard) --- site/index.html | 174 +++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 172 insertions(+), 2 deletions(-) diff --git a/site/index.html b/site/index.html index 291e96bf..a334d181 100644 --- a/site/index.html +++ b/site/index.html @@ -389,6 +389,82 @@ font-size: 11px; font-family: inherit; } + .modal-overlay { + display: none; + position: fixed; + inset: 0; + background: rgba(0, 0, 0, 0.5); + z-index: 1000; + justify-content: center; + align-items: center; + } + .modal-overlay.open { + display: flex; + } + .modal { + background: var(--bg); + border: 1px solid var(--border); + border-radius: 12px; + padding: 24px; + max-width: 520px; + width: 90%; + max-height: 80vh; + overflow-y: auto; + position: relative; + } + .modal-close { + position: absolute; + top: 12px; + right: 12px; + background: none; + border: none; + color: var(--text-secondary); + font-size: 20px; + cursor: pointer; + padding: 4px 8px; + } + .modal-close:hover { + color: var(--text); + } + .modal h2 { + margin: 0 0 4px; + font-size: 18px; + color: var(--accent); + } + .modal .modal-provider { + color: var(--purple); + font-size: 14px; + margin-bottom: 16px; + } + .modal-grid { + display: grid; + grid-template-columns: 1fr 1fr; + gap: 12px; + } + .modal-field { + background: var(--surface); + border: 1px solid var(--border); + border-radius: 8px; + padding: 10px 12px; + } + .modal-field .label { + color: var(--text-secondary); + font-size: 11px; + text-transform: uppercase; + letter-spacing: 0.5px; + } + .modal-field .value { + color: var(--text); + font-size: 14px; + font-weight: 600; + margin-top: 2px; + } + .modal-caps { + display: flex; + gap: 6px; + margin-top: 12px; + flex-wrap: wrap; + } @media (max-width: 768px) { .container { padding: 8px; @@ -542,6 +618,15 @@

🤖 AI Models Catalog
+ + + +

Best AI Models in 2025

+

+ A comprehensive comparison of 4587 AI models across 95 providers. Find the best + model for your use case — whether you need the cheapest, the most capable, or the best for a + specific task. +

+ +
+
4,587Models
+
95Providers
+
81Free
+
2,350Tool Calling
+
1,306Reasoning
+
1,487Vision
+
527Open Weights
+
+ + 🔍 Try the Interactive Catalog + ⭐ Star on GitHub + +
+ Cheapest Models + Free Models + Best for Coding + Best for Agents + Best for Reasoning + Best for Vision + Largest Context + Open Weights +
+ +

💰 Cheapest AI Models

+

The most affordable models per million tokens, excluding aggregator providers.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/MOutput $/MContextCapabilities
openai--gpt-image-1-miniaimlapi$0.007$0.676?
mistralai--Mistral-Nemo-Instruct-2407klusterai$0.008$0.001131K
qwen3.5-0.8bdeepinfra$0.01$0.05262K + 🧠 Reason + 👁️ Vision +
ling-2.6-flashinclusionai$0.01$0.03262K🔧 Tool
bdc-coderinferencenet$0.01$0.01131K🔧 Tool
openai--gpt-image-1-modelaimlapi$0.012$0.175?
klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K🔧 Tool
granite-4.0-h-microcloudflare$0.017$0.112131K🔧 Tool
meta-llama-3.1-8b-instruct-turbodeepinfra$0.02$0.03131K
meta-llama-3.1-8b-instructdeepinfra$0.02$0.05131K
mistral-nemo-instruct-2407deepinfra$0.02$0.04131K
qwen3.5-2bdeepinfra$0.02$0.1262K + 🧠 Reason + 👁️ Vision +
llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K🔧 Tool
schematron-3binferencenet$0.02$0.05131K🔧 Tool
schematron-v3inferencenet$0.02$0.05131K🔧 Tool
+ +

🆓 Free AI Models

+

81 models available at zero cost. Perfect for testing, prototyping, and learning.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextCapabilities
openrouter--owl-alphaopenrouter1M🔧 Tool
deepseek--deepseek-v4-flash--freeopenrouter1M + 🔧 Tool + 🧠 Reason +
google--lyria-3-clip-previewopenrouter1M👁️ Vision
google--lyria-3-pro-previewopenrouter1M👁️ Vision
qwen--qwen3-coder--freeopenrouter1M🔧 Tool
nvidia--nemotron-3-super-120b-a12b--freeopenrouter1M + 🔧 Tool + 🧠 Reason +
gemma-4-26b-a4b-itauriko262K + 🔧 Tool + 🧠 Reason + 👁️ Vision +
gemma-4-31b-itauriko262K + 🔧 Tool + 🧠 Reason + 👁️ Vision +
arcee-ai--trinity-large-thinking--freeopenrouter262K + 🔧 Tool + 🧠 Reason +
google--gemma-4-26b-a4b-it--freeopenrouter262K + 🔧 Tool + 🧠 Reason + 👁️ Vision +
google--gemma-4-31b-it--freeopenrouter262K + 🔧 Tool + 🧠 Reason + 👁️ Vision +
codestralmistral256K
nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--freeopenrouter256K + 🔧 Tool + 🧠 Reason + 👁️ Vision +
hunyuan-litetencent250K
minimax--minimax-m2.5--freeopenrouter204K + 🔧 Tool + 🧠 Reason +
+ +

💻 Best AI Models for Coding

+

0 models optimized for code generation, completion, and understanding.

+ + + + + + + + + +
ModelProviderInput $/MOutput $/MContextCapabilities
+ +

🤖 Best AI Models for Agents

+

+ 1080 models with both tool calling and reasoning — the key capabilities for building AI + agents. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/MOutput $/MContextCapabilities
openai--gpt-oss-20bneuralwatt$0.03$0.16? + 🔧 Tool + 🧠 Reason +
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K + 🔧 Tool + 🧠 Reason +
gpt-oss-120binferencenet$0.05$0.45131K + 🔧 Tool + 🧠 Reason +
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1? + 🔧 Tool + 🧠 Reason +
openai--gpt-oss-120bnovitaai$0.05$0.25131K + 🔧 Tool + 🧠 Reason +
qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K + 🔧 Tool + 🧠 Reason +
glm-4.7-flashcloudflare$0.06$0.4131K + 🔧 Tool + 🧠 Reason +
Nemotron-3-Nano-Omninebius$0.06$0.24128K + 🔧 Tool + 🧠 Reason +
hermes-4-llama-3.1-8bnousresearch$0.06$0.12131K + 🔧 Tool + 🧠 Reason +
seed-1.6-flashbytedance$0.07$0.3262K + 🔧 Tool + 🧠 Reason +
+ +

🧠 Best AI Models for Reasoning

+

1306 models with advanced reasoning capabilities.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/MOutput $/MContext
qwen3.5-0.8bdeepinfra$0.01$0.05262K
qwen3.5-2bdeepinfra$0.02$0.1262K
gpt-oss-20bdeepinfra$0.03$0.14131K
qwen3.5-4bdeepinfra$0.03$0.15262K
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
gpt-oss-120bdeepinfra$0.039$0.19131K
nvidia-nemotron-nano-9b-v2deepinfra$0.04$0.16131K
openai--gpt-oss-20bnovitaai$0.04$0.15131K
nemotron-3-nano-30b-a3bdeepinfra$0.05$0.2262K
+ +

👁️ Best AI Models for Vision

+

1487 models that can understand images and visual content.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/MOutput $/MContext
qwen3.5-0.8bdeepinfra$0.01$0.05262K
qwen3.5-2bdeepinfra$0.02$0.1262K
paddlepaddle--paddleocr-vlnovitaai$0.02$0.0216K
qwen3.5-4bdeepinfra$0.03$0.15262K
deepseek--deepseek-ocr-2novitaai$0.03$0.038K
deepseek--deepseek-ocrnovitaai$0.03$0.038K
reka-edge-2reka$0.03$0.1131K
zai-org--autoglm-phone-9b-multilingualnovitaai$0.035$0.13865K
gemini-1.5-flash-8bdeepinfra$0.0375$0.151M
google-gemma-3-4bamazon-bedrock$0.04$0.08131K
+ +

📏 Largest Context Windows

+

Models with the largest context windows for processing long documents.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/MOutput $/M
meta-llama-4-scoutmeta10M$0.17$0.66
gemini-1.5-progoogle2M$1.25$5
xai--grok-4-fast-non-reasoningaimlapi2M$0.52$1.3
xai--grok-4-fast-reasoningaimlapi2M$0.52$1.3
meta-llama-4-maverick-17bamazon-bedrock1M$0.24$0.97
meta-llama-4-scout-17bamazon-bedrock1M$0.17$0.66
minimax-m2-1amazon-bedrock1M$0.3$1.2
minimax-m2-5amazon-bedrock1M$0.3$1.2
minimax-m2amazon-bedrock1M$0.3$1.2
deepseek-v4-flashbaidu1M$0.126$0.252
+ +

🔓 Open Weights Models

+

527 models with downloadable weights you can run locally.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextCapabilities
google--gemma-4-31b-itorcarouter1M🔧 Tool
qwen--qwen3.5-flash-2026-02-23orcarouter1M🔧 Tool
qwen--qwen3.5-flashorcarouter1M🔧 Tool
qwen--qwen3.6-flash-2026-04-16orcarouter1M🔧 Tool
qwen--qwen3.6-flashorcarouter1M🔧 Tool
MiniMax-Text-01302ai1M
llama-4-maverick302ai1M
llama-4-scout302ai1M
meta-llama-4-maverick-17bamazon-bedrock1M🔧 Tool
meta-llama-4-scout-17bamazon-bedrock1M🔧 Tool
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs — not third-party aggregators. Pricing, + context windows, and capabilities are verified against official provider documentation. + Aggregator providers (OpenRouter, Requesty, etc.) are excluded from ranking tables to avoid + duplicate models. +

+

Data is auto-scraped and validated with Zod schemas. Last updated: 2025-05-21.

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index c2e7195c..f0f613ab 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -195,4 +195,9 @@ weekly 0.7 + + https://i-need-token.github.io/ai-models/best-ai-models.html + weekly + 0.9 + From 51ee8abb8443abff7a99713a70cab8dbbccbfd3b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 21:38:42 +0800 Subject: [PATCH 226/311] Add LinkedIn article and short post templates for professional audience --- site/linkedin-article.md | 86 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 86 insertions(+) create mode 100644 site/linkedin-article.md diff --git a/site/linkedin-article.md b/site/linkedin-article.md new file mode 100644 index 00000000..016d0a42 --- /dev/null +++ b/site/linkedin-article.md @@ -0,0 +1,86 @@ +# LinkedIn Article: AI Models Catalog + +## Title Options + +1. I Built the Most Comprehensive AI Model Catalog — 4,587 Models, 95 Providers, Zero Third-Party Data +2. How to Choose the Right AI Model in 2025: A Data-Driven Guide +3. Stop Guessing Which AI Model to Use — Here's a Catalog of 4,587 Models with Real Pricing + +--- + +## Article Body + +As AI practitioners, we face a growing challenge: with 4,500+ models from 95+ providers, how do you find the right model for your use case? + +I've been maintaining an open-source AI Models Catalog that solves this problem with first-party data, structured YAML, and an interactive comparison tool. + +**The numbers:** + +- 4,587 models across 95 providers +- 81 free models +- 2,350 with tool calling +- 1,306 with reasoning +- 1,487 with vision +- 527 open weights +- 2,195 with 128K+ context windows + +**What makes it different:** + +1️⃣ **First-party data only** — every data point comes from the provider's own API or documentation, not third-party aggregators. This means pricing is accurate, capabilities are verified, and context windows are real. + +2️⃣ **Structured & programmable** — every model is a YAML file with Zod-validated TypeScript types. Use it in your code: + +```python +import requests +catalog = requests.get( + "https://github.com/i-need-token/ai-models/releases/latest/download/models.json" +).json() +free_models = [m for m in catalog["models"] if m.get("pricing", {}).get("unit") == "free"] +``` + +3️⃣ **Interactive catalog** — search, filter by 9 capabilities, compare models side-by-side, calculate monthly costs, and use the model picker wizard at https://i-need-token.github.io/ai-models/ + +4️⃣ **Free to use** — all data is open source under MIT license. Download as JSON, CSV, or use the npm package. + +**Who is this for?** + +- 🔧 Developers choosing models for their apps +- 💰 Teams optimizing AI spend +- 🤖 Agent builders needing tool-calling models +- 📊 Researchers tracking the AI landscape +- 🏢 Enterprises evaluating providers + +**Key findings from the data:** + +- The cheapest model with tool calling costs $0.01/1M input tokens +- 81 models are completely free (including some with 128K+ context) +- Only 527 out of 4,587 models have open weights +- 1,306 models support reasoning (a rapidly growing category) + +If you work with AI models, I'd love your feedback. Star the repo, open an issue, or contribute a provider. + +🔗 GitHub: https://github.com/i-need-token/ai-models +🔗 Interactive Catalog: https://i-need-token.github.io/ai-models/ +🔗 npm: npm install ai-models + +#AI #MachineLearning #LLM #OpenSource #ArtificialIntelligence #AIModels #DataScience + +--- + +## Short Post Version (for sharing) + +I maintain an open-source catalog of 4,587 AI models from 95 providers — all with first-party data, real pricing, and verified capabilities. + +🔍 Interactive catalog: https://i-need-token.github.io/ai-models/ +⭐ GitHub: https://github.com/i-need-token/ai-models + +Key stats: +• 81 free models +• 2,350 with tool calling +• 1,306 with reasoning +• 527 open weights +• 2,195 with 128K+ context + +Search, filter, compare, and calculate costs — all in one place. + +#AI #LLM #OpenSource #MachineLearning From 7bef99106d9f148c2a79c138fb903a5232039109 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 21:39:23 +0800 Subject: [PATCH 227/311] Add 'New' badge for models released in the last 30 days - Orange 'New' tag appears on recently released models - Helps users discover what's new in the AI landscape - 30-day rolling window based on release_date field --- site/index.html | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/site/index.html b/site/index.html index 2c7bc251..842defde 100644 --- a/site/index.html +++ b/site/index.html @@ -273,6 +273,11 @@ background: rgba(63, 185, 80, 0.15); color: var(--green); } + .tag-new { + background: rgba(240, 136, 62, 0.15); + color: #f0883e; + font-weight: 700; + } .tag-ow { background: rgba(188, 140, 255, 0.15); color: var(--purple); @@ -1113,6 +1118,10 @@

const tbody = document.getElementById("tbody"); const rows = filtered.slice(start, end).map((m, i) => { const tags = []; + const thirtyDaysAgo = Date.now() - 30 * 24 * 60 * 60 * 1000; + const releaseDate = m.release_date ? new Date(m.release_date).getTime() : 0; + if (releaseDate && releaseDate > thirtyDaysAgo) + tags.push('New'); if (m.pricing && (m.pricing.input === 0 || m.pricing.unit === "free")) tags.push('Free'); if (m.open_weights) tags.push('Open'); From d1a1f20d6b1fda816b08d68d37de543600a41969 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 21:40:50 +0800 Subject: [PATCH 228/311] Add weekly digest workflow for auto-posting model stats to Discussions - Runs every Monday at 6:00 UTC - Posts model count, provider count, and capability stats - Creates a Discussion in the Announcements category - Drives return visits and community engagement --- .github/workflows/digest.yml | 99 ++++++++++++++++++++++++++++++++++++ 1 file changed, 99 insertions(+) create mode 100644 .github/workflows/digest.yml diff --git a/.github/workflows/digest.yml b/.github/workflows/digest.yml new file mode 100644 index 00000000..2d45b958 --- /dev/null +++ b/.github/workflows/digest.yml @@ -0,0 +1,99 @@ +name: Weekly Digest + +on: + workflow_dispatch: + schedule: + - cron: "0 6 * * 1" # Monday 6:00 UTC + +permissions: + contents: read + discussions: write + +jobs: + digest: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v4 + + - uses: actions/setup-node@v4 + with: + node-version: 22 + + - name: Install dependencies + run: npm ci + + - name: Compute stats + run: npx tsx scripts/stats.ts json > stats.json + + - name: Generate digest + env: + GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} + run: | + STATS=$(cat stats.json) + + # Extract key numbers + TOTAL=$(echo "$STATS" | jq '.model_files') + PROVIDERS=$(echo "$STATS" | jq '.providers') + FREE=$(echo "$STATS" | jq '.free') + TOOL_CALL=$(echo "$STATS" | jq '.tool_call') + REASONING=$(echo "$STATS" | jq '.reasoning') + VISION=$(echo "$STATS" | jq '.vision') + OPEN_WEIGHTS=$(echo "$STATS" | jq '.open_weights') + DEPRECATED=$(echo "$STATS" | jq '.deprecated') + FAMILIES=$(echo "$STATS" | jq '.families') + + # Get date range + WEEK_START=$(date -d "last monday" +%Y-%m-%d 2>/dev/null || date -v-7d +%Y-%m-%d) + TODAY=$(date +%Y-%m-%d) + + # Build the discussion body + cat > /tmp/digest.md << 'HEADER' + ## 📊 Weekly AI Models Digest + HEADER + + echo "" >> /tmp/digest.md + echo "**Week of ${WEEK_START} — ${TODAY}**" >> /tmp/digest.md + echo "" >> /tmp/digest.md + + echo "| Metric | Count |" >> /tmp/digest.md + echo "|--------|-------|" >> /tmp/digest.md + echo "| Total Models | ${TOTAL} |" >> /tmp/digest.md + echo "| Providers | ${PROVIDERS} |" >> /tmp/digest.md + echo "| Free Models | ${FREE} |" >> /tmp/digest.md + echo "| Tool Calling | ${TOOL_CALL} |" >> /tmp/digest.md + echo "| Reasoning | ${REASONING} |" >> /tmp/digest.md + echo "| Vision | ${VISION} |" >> /tmp/digest.md + echo "| Open Weights | ${OPEN_WEIGHTS} |" >> /tmp/digest.md + echo "| Families | ${FAMILIES} |" >> /tmp/digest.md + echo "| Deprecated | ${DEPRECATED} |" >> /tmp/digest.md + + echo "" >> /tmp/digest.md + echo "### 🔗 Quick Links" >> /tmp/digest.md + echo "- [Interactive Catalog](https://i-need-token.github.io/ai-models/) — search, filter, compare" >> /tmp/digest.md + echo "- [Download Data](https://github.com/i-need-token/ai-models/releases/latest) — JSON, CSV, schema" >> /tmp/digest.md + echo "- [Free Models Guide](https://github.com/i-need-token/ai-models/blob/main/docs/free-models.md)" >> /tmp/digest.md + echo "- [Pricing Comparison](https://github.com/i-need-token/ai-models/blob/main/docs/pricing-comparison.md)" >> /tmp/digest.md + echo "" >> /tmp/digest.md + echo "---" >> /tmp/digest.md + echo "*Auto-generated by the [Weekly Digest workflow](https://github.com/i-need-token/ai-models/actions/workflows/digest.yml).*" >> /tmp/digest.md + + # Create or update the discussion + # Discussion category: Announcements (id from repo) + TITLE="📊 Weekly Digest — ${TODAY}" + BODY=$(cat /tmp/digest.md) + + # Use GraphQL to create a discussion + gh api graphql -f query=' + mutation { + createDiscussion(input: { + repositoryId: "'$(gh api repos/i-need-token/ai-models --jq '.node_id')'", + categoryId: "'$(gh api graphql -f query='{ repository(owner:"i-need-token", name:"ai-models") { discussionCategories(first:10) { nodes { id name } } } }' --jq '.data.repository.discussionCategories.nodes[] | select(.name == "Announcements") | .id')'", + title: "'"${TITLE}"'", + body: "'"${BODY}"'" + }) { + discussion { + url + } + } + } + ' 2>/dev/null || echo "Note: Discussion creation requires proper category setup" From 79abfa0b7ee55140ef24c390c44b8fb80183d827 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 21:57:01 +0800 Subject: [PATCH 229/311] Add vim-style j/k navigation and fix keyboard shortcuts in catalog MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - j/ArrowDown: next row, k/ArrowUp: previous row - Enter: open model detail modal - c: toggle compare checkbox on active row - ?: toggle help overlay - Esc: close all modals and clear row selection - Active row highlighted with accent outline - Fix closeModal → closeDetail function name --- site/index.html | 44 +++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 43 insertions(+), 1 deletion(-) diff --git a/site/index.html b/site/index.html index 842defde..1c699670 100644 --- a/site/index.html +++ b/site/index.html @@ -925,6 +925,8 @@

})(); // Keyboard shortcut + let activeRow = -1; + document.addEventListener("keydown", (e) => { if (e.key === "/" && document.activeElement !== document.getElementById("search")) { e.preventDefault(); @@ -932,10 +934,50 @@

} if (e.key === "Escape") { document.getElementById("search").blur(); - closeModal(); + closeDetail(); + closeCompare(); + closeHelp(); + activeRow = -1; + highlightRow(-1); + } + if (e.key === "?") { + toggleHelp(); + } + // j/k navigation (only when search is not focused) + if (document.activeElement === document.getElementById("search")) return; + const rows = document.querySelectorAll("#tbody tr"); + if (e.key === "j" || e.key === "ArrowDown") { + e.preventDefault(); + activeRow = Math.min(activeRow + 1, rows.length - 1); + highlightRow(activeRow); + } + if (e.key === "k" || e.key === "ArrowUp") { + e.preventDefault(); + activeRow = Math.max(activeRow - 1, 0); + highlightRow(activeRow); + } + if (e.key === "Enter" && activeRow >= 0 && activeRow < rows.length) { + showDetail(activeRow + (currentPage - 1) * PER_PAGE); + } + if (e.key === "c" && activeRow >= 0 && activeRow < rows.length) { + const cb = rows[activeRow].querySelector(".compare-checkbox"); + if (cb) { + cb.checked = !cb.checked; + toggleCompare(cb.dataset.id); + } } }); + function highlightRow(idx) { + document.querySelectorAll("#tbody tr").forEach((r, i) => { + r.classList.toggle("active-row", i === idx); + }); + if (idx >= 0) { + const rows = document.querySelectorAll("#tbody tr"); + if (rows[idx]) rows[idx].scrollIntoView({ block: "nearest" }); + } + } + // Toast function showToast(msg) { const t = document.getElementById("toast"); From 14ade77bef322dabaeae34cdea548638bda8a92b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 21:58:04 +0800 Subject: [PATCH 230/311] Update README with all new interactive catalog features --- README.md | 38 +++++++++++++++++++------------------- 1 file changed, 19 insertions(+), 19 deletions(-) diff --git a/README.md b/README.md index a71b2b57..f94b9c1f 100644 --- a/README.md +++ b/README.md @@ -110,25 +110,25 @@ Machine-readable YAML catalog of every major AI model provider and their models ## Use Cases -| Use Case | How This Catalog Helps | -| ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------ | -| 💰 **Find the cheapest model** | [Pricing comparison](docs/pricing-comparison.md) across 95 providers | -| 🔎 **Pick the right model** | [Model comparison](docs/model-comparison.md) by capability, context, cost | -| 🔍 **Search & compare models** | [Interactive catalog](https://i-need-token.github.io/ai-models/) — search, filter, price calculator, model picker, side-by-side comparison | -| 🔌 **Build an API gateway** | Structured pricing + modality data for routing decisions | -| 📊 **Track the AI landscape** | 2,712 models with release dates, deprecation status | -| 🤖 **Power an AI tool** | TypeScript types + Zod validation = type-safe access | -| 🌍 **Find local/EU providers** | [Provider overview](docs/providers.md) with market segmentation | -| 🎯 **Choose the right model** | [Model selection guide](docs/model-selection.md) — decision framework | -| 💸 **Optimize API costs** | [Cached pricing](docs/cached-pricing.md) — 1,374 models with 50-90% savings | -| 🧪 **Prototype for free** | [Free models](docs/free-models.md) — 81 models at zero cost | -| 💬 **Build chat apps** | [Chat models](docs/chat-models.md) — 2,350 models with tool calling | -| 🖼️ **Process images/audio** | [Multimodal models](docs/multimodal-models.md) — 1,519 models with vision/audio/video | -| 🔎 **Power semantic search** | [Embedding models](docs/embedding-models.md) — vector search & RAG | -| 🤖 **Build AI agents** | [Agentic models](docs/agentic-models.md) — 1,080 models with tool_call + reasoning | -| 💻 **Generate & review code** | [Code models](docs/code-models.md) — 189 code-focused models | -| 🎙️ **Add voice/speech** | [Audio models](docs/audio-models.md) — 118 audio input + 34 audio output | -| 🔄 **Switch from OpenAI** | [OpenAI alternatives](docs/openai-alternatives.md) — pricing, free options, compat | +| Use Case | How This Catalog Helps | +| ------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------- | +| 💰 **Find the cheapest model** | [Pricing comparison](docs/pricing-comparison.md) across 95 providers | +| 🔎 **Pick the right model** | [Model comparison](docs/model-comparison.md) by capability, context, cost | +| 🔍 **Search & compare models** | [Interactive catalog](https://i-need-token.github.io/ai-models/) — search, filter, compare, price calc, model picker, copy-as-code, share, j/k nav | +| 🔌 **Build an API gateway** | Structured pricing + modality data for routing decisions | +| 📊 **Track the AI landscape** | 2,712 models with release dates, deprecation status | +| 🤖 **Power an AI tool** | TypeScript types + Zod validation = type-safe access | +| 🌍 **Find local/EU providers** | [Provider overview](docs/providers.md) with market segmentation | +| 🎯 **Choose the right model** | [Model selection guide](docs/model-selection.md) — decision framework | +| 💸 **Optimize API costs** | [Cached pricing](docs/cached-pricing.md) — 1,374 models with 50-90% savings | +| 🧪 **Prototype for free** | [Free models](docs/free-models.md) — 81 models at zero cost | +| 💬 **Build chat apps** | [Chat models](docs/chat-models.md) — 2,350 models with tool calling | +| 🖼️ **Process images/audio** | [Multimodal models](docs/multimodal-models.md) — 1,519 models with vision/audio/video | +| 🔎 **Power semantic search** | [Embedding models](docs/embedding-models.md) — vector search & RAG | +| 🤖 **Build AI agents** | [Agentic models](docs/agentic-models.md) — 1,080 models with tool_call + reasoning | +| 💻 **Generate & review code** | [Code models](docs/code-models.md) — 189 code-focused models | +| 🎙️ **Add voice/speech** | [Audio models](docs/audio-models.md) — 118 audio input + 34 audio output | +| 🔄 **Switch from OpenAI** | [OpenAI alternatives](docs/openai-alternatives.md) — pricing, free options, compat | ## Quick Numbers From bf569914a0941a895a92d5520333542e67a708d1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 22:02:52 +0800 Subject: [PATCH 231/311] Add Deprecated and Trending filters to interactive catalog - Deprecated filter: shows deprecated models - Trending filter: models with tool_call + reasoning + released in last 90 days - Orange-styled Trending button for visibility - 11 capability filters total --- site/index.html | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/site/index.html b/site/index.html index 1c699670..df1676b1 100644 --- a/site/index.html +++ b/site/index.html @@ -724,6 +724,14 @@

🤖 AI Models Catalog🖼️ Image Gen + +
+
⭐
+
Finding this useful?
+
+ Star the repo to help others discover it +
+ Star on GitHub + +

const checked = selectedModels.has(m.id) ? "checked" : ""; return ` - ${esc(m.id)} + ${esc(m.id)}📋 ${esc(m.provider)} ${ctx} ${inp} From b004739e1c0e52cc00a46b2c8a71850d5020b704 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 22:46:16 +0800 Subject: [PATCH 241/311] Add audio output filter to interactive catalog MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - New '🔊 Audio Out' filter button for 34 models with audio output - Filter checks modalities.output includes 'audio' - Completes the capability filter set alongside Audio In and Video In --- site/index.html | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/site/index.html b/site/index.html index 1d9ce323..b0702d9c 100644 --- a/site/index.html +++ b/site/index.html @@ -788,6 +788,7 @@

🤖 AI Models Catalog🖼️ Image Gen +

!(m.modalities && m.modalities.input && m.modalities.input.includes("video")) ) return false; + if ( + f === "audio_output" && + !(m.modalities && m.modalities.output && m.modalities.output.includes("audio")) + ) + return false; if (f === "open_weights" && !m.open_weights) return false; if (f === "reasoning" && !m.reasoning) return false; if (f === "tool_call" && !m.tool_call) return false; From 1f62fd5ba676c9b4c1c0b4ee2338772ca4424ee0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 22:48:51 +0800 Subject: [PATCH 242/311] Add SEO-optimized 'AI Models by Provider' page (95 providers, 20 detailed) - Targets 'OpenAI models list', 'Anthropic models', 'AI provider comparison' searches - Provider overview table with model counts, cheapest pricing, max context - 20 detailed provider sections with full model tables - Quick-nav links for top 12 providers - Aggregator providers labeled distinctly - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all other SEO pages --- site/ai-models-by-provider.html | 4480 +++++++++++++++++++++++++++++++ site/sitemap.xml | 5 + 2 files changed, 4485 insertions(+) create mode 100644 site/ai-models-by-provider.html diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html new file mode 100644 index 00000000..f135bb88 --- /dev/null +++ b/site/ai-models-by-provider.html @@ -0,0 +1,4480 @@ + + + + + + AI Models by Provider — All 95 Providers Listed (2025) | AI Models Catalog + + + + + + + + + + + + + + +

🏢 AI Models by Provider — All 95 Providers Listed

+

+ Browse 4,587 AI models across 95 providers. First-party data with real pricing, + context windows, and capabilities. +

+ +
+
95Providers
+
4,587Models
+
81Free Models
+
527Open Weights
+
+ + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
+ OpenAI + Anthropic + Google + Meta + DeepSeek + Mistral + xAI + AWS Bedrock + Groq + Together AI + Fireworks + Cerebras +
+ +

📊 Provider Overview

+

All 95 providers sorted by number of models. Click a provider to see their models.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ProviderModelsCheapest Input $/1MMax ContextTool CallFree
+ nanogpt (aggregator) + 547Aggregator?0
+ aihubmix (aggregator) + 476Aggregator?132
+ openrouter (aggregator) + 356Aggregator10M263✅
+ martian (aggregator) + 304Aggregator?0
+ requesty (aggregator) + 277Aggregator1M251
+ 302ai (aggregator) + 268Aggregator2M190
+ auriko (aggregator) + 181Aggregator1M154✅
+ llmgateway (aggregator) + 163Aggregator?158✅
+ aimlapi + 147$0.0072M21✅
+ fastrouter (aggregator) + 120Aggregator2M94✅
+ orcarouter (aggregator) + 120Aggregator1M102
+ cortecs (aggregator) + 105Aggregator?97
+ novitaai + 104$0.021M72✅
+ vultr + 98$0.551M11
+ deepinfra + 88$0.011M0
+ venice (aggregator) + 75Aggregator2M64
+ jiekou (aggregator) + 73Aggregator2M73
+ meganova (aggregator) + 63Aggregator1M60✅
+ alibaba + 62$0.151M62
+ ppio + 60$0.21451M46✅
+ amazon-bedrock + 57$0.0351M37
+ google-vertex + 38$0.071M32
+ siliconflow-cn + 37$0.5262K2
+ stepfun + 31$0.7256K0✅
+ cloudflare + 30$0.017327K15
+ databricks + 29$0.05200K4
+ gmicloud + 29$0.071M11
+ openai + 28$0.021M18
+ siliconflow + 27$0.041M24
+ togetherai + 24$0.03262K22
+ nebius + 23$0.021M21
+ google + 21$0.0752M8✅
+ minimax + 21$2.1204K0
+ voyage + 21$0.02?0✅
+ digitalocean + 20$0.051M14
+ inferencenet + 20$0.01131K15
+ zhipuai + 20$0.11M20✅
+ tencent-tokenhub + 19$11M16
+ mistral + 16$0.04256K12✅
+ moonshotai + 16$2262K0
+ neuralwatt + 14$0.03?14
+ tencent + 14$0.5250K3✅
+ scaleway + 13$0.15131K6
+ chutes + 12$0.08262K12
+ clarifai + 12$0.091M9
+ cloudferro-sherlock + 12$0.261M5
+ groq + 12$0.05131K8
+ klusterai + 12$0.0081M4
+ meta + 12$0.110M9
+ microsoft + 12$0.075128K6
+ ovhcloud + 12$0.05262K0
+ anthropic + 11$11M11
+ baichuan + 11$0.98131K0✅
+ cerebras + 11$0.1131K9✅
+ hpc-ai + 11$0.141M11
+ hyperbolic + 11$0.1163K0
+ fireworks + 10$0.071M10
+ baseten + 9$0.11M9
+ baidu + 8$0.1261M7✅
+ evroc + 8$0.1131K3
+ friendli + 8$0.1262K8
+ upstage + 8$0.1128K3
+ amazon + 7$0.0351M7
+ arcee + 7$0.04262K6✅
+ berget + 7$0.2?7
+ morph + 7$0.21M5
+ nousresearch + 7$0.06131K7
+ sambanova + 7$0.22196K0
+ dinference + 6$0.07204K3
+ iflytek + 6$0.8262K0✅
+ submodel + 6$0.1262K0
+ textsynth + 6$0.2131K0
+ writer + 6$0.61M3
+ xai + 6$0.2131K6
+ 01ai + 5$132K4
+ aion + 5$0.7131K0
+ bytedance + 5$0.07262K4
+ inception + 5$0.25128K3
+ mixlayer + 5$0.1131K5✅
+ privatemode + 5$0.43131K3
+ xiaomi + 5$0.11M5
+ deepseek + 4$0.141M4
+ perplexity + 4$1200K4
+ inclusionai + 3$0.01262K3
+ ai21 + 2$0.2256K0
+ reka + 2$0.03131K1
+ wafer + 2$0.6262K2
+ +

🏢 OpenAI

+

GPT-4, GPT-4o, o1, o3 — the industry standard for LLMs. 28 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
text-embedding-3-small$0.02$08K
gpt-4.1-nano$0.1$0.41M✅
text-embedding-ada-002$0.1$08K
text-embedding-3-large$0.13$08K
gpt-4o-mini$0.15$0.6128K✅
gpt-4.1-mini$0.4$1.61M✅
gpt-3.5-turbo$0.5$1.516K✅
o3-mini$1.1$4.4200K✅✅
o4-mini$1.1$4.4200K✅✅
codex-mini$1.5$6192K✅
o1-mini$1.5$6128K✅✅
gpt-4.1$2$81M✅
gpt-4o-audio$2.5$10128K✅
gpt-4o$2.5$10128K✅
gpt-3.5-turbo-16k$3$416K✅
gpt-4o-realtime$5$20128K✅
gpt-4-turbo$10$30128K✅
o3$10$40200K✅✅
o1-realtime$15$60200K✅✅
o1$15$60200K✅✅
gpt-4$30$608K✅
gpt-4-32k$60$12032K
o1-pro$150$600200K✅✅
dall-e-2$?$??
dall-e-3$?$??
tts-1-hd$?$??
tts-1$?$??
whisper-1$?$??
+ +

🏢 Anthropic

+

Claude — known for safety, reasoning, and long context. 11 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
claude-haiku-4-5$1$5200K✅✅
claude-sonnet-4-0$3$151M✅✅
claude-sonnet-4-5$3$151M✅✅
claude-sonnet-4-6$3$151M✅✅
claude-opus-4-5$5$25200K✅✅
claude-opus-4-6$5$251M✅✅
claude-opus-4-7$5$251M✅✅
claude-opus-4-0$15$75200K✅✅
claude-opus-4-1$15$75200K✅✅
claude-opus-4-6-fast$30$1501M✅✅
claude-opus-4-7-fast$30$1501M✅✅
+ +

🏢 Google

+

Gemini — multimodal models with massive context windows. 21 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
gemini-1.5-flash-8b$0.075$0.31M✅
gemini-1.5-flash$0.075$0.31M✅
gemini-2.0-flash-lite$0.075$0.31M✅
gemini-2.0-flash$0.1$0.41M✅
gemini-2.5-flash-lite$0.1$0.41M✅
gemini-2.5-flash$0.15$3.51M✅✅
gemini-1.5-pro$1.25$52M✅
gemini-2.5-pro$1.25$101M✅✅
chirp-3.0-HD$?$??
gemma-3-12b-itFree131K
gemma-3-1b-itFree131K
gemma-3-27b-itFree131K
gemma-3-4b-itFree131K
gemma-3n-E2B-itFree131K
gemma-3n-E4B-itFree131K
imagen-3.0-fast-generate$?$??
imagen-3.0-generate$?$??
imagen-4.0-fast-generate$?$??
imagen-4.0-generate$?$??
lyria-2.0$?$??
veo-2.0-generate$?$??
+ +

🏢 Meta

+

Llama — open-weight models you can run anywhere. 12 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
meta-llama-3.2-1b$0.1$0.1128K
meta-llama-3.2-3b$0.15$0.15128K
meta-llama-3.2-11b-vision$0.16$0.16128K✅
meta-llama-4-scout$0.17$0.6610M✅
meta-llama-3.1-8b$0.22$0.22128K✅
meta-llama-4-maverick$0.24$0.971M✅
meta-llama-3-8b$0.3$0.68K
meta-llama-3.1-70b$0.72$0.72128K✅
meta-llama-3.2-90b-vision$0.72$0.72128K✅
meta-llama-3.3-70b$0.72$0.72128K✅
meta-llama-3.1-405b$2.4$2.4128K✅
meta-llama-3-70b$2.65$3.58K✅
+ +

🏢 DeepSeek

+

High-performance reasoning at competitive prices. 4 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
deepseek-chat$0.14$0.281M✅
deepseek-reasoner$0.14$0.281M✅✅
deepseek-v4-flash$0.14$0.281M✅✅
deepseek-v4-pro$0.435$0.871M✅✅
+ +

🏢 Mistral

+

European AI with open and commercial models. 16 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
ministral-3b$0.04$0.04128K✅
voxtral-mini$0.04$0.04128K
ministral-8b$0.1$0.1128K✅
voxtral-small$0.1$0.3128K
mistral-7b$0.15$0.232K
mistral-nemo$0.15$0.15128K✅
mistral-small$0.2$0.6128K✅
mistral-medium$0.4$2128K✅
mixtral-8x7b$0.45$0.732K✅
magistral-small$0.5$1.5128K✅✅
mixtral-8x22b$0.8$1.264K✅
mistral-large$2$6128K✅
pixtral-large$2$6128K✅
mistral-large-2407$4$12128K✅
codestralFree256K
devstralFree128K✅
+ +

🏢 xAI

+

Grok — models with real-time knowledge. 6 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
xai-grok-4-fast$0.2$0.5131K✅
xai-grok-4.1$0.2$0.5131K✅✅
xai-grok-3-mini$0.25$1.27131K✅✅
xai-grok-4.2$2$6131K✅✅
xai-grok-3$3$15131K✅✅
xai-grok-4$3$15131K✅✅
+ +

🏢 AWS Bedrock

+

Managed access to multiple foundation models. 57 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
amazon-nova-micro$0.035$0.14128K✅
google-gemma-3-4b$0.04$0.08131K
mistral-voxtral-mini$0.04$0.04128K
amazon-nova-lite$0.06$0.24300K✅
nvidia-nemotron-nano-2$0.06$0.234K
nvidia-nemotron-nano-3-30b$0.06$0.244K
openai-gpt-oss-20b$0.07$0.3131K✅
openai-gpt-oss-safeguard-20b$0.07$0.2131K✅
zai-glm-4-7-flash$0.07$0.4131K✅
google-gemma-3-12b$0.09$0.29131K
meta-llama-3-2-1b$0.1$0.1128K
mistral-ministral-3b$0.1$0.1128K
mistral-voxtral-small$0.1$0.3128K
meta-llama-3-2-3b$0.15$0.15128K
mistral-ministral-8b$0.15$0.15128K
mistral-mistral-7b$0.15$0.232K
nvidia-nemotron-3-super-120b$0.15$0.654K
openai-gpt-oss-120b$0.15$0.6131K✅
openai-gpt-oss-safeguard-120b$0.15$0.6131K✅
qwen-qwen3-32b$0.15$0.6131K✅
qwen-qwen3-coder-30b-a3b$0.15$0.6131K✅
writer-palmyra-vision-7b$0.15$0.68K
meta-llama-3-2-11b$0.16$0.16128K✅
meta-llama-4-scout-17b$0.17$0.661M✅
mistral-ministral-14b$0.2$0.2128K
nvidia-nemotron-nano-2-vl$0.2$0.64K
meta-llama-3-1-8b$0.22$0.22128K✅
google-gemma-3-27b$0.23$0.38131K
meta-llama-4-maverick-17b$0.24$0.971M✅
meta-llama-3-8b$0.3$0.68K
minimax-m2-1$0.3$1.21M✅
minimax-m2-5$0.3$1.21M✅
minimax-m2$0.3$1.21M✅
amazon-nova-2-lite$0.33$2.7564K✅
mistral-devstral$0.4$2128K✅
mistral-mixtral-8x7b$0.45$0.732K
mistral-magistral-small$0.5$1.5128K✅
mistral-mistral-large-3$0.5$1.5128K✅
qwen-qwen3-coder-next$0.5$1.2131K✅
qwen-qwen3-vl-235b-a22b$0.53$2.66131K✅
kimi-k2-thinking$0.6$2.5131K✅
moonshot-kimi-k2-5$0.6$3131K✅
zai-glm-4-7$0.6$2.2131K✅
deepseek-v3-2$0.62$1.8565K✅
meta-llama-3-1-70b$0.72$0.72128K✅
meta-llama-3-2-90b$0.72$0.72128K✅
meta-llama-3-3-70b$0.72$0.72128K✅
amazon-nova-pro$0.8$3.2300K✅
meta-llama-3-1-70b-latency-optimized$0.9$0.9128K✅
amazon-nova-pro-latency-optimized$1$4300K✅
mistral-mistral-small$1$3128K✅
zai-glm-5$1$3.2131K✅
deepseek-r1$1.35$5.465K
mistral-pixtral-large$2$6128K✅
amazon-nova-premier$2.5$12.51M✅
meta-llama-3-70b$2.65$3.58K
mistral-mistral-large$4$12128K✅
+ +

🏢 Groq

+

Ultra-fast inference with LPU hardware. 12 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
llama-3.1-8b-instant$0.05$0.08131K✅
gpt-oss-20b$0.075$0.3131K✅
gpt-oss-safeguard-20b$0.075$0.3131K✅
llama-4-scout-17b-16e-instruct$0.11$0.34131K✅
gpt-oss-120b$0.15$0.6131K✅
qwen3-32b$0.29$0.59131K✅
llama-3.3-70b-versatile$0.59$0.79131K✅
kimi-k2-instruct-0905$1$3131K✅
orpheus-ar-sa$?$??
orpheus-en$?$??
whisper-large-v3-turbo$?$??
whisper-large-v3$?$??
+ +

🏢 Together AI

+

Open-weight model hosting platform. 24 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
liquid-ai--LFM2-24B-A2B$0.03$0.12131K✅
openai--gpt-oss-20b$0.05$0.2131K✅
google--gemma-3n-E4B-it$0.06$0.12131K
Qwen--Qwen3.5-9B$0.1$0.15131K✅
meta-llama--Meta-Llama-3.1-8B-Instruct-Lite$0.1$0.1131K✅
essential-ai--Rnj-1-Instruct$0.15$0.15131K
openai--gpt-oss-120b$0.15$0.6131K✅
Qwen--Qwen3-235B-A22B-FP8-Throughput$0.2$0.6131K✅
MiniMaxAI--MiniMax-M2.5$0.3$1.2131K✅
MiniMaxAI--MiniMax-M2.7$0.3$1.2131K✅
Qwen--Qwen2.5-7B-Instruct-Turbo$0.3$0.3131K✅
google--gemma-4-31B-it$0.39$0.97131K✅
Qwen--Qwen3-Coder-Next$0.5$1.2131K✅
Qwen--Qwen3.6-Plus$0.5$3131K✅
moonshotai--Kimi-K2.5$0.5$2.8131K✅
Qwen--Qwen3.5-397B-A17B$0.6$3.6131K✅
deepseek-ai--DeepSeek-V3.1$0.6$1.7131K✅
meta-llama--Llama-3.3-70B-Instruct-Turbo$0.88$0.88131K✅
zai-org--GLM-5$1$3.2131K✅
moonshotai--Kimi-K2.6$1.2$4.5262K✅
cogito-ai--Cogito-v2.1-671B$1.25$1.25131K✅✅
zai-org--GLM-5.1$1.4$4.4131K✅
Qwen--Qwen3-Coder-480B-A35B-Instruct$2$2131K✅
deepseek-ai--DeepSeek-V4-Pro$2.1$4.4131K✅✅
+ +

🏢 Fireworks

+

Fast inference for open-source models. 10 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
gpt-oss-20b$0.07$0.3131K✅
gpt-oss-120b$0.15$0.6131K✅
llama4-scout-17b-16e-instruct$0.18$0.59131K✅
minimax-m2.5$0.3$1.2196K✅
minimax-m2.7$0.3$1.2196K✅
qwen3.6-plus$0.5$3131K✅
kimi-k2.5$0.6$3262K✅
kimi-k2.6$0.95$4262K✅
glm-5.1$1.4$4.4202K✅
deepseek-v4-pro$1.74$3.481M✅✅
+ +

🏢 Cerebras

+

Wafer-scale inference at extreme speed. 11 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
llama3.1-8b$0.1$0.1131K✅
gpt-oss-120b$0.35$0.75131K✅
qwen3-235b-instruct$0.6$1.2131K✅
zai-glm-4.7$2.25$2.75131K✅
deepseek-r1-distill-llama-70bFree131K✅
deepseek-r1-distill-llama-8bFree131K✅
llama-3.3-70bFree131K✅
llama-4-scout-17b-16e-instructFree131K✅
qwen-2.5-32bFree131K✅
qwen-2.5-coder-32bFree131K✅
qwen3-32bFree131K✅
+ +

🏢 Databricks

+

DBRX and enterprise AI models. 29 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
databricks-gpt-5-nano$0.05$0.4200K
databricks-gpt-oss-20b$0.07$0.3131K
databricks-gemma-3-12b$0.15$0.5131K
databricks-gpt-oss-120b$0.15$0.6131K
databricks-meta-llama-3-1-8b-instruct$0.15$0.45131K✅
databricks-qwen3-next-80b-a3b-instruct$0.15$1.2131K✅
databricks-gpt-5-4-nano$0.2$1.25128K
databricks-gemini-3-1-flash-lite$0.25$1.5128K
databricks-gpt-5-1-codex-mini$0.25$2200K
databricks-gpt-5-mini$0.25$2200K
databricks-gemini-2-5-flash$0.3$2.5128K
databricks-llama-4-maverick$0.5$1.5131K✅
databricks-meta-llama-3-3-70b-instruct$0.5$1.5131K✅
databricks-gemini-3-flash$0.63$3.75128K
databricks-gpt-5-4-mini$0.75$4.5128K
databricks-claude-haiku-4-5$1$5200K
databricks-gemini-2-5-pro$1.25$10128K
databricks-gpt-5-1-codex-max$1.25$10200K
databricks-gpt-5-1$1.25$10200K
databricks-gpt-5$1.25$10200K
databricks-gpt-5-2-codex$1.75$14200K
databricks-gpt-5-2$1.75$14200K
databricks-gemini-3-1-pro$2.5$15128K
databricks-gpt-5-4$2.5$15128K
databricks-claude-sonnet-4-5$3$15200K
databricks-claude-sonnet-4$3$15200K
databricks-claude-opus-4-5$5$25200K
databricks-gpt-5-5$5$30128K
databricks-claude-opus-4-1$15$75200K
+ +

🏢 Alibaba (Qwen)

+

Qwen — multilingual models from Alibaba Cloud. 62 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
qwen-flash$0.15$1.5?✅✅
qwen3.5-flash-2026-02-23$0.2$21M✅
qwen3.5-flash$0.2$21M✅
qwen-flash-character$0.25$1.5?✅✅
qwen-turbo$0.3$0.6?✅✅
qwen3-0.6b$0.3$1.2?✅✅
qwen3-1.7b$0.3$1.2?✅✅
qwen3-4b$0.3$1.2?✅✅
qwen-omni-turbo$0.4$25?✅✅
qwen3.5-35b-a3b$0.4$3.2256K✅
qwen-long-2025-01-25$0.5$2?✅✅
qwen-long-latest$0.5$2?✅✅
qwen-long$0.5$2?✅✅
qwen2.5-7b-instruct-1m$0.5$1?✅✅
qwen2.5-7b-instruct$0.5$1?✅✅
qwen3-8b$0.5$2?✅✅
qwen-mt-lite$0.6$1.6?✅✅
qwen2.5-omni-7b$0.6$38?✅✅
qwen3.5-27b$0.6$4.8256K✅
qwen-mt-flash$0.7$1.95?✅✅
qwen-mt-turbo$0.7$1.95?✅✅
qwen3-30b-a3b-instruct-2507$0.75$3?✅✅
qwen3-30b-a3b$0.75$3?✅✅
qwen-plus-character$0.8$2?✅✅
qwen-plus$0.8$2?✅✅
qwen3.5-122b-a10b$0.8$6.4256K✅
qwen3.5-plus-2026-02-15$0.8$4.81M✅
qwen3.5-plus$0.8$4.81M✅
qwen2.5-14b-instruct-1m$1$3?✅✅
qwen2.5-14b-instruct$1$3?✅✅
qwen3-14b$1$4?✅✅
qwen3-coder-flash-2025-07-28$1$4?✅✅
qwen3-coder-flash$1$4?✅✅
qwen3-coder-next$1$4?✅✅
qwen3-next-80b-a3b-instruct$1$4?✅✅
qwen2.5-vl-3b-instruct$1.2$3.6?✅✅
qwen3.5-397b-a17b$1.2$7.2256K✅
qwen3.6-flash-2026-04-16$1.2$7.21M✅
qwen3.6-flash$1.2$7.21M✅✅
qwen3-coder-30b-a3b-instruct$1.5$6?✅✅
qwen-mt-plus$1.8$5.4?✅✅
qwen2.5-32b-instruct$2$6?✅✅
qwen2.5-vl-7b-instruct$2$5?✅✅
qwen3-235b-a22b-instruct-2507$2$8?✅✅
qwen3-235b-a22b$2$8?✅✅
qwen3-32b$2$8?✅✅
qwen3.6-plus-2026-04-02$2$121M✅
qwen3.6-plus$2$121M✅✅
qwen-max$2.4$9.6?✅✅
qwen3-max-2026-01-23$2.5$10?✅✅
qwen3-max$2.5$10?✅✅
qwen-plus-character-ja$3.67$10.275?✅✅
qwen2.5-72b-instruct$4$12?✅✅
qwen3-coder-plus-2025-07-22$4$16?✅✅
qwen3-coder-plus-2025-09-23$4$16?✅✅
qwen3-coder-plus$4$16?✅✅
qwen3-coder-480b-a35b-instruct$6$24?✅✅
qwen3-max-2025-09-23$6$24?✅✅
qwen3-max-preview$6$24?✅✅
qwen2.5-vl-32b-instruct$8$24?✅✅
qwen3.6-max-preview$9$54256K✅✅
qwen2.5-vl-72b-instruct$16$48?✅✅
+ +

🏢 ByteDance

+

Doubao — models from the TikTok parent company. 5 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
seed-1.6-flash$0.07$0.3262K✅✅
seed-2.0-mini$0.1$0.4262K✅✅
ui-tars-1.5-7b$0.1$0.2128K
seed-1.6$0.25$2262K✅✅
seed-2.0-lite$0.25$2262K✅✅
+ +

🏢 MiniMax

+

Chinese AI startup with competitive models. 21 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
M2-her$2.1$8.464K
MiniMax-M2.1$2.1$8.4204K
MiniMax-M2.5$2.1$8.4204K
MiniMax-M2.7$2.1$8.4204K
MiniMax-M2$2.1$8.4204K
MiniMax-M2.1-highspeed$4.2$16.8204K
MiniMax-M2.5-highspeed$4.2$16.8204K
MiniMax-M2.7-highspeed$4.2$16.8204K
MiniMax-Hailuo-02$?$??
MiniMax-Hailuo-2.3-Fast$?$??
MiniMax-Hailuo-2.3$?$??
image-01-live$?$??
image-01$?$??
music-2.6$?$??
music-cover$?$??
speech-02-hd$?$??
speech-02-turbo$?$??
speech-2.6-hd$?$??
speech-2.6-turbo$?$??
speech-2.8-hd$?$??
speech-2.8-turbo$?$??
+ +

🏢 Moonshot AI

+

Kimi — long-context Chinese models. 16 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
moonshot-v1-8k-vision-preview$2$108K
moonshot-v1-8k$2$108K
kimi-k2-0711-preview$4$16131K
kimi-k2-0905-preview$4$16262K
kimi-k2-thinking$4$16262K✅
kimi-k2.5$4$21262K✅
kimi-vl-a3b-thinking$4$21131K✅
kimi-vl-a3b$4$21131K
moonshot-v1-32k-vision-preview$5$2032K
moonshot-v1-32k$5$2032K
kimi-k2.6-long$6.5$27262K✅
kimi-k2.6$6.5$27262K✅
kimi-k2-thinking-turbo$8$58262K✅
kimi-k2-turbo-preview$8$58262K
moonshot-v1-128k-vision-preview$10$30131K
moonshot-v1-128k$10$30131K
+ +

🏢 StepFun

+

Step — Chinese AI models with strong capabilities. 31 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
step-3.5-flash-2603$0.7$2.1256K
step-3.5-flash$0.7$2.1256K
step-2-mini$1$232K
step-3$1.5$464K
step-1o-turbo-vision$2.5$832K
step-r1-v-mini$2.5$8100K
step-1-8k$5$208K
step-1v-8k$5$208K
step-audio-2$10$70?
stepaudio-2.5-chat$10$25?
stepaudio-2.5-realtime$10$70?
step-1-32k$15$7032K
step-1o-vision-32k$15$7032K
step-1v-32k$15$7032K
step-1o-audio$25$60?
step-2-16k-exp$38$12016K
step-2-16k$38$12016K
step-1x-editFree?
step-1x-medium$?$??
step-2x-largeFree?
step-asr-1.1-stream$?$??
step-asr-1.1$?$??
step-asr$?$??
step-audio-r1.1Free?
step-guiFree?
step-image-edit-2$?$??
step-tts-2$?$??
step-tts-mini$?$??
stepaudio-2-asr-pro$?$??
stepaudio-2.5-asr$?$??
stepaudio-2.5-tts$?$??
+ +

🏢 Baidu

+

ERNIE — models from China's search giant. 8 models available.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInput $/1MOutput $/1MContextTool CallReasoning
deepseek-v4-flash$0.126$0.2521M✅✅
deepseek-v3.2$0.252$0.378131K✅✅
minimax-m2.5$0.27$1.08196K✅✅
qianfan-ocr-fast$0.6799999999999999$2.8165K
glm-5$0.7$2.24202K✅✅
glm-5.1$0.98$3.08202K✅✅
deepseek-v4-pro$1.521$3.042716K✅✅
cobuddyFree131K✅
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs — not third-party aggregators. Pricing, + context windows, and capabilities are verified against official provider documentation. + Aggregator providers (OpenRouter, Requesty, etc.) are labeled as such — they provide access to + other providers' models. +

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index b563cf98..524f68a2 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -215,4 +215,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/ai-models-by-provider.html + weekly + 0.9 + From 1d9e0416aae7938bc6ac886fe0a007d44c6102c0 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 22:51:17 +0800 Subject: [PATCH 243/311] Add SEO-optimized 'Context Window Comparison' page - Targets 'AI model context window', 'largest context LLM' searches - Top 20 largest context windows ranked - 7 context tier sections (1M+, 512K, 256K, 128K, 64K, 32K, 8K) - Cheapest model per context tier table - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all other SEO pages --- site/context-window-comparison.html | 1556 +++++++++++++++++++++++++++ site/sitemap.xml | 5 + 2 files changed, 1561 insertions(+) create mode 100644 site/context-window-comparison.html diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html new file mode 100644 index 00000000..2ef4ff95 --- /dev/null +++ b/site/context-window-comparison.html @@ -0,0 +1,1556 @@ + + + + + + + AI Model Context Window Comparison — Largest Context LLMs (2025) | AI Models Catalog + + + + + + + + + + + + + + + +

📏 AI Model Context Window Comparison

+

+ Compare context windows across 4,587 AI models. Find the largest context LLMs for your + use case — from 1M+ token monsters to compact 8K models. +

+ +
+
4,587Models
+
2,195128K+ Context
+
95Providers
+
+ + 🔍 Interactive Catalog + ⭐ Star on GitHub + +

🏆 Top 20 Largest Context Windows

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
#ModelProviderContextInput $/1MTool Call
1meta-llama-4-scoutmeta10M$0.17✅
2gemini-1.5-progoogle2M$1.25✅
3xai--grok-4-fast-non-reasoningaimlapi2M$0.52✅
4xai--grok-4-fast-reasoningaimlapi2M$0.52✅
5meta-llama-4-maverick-17bamazon-bedrock1M$0.24✅
6meta-llama-4-scout-17bamazon-bedrock1M$0.17✅
7minimax-m2-1amazon-bedrock1M$0.3✅
8minimax-m2-5amazon-bedrock1M$0.3✅
9minimax-m2amazon-bedrock1M$0.3✅
10deepseek-v4-flashbaidu1M$0.126✅
11minimax-m2-5baseten1M$0.3✅
12gpt-5-1clarifai1M$1.5625✅
13deepseek-v4-flashdeepinfra1M$0.14
14llama-4-maverick-17b-128e-instruct-fp8deepinfra1M$0.15
15mimo-v2.5-prodeepinfra1M$1
16llama-4-maverickdigitalocean1M$0.25✅
17deepseek-v4-profireworks1M$1.74✅
18meta-llama--Llama-4-Maverick-17B-128E-Instruct-FP8gmicloud1M$0.25✅
19gemini-1.5-flash-8bgoogle1M$0.075✅
20gemini-1.5-flashgoogle1M$0.075✅
+ +

📊 1M+ Tokens (93 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
meta-llama-4-scoutmeta10M$0.17$0.66✅
gemini-1.5-progoogle2M$1.25$5✅
xai--grok-4-fast-non-reasoningaimlapi2M$0.52$1.3✅
xai--grok-4-fast-reasoningaimlapi2M$0.52$1.3✅
meta-llama-4-maverick-17bamazon-bedrock1M$0.24$0.97✅
meta-llama-4-scout-17bamazon-bedrock1M$0.17$0.66✅
minimax-m2-1amazon-bedrock1M$0.3$1.2✅
minimax-m2-5amazon-bedrock1M$0.3$1.2✅
minimax-m2amazon-bedrock1M$0.3$1.2✅
deepseek-v4-flashbaidu1M$0.126$0.252✅✅
minimax-m2-5baseten1M$0.3$1.2✅
gpt-5-1clarifai1M$1.5625$12.5✅
deepseek-v4-flashdeepinfra1M$0.14$0.28✅
llama-4-maverick-17b-128e-instruct-fp8deepinfra1M$0.15$0.6
mimo-v2.5-prodeepinfra1M$1$3✅
... and 78 more models
+ +

📊 512K–1M Tokens (1 models)

+ + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
deepseek-v4-probaidu716K$1.521$3.042✅✅
+ +

📊 256K–512K Tokens (187 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
openai--gpt-5-chataimlapi400K$1.625$13
openai--gpt-5-miniaimlapi400K$0.325$2.6✅
openai--gpt-5-nanoaimlapi400K$0.065$0.52✅
openai--gpt-5.1-chat-latestaimlapi400K$1.625$13✅
openai--gpt-5.1aimlapi400K$1.625$13✅
openai--gpt-5.2aimlapi400K$2.275$18.2✅
openai--gpt-5aimlapi400K$1.625$13✅
llama-4-scout-17b-16e-instructcloudflare327K$0.27$0.85✅
llama-4-scout-17b-16e-instructdeepinfra327K$0.08$0.3
meta-llama--Llama-4-Scout-17B-16E-Instructgmicloud327K$0.08$0.5✅
llama-4-scout-17b-16e-instructvultr327K$0.55$2.75✅
llama-4-scout-17b-16evultr327K$0.55$2.75
amazon-nova-liteamazon300K$0.06$0.24✅
amazon-nova-proamazon300K$0.8$3.2✅
amazon-nova-liteamazon-bedrock300K$0.06$0.24✅
... and 172 more models
+ +

📊 128K–256K Tokens (685 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
hunyuan-litetencent250KFree
hunyuan-a13btencent224K$0.5$2✅
minimax-m2.5dinference204K$0.22$0.88
minimax--minimax-m2.5hpc-ai204K$0.3$1.2✅✅
MiniMax-M2.1-highspeedminimax204K$4.2$16.8
MiniMax-M2.1minimax204K$2.1$8.4
MiniMax-M2.5-highspeedminimax204K$4.2$16.8
MiniMax-M2.5minimax204K$2.1$8.4
MiniMax-M2.7-highspeedminimax204K$4.2$16.8
MiniMax-M2.7minimax204K$2.1$8.4
MiniMax-M2minimax204K$2.1$8.4
minimax--minimax-m2.1novitaai204K$0.3$1.2✅
minimax--minimax-m2.5-highspeednovitaai204K$0.6$2.4✅✅
minimax--minimax-m2.5novitaai204K$0.3$1.2✅✅
minimax--minimax-m2.7novitaai204K$0.3$1.2✅✅
... and 670 more models
+ +

📊 64K–128K Tokens (56 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
sonarperplexity127K$1$1✅
baidu--ernie-4.5-300b-a47b-paddlenovitaai123K$0.28$1.1
baidu--ernie-4.5-vl-424b-a47bnovitaai123K$0.42$1.25✅
baidu--ernie-4.5-300b-a47b-paddleppio123K$2$7
baidu--ernie-4.5-vl-424b-a47bppio123K$3$9
baidu--ernie-4.5-0.3baimlapi120KFree✅
baidu--ernie-4.5-21B-a3bnovitaai120K$0.07$0.28✅
baidu--ernie-4.5-0.3bppio120KFree
baidu--ernie-4.5-21B-a3bppio120K$0.5$2
qwen3.6-27bvultr120K$0.55$2.75
step-r1-v-ministepfun100K$2.5$8
google--gemma-3-27b-itnovitaai98K$0.119$0.2
Gemma-3-27b-itnebius96K$0.1$0.3✅✅
gemma-3-27bprivatemode96K$0.77$1.27
gemma-4-31bprivatemode96K$0.77$1.27
... and 41 more models
+ +

📊 32K–64K Tokens (74 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
mistralai--mistral-nemonovitaai60K$0.04$0.17
Qwen--Qwen3-32B-TEEchutes40K$0.08$0.24✅✅
qwen3-30b-a3b-fp8cloudflare40K$0.051$0.335✅✅
qwen3-14bdeepinfra40K$0.12$0.24✅
qwen3-30b-a3bdeepinfra40K$0.09$0.45✅
qwen3-32bdeepinfra40K$0.08$0.28✅
Qwen--Qwen3-30B-A3B-Instructevroc40K$0.1$0.8
Qwen--Qwen3-VL-30B-A3B-Instructevroc40K$0.2$0.8
Qwen--Qwen3-30B-A3Bgmicloud40K$0.08$0.25
Qwen--Qwen3-32B-FP8gmicloud40K$0.1$0.6
Qwen--Qwen3-235B-A22B-FP8klusterai40K$0.13$2✅✅
mistralai--Magistral-Small-2506klusterai40K$0.1$0.3
qwen--qwen3-235b-a22b-fp8novitaai40K$0.2$0.8✅
qwen--qwen3-30b-a3b-fp8novitaai40K$0.09$0.45✅✅
qwen--qwen3-32b-fp8novitaai40K$0.1$0.45✅
... and 59 more models
+ +

📊 8K–32K Tokens (79 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
baidu--ernie-4.5-vl-28b-a3bnovitaai30K$0.14$0.56✅✅
baidu--ernie-4.5-vl-28b-a3bppio30K$1$4
gpt-oss-120bvultr30K$0.55$2.75
gpt-oss-20bvultr30K$0.55$2.75
hunyuan-large-role-latesttencent28K$2.4$9.6
hunyuan-t1-visiontencent28K$3$9✅
hunyuan-roletencent-tokenhub28K$2.4$9.6
hunyuan-turbos-vision-videotencent24K$3$9
hunyuan-turbos-visiontencent24K$3$9
hunyuan-vision-1.5-instructtencent24K$3$9
autoglm-phonezhipuai20KFree✅
gpt-3.5-turbo-16kopenai16K$3$4✅
gpt-3.5-turboopenai16K$0.5$1.5✅
yi-lightning01ai16K$1$1✅
yi-medium01ai16K$2.5$2.5✅
... and 64 more models
+ +

📊 Under 8K Tokens (13 models)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MOutput $/1MTool CallReasoning
nvidia-nemotron-3-super-120bamazon-bedrock4K$0.15$0.65
nvidia-nemotron-nano-2-vlamazon-bedrock4K$0.2$0.6
nvidia-nemotron-nano-2amazon-bedrock4K$0.06$0.23
nvidia-nemotron-nano-3-30bamazon-bedrock4K$0.06$0.24
llama-2-7b-chat-fp16cloudflare4K$0.556$6.667
mythomax-l2-13bdeepinfra4K$0.4$0.4
nvidia-nemotron-3-super-120bdigitalocean4K$0.3$0.65
nemotron3-superinferencenet4K$2.5$5
gryphe--mythomax-l2-13bnovitaai4K$0.09$0.09
nemotron-3-super-120b-a12b-bf16vultr4K$0.55$2.75
hunyuan-translation-litetencent4K$1$3
hunyuan-translationtencent4K$1.2$3.6
EleutherAI--gpt-j-6Btextsynth2K$0.2$2
+ +

💰 Cheapest Models by Context Tier

+

Find the most affordable model in each context window tier.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Context TierCheapest ModelProviderContextInput $/1M
1M+ Tokensgemini-1.5-flash-8bdeepinfra1M$0.0375
512K–1M Tokensdeepseek-v4-probaidu716K$1.521
256K–512K Tokensqwen3.5-0.8bdeepinfra262K$0.01
128K–256K Tokensmistralai--Mistral-Nemo-Instruct-2407klusterai131K$0.008
64K–128K Tokenszai-org--autoglm-phone-9b-multilingualnovitaai65K$0.035
32K–64K Tokensmeta-llama--llama-3.2-3b-instructnovitaai32K$0.03
8K–32K TokensGemma-2-2b-itnebius8K$0.02
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs — not third-party aggregators. Context + windows are as reported by each provider. Aggregator providers are excluded from ranking + tables to avoid duplicate models. +

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index 524f68a2..00e25bdd 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -220,4 +220,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/context-window-comparison.html + weekly + 0.9 + From 2b8d30f1c1b4b24a481fffc5944b190f38e65b65 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 22:54:17 +0800 Subject: [PATCH 244/311] Add SEO-optimized 'Best AI Models for Coding' page - Targets 'best AI model for coding', 'best LLM for programming' searches - 6 sections: flagship, best value, free, open weights, large context, agentic - Focus on tool_call + reasoning + large context for coding use cases - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all other SEO pages --- site/best-ai-models-for-coding.html | 774 ++++++++++++++++++++++++++++ site/sitemap.xml | 5 + 2 files changed, 779 insertions(+) create mode 100644 site/best-ai-models-for-coding.html diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html new file mode 100644 index 00000000..fafe619c --- /dev/null +++ b/site/best-ai-models-for-coding.html @@ -0,0 +1,774 @@ + + + + + + Best AI Models for Coding — Top 20 Code LLMs Compared (2025) | AI Models Catalog + + + + + + + + + + + + + + +

💻 Best AI Models for Coding (2025)

+

+ Compare the top AI models for code generation, debugging, and software development. Real + pricing, context windows, and capabilities from first-party data. +

+ +
+
189Code Models
+
2,350Tool Calling
+
1,306Reasoning
+
81Free Models
+
+ + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
+ 💡 What makes a good coding model? Tool calling for agentic workflows, large + context for codebases, reasoning for complex logic, and structured output for parsing. We rank + models by these capabilities. +
+ +

🏆 Top Coding Models — Flagship Tier

+

The most capable models for complex coding tasks. Higher price, highest quality.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContextTool CallReasoning
gpt-4.1openai$2$81M✅
gpt-4oopenai$2.5$10128K✅
gemini-2.5-prodeepinfra$1.25$101M✅
deepseek-r1amazon-bedrock$1.35$5.465K
+ +

💰 Best Value for Coding

+

Great coding performance at lower prices. Perfect for high-volume code generation.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContextTool CallReasoning
gpt-4o-miniopenai$0.15$0.6128K✅
gemini-2.5-flashdeepinfra$0.3$2.51M✅
deepseek-v3deepinfra$0.32$0.89163K
deepseek-r1amazon-bedrock$1.35$5.465K
+ +

🆓 Free Models for Coding

+

Zero-cost models for learning, prototyping, and personal projects.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextTool CallReasoning
openrouter--owl-alphaopenrouter1M✅
deepseek--deepseek-v4-flash--freeopenrouter1M✅✅
qwen--qwen3-coder--freeopenrouter1M✅
nvidia--nemotron-3-super-120b-a12b--freeopenrouter1M✅✅
gemma-4-26b-a4b-itauriko262K✅✅
gemma-4-31b-itauriko262K✅✅
arcee-ai--trinity-large-thinking--freeopenrouter262K✅✅
google--gemma-4-26b-a4b-it--freeopenrouter262K✅✅
google--gemma-4-31b-it--freeopenrouter262K✅✅
nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--freeopenrouter256K✅✅
+ +

🔓 Open-Weight Models for Coding

+

Download and run locally for full privacy and zero API costs at scale.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextTool CallReasoning
google--gemma-4-31b-itorcarouter1M✅
qwen--qwen3.5-flash-2026-02-23orcarouter1M✅
qwen--qwen3.5-flashorcarouter1M✅
qwen--qwen3.6-flash-2026-04-16orcarouter1M✅
qwen--qwen3.6-flashorcarouter1M✅
meta-llama-4-maverick-17bamazon-bedrock1M✅
meta-llama-4-scout-17bamazon-bedrock1M✅
minimax-m2-1amazon-bedrock1M✅
minimax-m2-5amazon-bedrock1M✅
minimax-m2amazon-bedrock1M✅
+ +

📏 Large Context for Codebases

+

+ Models with 128K+ context for working with large codebases, multiple files, and long + conversations. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MTool Call
ling-2.6-flashinclusionai262K$0.01✅
bdc-coderinferencenet131K$0.01✅
klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai131K$0.015✅
granite-4.0-h-microcloudflare131K$0.017✅
llama-3.1-8b-instruct--fp-16inferencenet131K$0.02✅
schematron-3binferencenet131K$0.02✅
schematron-v3inferencenet131K$0.02✅
gpt-oss-20binferencenet131K$0.03✅
schematron-v2-turboinferencenet131K$0.03✅
qwen--qwen3-4b-fp8novitaai128K$0.03✅
liquid-ai--LFM2-24B-A2Btogetherai131K$0.03✅
amazon-nova-microamazon128K$0.035✅
amazon-nova-microamazon-bedrock128K$0.035✅
mistral-nemo-12b-instruct--fp-8inferencenet131K$0.0375✅
klusterai--Meta-Llama-3.3-70B-Instruct-Turboklusterai131K$0.038✅
+ +

🤖 Agentic Coding Models

+

+ Models with tool calling + reasoning — the key capabilities for AI coding agents (Cursor, + Copilot, Devin-style). +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
gpt-oss-120binferencenet$0.05$0.45131K
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
openai--gpt-oss-120bnovitaai$0.05$0.25131K
qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K
glm-4.7-flashcloudflare$0.06$0.4131K
Nemotron-3-Nano-Omninebius$0.06$0.24128K
hermes-4-llama-3.1-8bnousresearch$0.06$0.12131K
seed-1.6-flashbytedance$0.07$0.3262K
ring-2.6-1tinclusionai$0.07$0.62262K
zai-org--glm-4.7-flashnovitaai$0.07$0.4200K
microsoft-phi-4-mini-reasoningmicrosoft$0.075$0.3128K
Qwen--Qwen3-32B-TEEchutes$0.08$0.2440K
gpt-oss-120bclarifai$0.09$0.36131K
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs. Models are selected based on capabilities + relevant to coding: tool calling (for agentic workflows), reasoning (for complex logic), large + context (for codebases), and structured output (for parsing). Aggregator providers are + excluded from ranking tables. +

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index 00e25bdd..2999e79b 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -225,4 +225,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html + weekly + 0.9 + From fafe2c9bd39ffa20f414d6094f9136da05ae25a9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 22:57:01 +0800 Subject: [PATCH 245/311] Add SEO-optimized 'Best AI Models for Agents' page - Targets 'best AI model for agents', 'agentic LLM', 'tool calling model' searches - 6 sections: full stack (TC+reasoning+structured), TC+reasoning, cheapest TC, free TC, open-weight TC, large context TC - 1,080+ agentic models with tool calling capability - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all 7 other SEO pages --- site/best-ai-models-for-agents.html | 899 ++++++++++++++++++++++++++++ site/sitemap.xml | 5 + 2 files changed, 904 insertions(+) create mode 100644 site/best-ai-models-for-agents.html diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html new file mode 100644 index 00000000..e073fdcf --- /dev/null +++ b/site/best-ai-models-for-agents.html @@ -0,0 +1,899 @@ + + + + + + Best AI Models for Agents — Top Agentic LLMs Compared (2025) | AI Models Catalog + + + + + + + + + + + + + + +

🤖 Best AI Models for Agents (2025)

+

+ Compare the top AI models for building autonomous agents. 1,080+ models with tool + calling — the key capability for agentic workflows. +

+ +
+
1,080Agentic Models
+
2,350Tool Calling
+
1,306Reasoning
+
829Structured Output
+
+ + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
+ 💡 What makes a model "agentic"? The three key capabilities are: + Tool calling (invoke APIs/functions), Reasoning (plan multi-step actions), and + Structured output (return parseable JSON). Models with all three are the most capable + agents. +
+ +

🏆 Top Agentic Models — Full Stack (Tool Call + Reasoning + Structured Output)

+

Models with all three agentic capabilities. Best for complex autonomous workflows.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
openai--gpt-oss-20bneuralwatt$0.03$0.16?
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
openai--gpt-oss-120bnovitaai$0.05$0.25131K
Nemotron-3-Nano-Omninebius$0.06$0.24128K
hermes-4-llama-3.1-8bnousresearch$0.06$0.12131K
zai-org--glm-4.7-flashnovitaai$0.07$0.4200K
Qwen--Qwen3-32B-TEEchutes$0.08$0.2440K
Gemma-3-27b-itnebius$0.1$0.396K
Qwen3-32Bnebius$0.1$0.3128K
xiaomimimo--mimo-v2-flashnovitaai$0.1$0.3262K
Qwen--Qwen3-235B-A22B-Thinking-2507chutes$0.11$0.6262K
deepseek-v4-flashbaidu$0.126$0.2521M
google--gemma-4-31B-turbo-TEEchutes$0.13$0.38131K
Hermes-4-70Bnebius$0.13$0.4128K
google--gemma-4-26b-a4b-itnovitaai$0.13$0.4262K
+ +

🔧 Tool Calling + Reasoning

+

+ Models that can both call tools and reason about when/how to use them. Essential for + ReAct-style agents. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
gpt-oss-120binferencenet$0.05$0.45131K
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
openai--gpt-oss-120bnovitaai$0.05$0.25131K
qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K
glm-4.7-flashcloudflare$0.06$0.4131K
Nemotron-3-Nano-Omninebius$0.06$0.24128K
hermes-4-llama-3.1-8bnousresearch$0.06$0.12131K
seed-1.6-flashbytedance$0.07$0.3262K
ring-2.6-1tinclusionai$0.07$0.62262K
zai-org--glm-4.7-flashnovitaai$0.07$0.4200K
microsoft-phi-4-mini-reasoningmicrosoft$0.075$0.3128K
Qwen--Qwen3-32B-TEEchutes$0.08$0.2440K
gpt-oss-120bclarifai$0.09$0.36131K
+ +

💰 Cheapest Tool Calling Models

+

Most affordable models with tool calling for budget-conscious agent deployments.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
ling-2.6-flashinclusionai$0.01$0.03262K
bdc-coderinferencenet$0.01$0.01131K
klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
granite-4.0-h-microcloudflare$0.017$0.112131K
llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K
schematron-3binferencenet$0.02$0.05131K
schematron-v3inferencenet$0.02$0.05131K
gpt-oss-20binferencenet$0.03$0.15131K
schematron-v2-turboinferencenet$0.03$0.15131K
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
liquid-ai--LFM2-24B-A2Btogetherai$0.03$0.12131K
amazon-nova-microamazon$0.035$0.14128K
amazon-nova-microamazon-bedrock$0.035$0.14128K
mistral-nemo-12b-instruct--fp-8inferencenet$0.0375$0.1131K
+ +

🆓 Free Models with Tool Calling

+

Zero-cost models for building and testing agents.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextReasoningStructured Output
openrouter--owl-alphaopenrouter1M✅
deepseek--deepseek-v4-flash--freeopenrouter1M✅
qwen--qwen3-coder--freeopenrouter1M
nvidia--nemotron-3-super-120b-a12b--freeopenrouter1M✅✅
gemma-4-26b-a4b-itauriko262K✅✅
gemma-4-31b-itauriko262K✅✅
arcee-ai--trinity-large-thinking--freeopenrouter262K✅
google--gemma-4-26b-a4b-it--freeopenrouter262K✅✅
google--gemma-4-31b-it--freeopenrouter262K✅✅
nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--freeopenrouter256K✅
+ +

🔓 Open-Weight Models with Tool Calling

+

Run agent models locally for full privacy and zero API costs at scale.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextReasoningStructured Output
google--gemma-4-31b-itorcarouter1M
qwen--qwen3.5-flash-2026-02-23orcarouter1M
qwen--qwen3.5-flashorcarouter1M
qwen--qwen3.6-flash-2026-04-16orcarouter1M
qwen--qwen3.6-flashorcarouter1M
meta-llama-4-maverick-17bamazon-bedrock1M
meta-llama-4-scout-17bamazon-bedrock1M
minimax-m2-1amazon-bedrock1M
minimax-m2-5amazon-bedrock1M
minimax-m2amazon-bedrock1M
+ +

📏 Large Context + Tool Calling

+

+ Models with 128K+ context and tool calling for agents that need to process large documents or + maintain long conversation history. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MReasoning
ling-2.6-flashinclusionai262K$0.01
bdc-coderinferencenet131K$0.01
klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai131K$0.015
granite-4.0-h-microcloudflare131K$0.017
llama-3.1-8b-instruct--fp-16inferencenet131K$0.02
schematron-3binferencenet131K$0.02
schematron-v3inferencenet131K$0.02
gpt-oss-20binferencenet131K$0.03
schematron-v2-turboinferencenet131K$0.03
qwen--qwen3-4b-fp8novitaai128K$0.03✅
liquid-ai--LFM2-24B-A2Btogetherai131K$0.03
amazon-nova-microamazon128K$0.035
amazon-nova-microamazon-bedrock128K$0.035
mistral-nemo-12b-instruct--fp-8inferencenet131K$0.0375
klusterai--Meta-Llama-3.3-70B-Instruct-Turboklusterai131K$0.038
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs. Agentic capability is defined by tool + calling (function calling), reasoning (chain-of-thought), and structured output (JSON mode). + Aggregator providers are excluded from ranking tables to avoid duplicate models. +

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index 2999e79b..bad78003 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -230,4 +230,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html + weekly + 0.9 + From facefa952140a2c178785afd47f2fb6a82bdad8b Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 23:18:34 +0800 Subject: [PATCH 246/311] Add SEO-optimized 'Reasoning Models Comparison' page - Targets 'AI reasoning model', 'o1 vs R1', 'reasoning LLM comparison' searches - 6 sections: flagship head-to-head, cheapest, free, open weights, reasoning+tool_call, large context reasoning - 1,306 reasoning models with real pricing and capabilities - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all 8 other SEO pages --- site/reasoning-models-comparison.html | 876 ++++++++++++++++++++++++++ site/sitemap.xml | 5 + 2 files changed, 881 insertions(+) create mode 100644 site/reasoning-models-comparison.html diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html new file mode 100644 index 00000000..e896096d --- /dev/null +++ b/site/reasoning-models-comparison.html @@ -0,0 +1,876 @@ + + + + + + AI Reasoning Models Compared — o1, R1, Claude, Gemini (2025) | AI Models Catalog + + + + + + + + + + + + + + +

🧠 AI Reasoning Models Compared (2025)

+

+ Compare 1,306 reasoning models across 95 providers. Find the best chain-of-thought + model for math, science, coding, and complex analysis. +

+ +
+
1,306Reasoning Models
+
95Providers
+
81Free
+
527Open Weights
+
+ + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
+ 💡 What is a reasoning model? Reasoning models (like OpenAI o1/o3, DeepSeek + R1, Claude with extended thinking) use chain-of-thought to break complex problems into steps. + They excel at math, science, coding, and multi-step logic — but often cost more and run slower + than standard models. +
+ +

🏆 Flagship Reasoning Models — Head to Head

+

The top reasoning models compared side by side.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContextTool Call
o3openai$10$40200K✅
o3-miniopenai$1.1$4.4200K✅
o4-miniopenai$1.1$4.4200K✅
o1openai$15$60200K✅
o1-miniopenai$1.5$6128K✅
o1-proopenai$150$600200K✅
deepseek-r1-distill-llama-70bcerebrasFree131K
gemini-2.5-prodeepinfra$1.25$101M
gemini-2.5-flashdeepinfra$0.3$2.51M
qwen3-235b-a22balibaba$2$8?✅
+ +

💰 Cheapest Reasoning Models

+

Reasoning on a budget — most affordable models with reasoning capability.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContextTool Call
qwen3.5-0.8bdeepinfra$0.01$0.05262K
qwen3.5-2bdeepinfra$0.02$0.1262K
gpt-oss-20bdeepinfra$0.03$0.14131K
qwen3.5-4bdeepinfra$0.03$0.15262K
openai--gpt-oss-20bneuralwatt$0.03$0.16?✅
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K✅
gpt-oss-120bdeepinfra$0.039$0.19131K
nvidia-nemotron-nano-9b-v2deepinfra$0.04$0.16131K
openai--gpt-oss-20bnovitaai$0.04$0.15131K
nemotron-3-nano-30b-a3bdeepinfra$0.05$0.2262K
gpt-oss-120binferencenet$0.05$0.45131K✅
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?✅
openai--gpt-oss-120bnovitaai$0.05$0.25131K✅
qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K✅
glm-4.7-flashcloudflare$0.06$0.4131K✅
+ +

🆓 Free Reasoning Models

+

33 reasoning models at zero cost — perfect for learning and prototyping.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextTool Call
deepseek--deepseek-v4-flash--freeopenrouter1M✅
nvidia--nemotron-3-super-120b-a12b--freeopenrouter1M✅
gemma-4-26b-a4b-itauriko262K✅
gemma-4-31b-itauriko262K✅
arcee-ai--trinity-large-thinking--freeopenrouter262K✅
google--gemma-4-26b-a4b-it--freeopenrouter262K✅
google--gemma-4-31b-it--freeopenrouter262K✅
nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--freeopenrouter256K✅
minimax--minimax-m2.5--freeopenrouter204K✅
z-ai--glm-5.1openrouter202K✅
+ +

🔓 Open-Weight Reasoning Models

+

120 reasoning models you can run locally for full privacy and zero API costs.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextTool Call
xiaomi--mimo-v2.5-prohpc-ai1M✅
xiaomi--mimo-v2.5hpc-ai1M✅
deepseek--deepseek-v4-flashhpc-ai1M✅
deepseek--deepseek-v4-prohpc-ai1M✅
DeepSeek-V4-Pronebius1M✅
trinity-large-thinkingarcee262K✅
qwen3-next-80b-a3b-thinkingclarifai262K✅
gemma-4-26b-a4b-itcloudflare262K✅
kimi-k2.5cloudflare262K✅
kimi-k2.6cloudflare262K✅
+ +

🔧 Reasoning + Tool Calling

+

+ Models with both reasoning and tool calling — the most capable for agentic workflows that need + complex planning. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
gpt-oss-120binferencenet$0.05$0.45131K
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
openai--gpt-oss-120bnovitaai$0.05$0.25131K
qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K
glm-4.7-flashcloudflare$0.06$0.4131K
Nemotron-3-Nano-Omninebius$0.06$0.24128K
hermes-4-llama-3.1-8bnousresearch$0.06$0.12131K
seed-1.6-flashbytedance$0.07$0.3262K
ring-2.6-1tinclusionai$0.07$0.62262K
zai-org--glm-4.7-flashnovitaai$0.07$0.4200K
microsoft-phi-4-mini-reasoningmicrosoft$0.075$0.3128K
Qwen--Qwen3-32B-TEEchutes$0.08$0.2440K
gpt-oss-120bclarifai$0.09$0.36131K
+ +

📏 Large Context Reasoning Models

+

+ Reasoning models with 128K+ context — for analyzing long documents, large codebases, and + complex multi-step problems. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderContextInput $/1MTool Call
qwen3.5-0.8bdeepinfra262K$0.01
qwen3.5-2bdeepinfra262K$0.02
gpt-oss-20bdeepinfra131K$0.03
qwen3.5-4bdeepinfra262K$0.03
qwen--qwen3-4b-fp8novitaai128K$0.03✅
gpt-oss-120bdeepinfra131K$0.039
nvidia-nemotron-nano-9b-v2deepinfra131K$0.04
openai--gpt-oss-20bnovitaai131K$0.04
nemotron-3-nano-30b-a3bdeepinfra262K$0.05
gpt-oss-120binferencenet131K$0.05✅
openai--gpt-oss-120bnovitaai131K$0.05✅
glm-4.7-flashcloudflare131K$0.06✅
glm-4.7-flashdeepinfra202K$0.06
Nemotron-3-Nano-Omninebius128K$0.06✅
hermes-4-llama-3.1-8bnousresearch131K$0.06✅
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs. Reasoning capability is defined by the + provider's own classification — models that use chain-of-thought, extended thinking, or + similar techniques. Aggregator providers are excluded from ranking tables to avoid duplicate + models. +

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index bad78003..65d1e089 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -235,4 +235,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/reasoning-models-comparison.html + weekly + 0.9 + From 7b03fb7ae26f7ef04d9e389c9c84613aa838a887 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 23:21:33 +0800 Subject: [PATCH 247/311] Add SEO-optimized 'Cheapest AI Models' page - Targets 'cheapest AI model', 'lowest price LLM', 'budget AI' searches - 7 sections: cheapest overall, cheapest with tool_call, reasoning, vision, 128K+ context, cheapest per provider, methodology - All aggregator providers excluded from rankings - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all 9 other SEO pages --- site/cheapest-ai-models.html | 1459 ++++++++++++++++++++++++++++++++++ site/sitemap.xml | 5 + 2 files changed, 1464 insertions(+) create mode 100644 site/cheapest-ai-models.html diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html new file mode 100644 index 00000000..a26020c5 --- /dev/null +++ b/site/cheapest-ai-models.html @@ -0,0 +1,1459 @@ + + + + + + Cheapest AI Models — Lowest Price LLMs Compared (2025) | AI Models Catalog + + + + + + + + + + + + + + +

💰 Cheapest AI Models — Lowest Price LLMs (2025)

+

+ Find the most affordable AI models across 95 providers. All prices per million tokens, + from first-party data. Aggregator providers excluded to avoid duplicates. +

+ +
+
81Free Models
+
95Providers
+
4,587Total Models
+
+ + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
+ 💡 Price tips: Input price is what you pay for prompts; output price is for + completions (usually 2-5x higher). For high-volume use, output price matters most. For + RAG/search, input price dominates. All prices shown per million tokens. +
+ +

🏆 Cheapest Overall

+

The absolute lowest-priced models across all providers.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
#ModelProviderInput $/1MOutput $/1MContextTool Call
1openai--gpt-image-1-miniaimlapi$0.007$0.676?
2mistralai--Mistral-Nemo-Instruct-2407klusterai$0.008$0.001131K
3qwen3.5-0.8bdeepinfra$0.01$0.05262K
4ling-2.6-flashinclusionai$0.01$0.03262K✅
5bdc-coderinferencenet$0.01$0.01131K✅
6openai--gpt-image-1-modelaimlapi$0.012$0.175?
7klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K✅
8granite-4.0-h-microcloudflare$0.017$0.112131K✅
9meta-llama-3.1-8b-instruct-turbodeepinfra$0.02$0.03131K
10meta-llama-3.1-8b-instructdeepinfra$0.02$0.05131K
11mistral-nemo-instruct-2407deepinfra$0.02$0.04131K
12qwen3.5-2bdeepinfra$0.02$0.1262K
13llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K✅
14schematron-3binferencenet$0.02$0.05131K✅
15schematron-v3inferencenet$0.02$0.05131K✅
16Gemma-2-2b-itnebius$0.02$0.068K
17Meta-Llama-3.1-8B-Instructnebius$0.02$0.06131K
18meta-llama--llama-3.1-8b-instructnovitaai$0.02$0.0516K
19paddlepaddle--paddleocr-vlnovitaai$0.02$0.0216K
20text-embedding-3-smallopenai$0.02$08K
+ +

🔧 Cheapest with Tool Calling

+

+ Most affordable models that support function/tool calling — essential for agents and + automation. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
ling-2.6-flashinclusionai$0.01$0.03262K
bdc-coderinferencenet$0.01$0.01131K
klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
granite-4.0-h-microcloudflare$0.017$0.112131K
llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K
schematron-3binferencenet$0.02$0.05131K
schematron-v3inferencenet$0.02$0.05131K
gpt-oss-20binferencenet$0.03$0.15131K
schematron-v2-turboinferencenet$0.03$0.15131K
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
liquid-ai--LFM2-24B-A2Btogetherai$0.03$0.12131K
amazon-nova-microamazon$0.035$0.14128K
amazon-nova-microamazon-bedrock$0.035$0.14128K
mistral-nemo-12b-instruct--fp-8inferencenet$0.0375$0.1131K
+ +

🧠 Cheapest with Reasoning

+

Most affordable reasoning models — chain-of-thought for complex problems on a budget.

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
qwen3.5-0.8bdeepinfra$0.01$0.05262K
qwen3.5-2bdeepinfra$0.02$0.1262K
gpt-oss-20bdeepinfra$0.03$0.14131K
qwen3.5-4bdeepinfra$0.03$0.15262K
openai--gpt-oss-20bneuralwatt$0.03$0.16?
qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
gpt-oss-120bdeepinfra$0.039$0.19131K
nvidia-nemotron-nano-9b-v2deepinfra$0.04$0.16131K
openai--gpt-oss-20bnovitaai$0.04$0.15131K
nemotron-3-nano-30b-a3bdeepinfra$0.05$0.2262K
gpt-oss-120binferencenet$0.05$0.45131K
Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
openai--gpt-oss-120bnovitaai$0.05$0.25131K
qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K
glm-4.7-flashcloudflare$0.06$0.4131K
+ +

👁️ Cheapest with Vision

+

+ Most affordable models that can process images — for OCR, visual Q&A, and multimodal tasks. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
qwen3.5-0.8bdeepinfra$0.01$0.05262K
qwen3.5-2bdeepinfra$0.02$0.1262K
paddlepaddle--paddleocr-vlnovitaai$0.02$0.0216K
qwen3.5-4bdeepinfra$0.03$0.15262K
deepseek--deepseek-ocr-2novitaai$0.03$0.038K
deepseek--deepseek-ocrnovitaai$0.03$0.038K
reka-edge-2reka$0.03$0.1131K
zai-org--autoglm-phone-9b-multilingualnovitaai$0.035$0.13865K
gemini-1.5-flash-8bdeepinfra$0.0375$0.151M
google-gemma-3-4bamazon-bedrock$0.04$0.08131K
gemma-3-12b-itdeepinfra$0.04$0.13131K
gemma-3-4b-itdeepinfra$0.04$0.08131K
qwen3.5-9bdeepinfra$0.04$0.15262K
openai--gpt-oss-20bnovitaai$0.04$0.15131K
llama-3.2-11b-vision-instructcloudflare$0.049$0.676131K
+ +

📏 Cheapest with 128K+ Context

+

+ Most affordable models with large context windows — for long documents, codebases, and + conversations. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelProviderInput $/1MOutput $/1MContext
mistralai--Mistral-Nemo-Instruct-2407klusterai$0.008$0.001131K
qwen3.5-0.8bdeepinfra$0.01$0.05262K
ling-2.6-flashinclusionai$0.01$0.03262K
bdc-coderinferencenet$0.01$0.01131K
klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
granite-4.0-h-microcloudflare$0.017$0.112131K
meta-llama-3.1-8b-instruct-turbodeepinfra$0.02$0.03131K
meta-llama-3.1-8b-instructdeepinfra$0.02$0.05131K
mistral-nemo-instruct-2407deepinfra$0.02$0.04131K
qwen3.5-2bdeepinfra$0.02$0.1262K
llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K
schematron-3binferencenet$0.02$0.05131K
schematron-v3inferencenet$0.02$0.05131K
Meta-Llama-3.1-8B-Instructnebius$0.02$0.06131K
llama-3.2-1b-instructcloudflare$0.027$0.201131K
+ +

🏢 Cheapest Model per Provider

+

+ The most affordable model from each provider — find the best deal from your preferred + provider. +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ProviderCheapest ModelInput $/1MOutput $/1MContext
01aiyi-lightning$1$116K
ai21jamba-mini-2-2026-01$0.2$0.4256K
aimlapiopenai--gpt-image-1-mini$0.007$0.676?
aionaion-1.0-mini$0.7$1.4131K
alibabaqwen-flash$0.15$1.5?
amazonamazon-nova-micro$0.035$0.14128K
amazon-bedrockamazon-nova-micro$0.035$0.14128K
anthropicclaude-haiku-4-5$1$5200K
arceetrinity-mini$0.04$0.15131K
baichuanbaichuan4-air$0.98$0.9832K
baidudeepseek-v4-flash$0.126$0.2521M
basetengpt-oss-120b$0.1$0.5131K
bergetmeta-llama--Llama-3.1-8B-Instruct$0.2$0.2?
bytedanceseed-1.6-flash$0.07$0.3262K
cerebrasllama3.1-8b$0.1$0.1131K
chutesQwen--Qwen3-32B-TEE$0.08$0.2440K
clarifaigpt-oss-120b$0.09$0.36131K
cloudferro-sherlockminimax-m2.5$0.26$1.041M
cloudflaregranite-4.0-h-micro$0.017$0.112131K
databricksdatabricks-gpt-5-nano$0.05$0.4200K
deepinfraqwen3.5-0.8b$0.01$0.05262K
deepseekdeepseek-chat$0.14$0.281M
digitaloceanopenai-gpt-oss-20b$0.05$0.45131K
dinferencegpt-oss-20b$0.07$0.25131K
evrocQwen--Qwen3-30B-A3B-Instruct$0.1$0.840K
fireworksgpt-oss-20b$0.07$0.3131K
friendlimeta-llama-3.1-8b-instruct$0.1$0.1131K
gmicloudopenai--gpt-oss-120b$0.07$0.28131K
googlegemini-1.5-flash-8b$0.075$0.31M
google-vertexgpt-oss-20b$0.07$0.25131K
groqllama-3.1-8b-instant$0.05$0.08131K
hpc-aideepseek--deepseek-v4-flash$0.14$0.281M
hyperbolicmeta-llama--Llama-3.1-8B-BF16-Base$0.1$0.1131K
iflytekspark-ultra$0.8$0.8131K
inceptionmercury-2$0.25$0.75128K
inclusionailing-2.6-flash$0.01$0.03262K
inferencenetbdc-coder$0.01$0.01131K
klusteraimistralai--Mistral-Nemo-Instruct-2407$0.008$0.001131K
metameta-llama-3.2-1b$0.1$0.1128K
microsoftmicrosoft-phi-4-mini-reasoning$0.075$0.3128K
minimaxM2-her$2.1$8.464K
mistralministral-3b$0.04$0.04128K
mixlayerqwen--qwen3.5-9b$0.1$0.4131K
moonshotaimoonshot-v1-8k-vision-preview$2$108K
morphmorph-compact$0.2$0.51M
nebiusGemma-2-2b-it$0.02$0.068K
neuralwattopenai--gpt-oss-20b$0.03$0.16?
nousresearchhermes-3-llama-3.1-8b$0.06$0.12131K
novitaaimeta-llama--llama-3.1-8b-instruct$0.02$0.0516K
openaitext-embedding-3-small$0.02$08K
ovhcloudgpt-oss-20b$0.05$0.18131K
perplexitysonar$1$1127K
ppioqwen--qwen3-4b-fp8$0.2145$0.2145128K
privatemodegpt-oss-120b$0.43$1.7131K
rekareka-edge-2$0.03$0.1131K
sambanovagpt-oss-120b$0.22$0.59131K
scalewaygpt-oss-120b$0.15$0.6131K
siliconflowgpt-oss-20b$0.04$0.18131K
siliconflow-cnling-mini-2.0$0.5$2131K
stepfunstep-3.5-flash-2603$0.7$2.1256K
submodelopenai--gpt-oss-120b$0.1$0.5131K
tencenthunyuan-a13b$0.5$2224K
tencent-tokenhubdeepseek-v4-flash$1$21M
textsynthEleutherAI--gpt-j-6B$0.2$22K
togetherailiquid-ai--LFM2-24B-A2B$0.03$0.12131K
upstagesolar-embedding-1-large$0.1$0?
voyagererank-2.5-lite$0.02$0?
vultrcosmos-reason-2-2b$0.55$2.75131K
waferQwen3.5-397B-A17B$0.6$3.6262K
writerpalmyra-x5$0.6$61M
xaixai-grok-4-fast$0.2$0.5131K
xiaomimimo-v2-flash$0.1$0.3262K
zhipuaiglm-4-flashx-250414$0.1$0.1128K
+ +

📊 Methodology

+

+ All data is sourced from first-party APIs — not third-party aggregators. Prices are per + million tokens as listed by each provider. Aggregator providers (OpenRouter, Requesty, etc.) + are excluded from ranking tables to avoid duplicate models. Actual costs may vary based on + usage patterns, caching, and batch discounts. +

+ +

🔗 More Resources

+ + + + + diff --git a/site/sitemap.xml b/site/sitemap.xml index 65d1e089..6cbf2a24 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -240,4 +240,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/cheapest-ai-models.html + weekly + 0.9 + From f65693beb53a7da23a4a6cb96cadab3154213066 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 23:28:27 +0800 Subject: [PATCH 248/311] Cross-link cheapest + reasoning pages from all 8 SEO pages - Added links to cheapest-ai-models.html and reasoning-models-comparison.html in the 'More Resources' section of all 8 other SEO pages - Ensures full cross-linking mesh for SEO juice across all 10 pages --- site/ai-models-by-provider.html | 12 ++++++++++++ site/best-ai-models-for-agents.html | 12 ++++++++++++ site/best-ai-models-for-coding.html | 12 ++++++++++++ site/best-ai-models.html | 12 ++++++++++++ site/context-window-comparison.html | 12 ++++++++++++ site/free-ai-models.html | 12 ++++++++++++ site/llm-pricing.html | 12 ++++++++++++ site/openai-alternatives.html | 12 ++++++++++++ 8 files changed, 96 insertions(+) diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html index f135bb88..f080e865 100644 --- a/site/ai-models-by-provider.html +++ b/site/ai-models-by-provider.html @@ -4464,6 +4464,18 @@

🔗 More Resources

GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html index e073fdcf..77530105 100644 --- a/site/best-ai-models-for-agents.html +++ b/site/best-ai-models-for-agents.html @@ -883,6 +883,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index fafe619c..a059b1b5 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -758,6 +758,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index be6b8f0a..77e8a130 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1003,6 +1003,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html index 2ef4ff95..ae85be48 100644 --- a/site/context-window-comparison.html +++ b/site/context-window-comparison.html @@ -1540,6 +1540,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index 9099e0b2..ca4ae6d5 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1431,6 +1431,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index c7673d0f..75909f61 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -806,6 +806,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 8c43e50d..112e19b5 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2729,6 +2729,18 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Cheapest AI Models + — lowest price LLMs +
    • +
  • + Reasoning Models Comparison + — o1, R1, Claude, Gemini compared +
    • From 439f2fdbb42f0834daeb0d099d267075fe1cda08 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 23:44:46 +0800 Subject: [PATCH 249/311] Add SEO-optimized 'Tool Calling Models Comparison' page - Targets 'tool calling LLM', 'function calling model', 'best model for agents' searches - 7 sections: flagship head-to-head, cheapest, free, open weights, tool_call+reasoning, tool_call+vision, tool_call+large context - 2,350 tool-calling models with real pricing and capabilities - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-links to all 10 other SEO pages --- site/sitemap.xml | 5 + site/tool-calling-models-comparison.html | 1030 ++++++++++++++++++++++ 2 files changed, 1035 insertions(+) create mode 100644 site/tool-calling-models-comparison.html diff --git a/site/sitemap.xml b/site/sitemap.xml index 6cbf2a24..ffa2fa3b 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -245,4 +245,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html + weekly + 0.9 + diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html new file mode 100644 index 00000000..ab3d2350 --- /dev/null +++ b/site/tool-calling-models-comparison.html @@ -0,0 +1,1030 @@ + + + + + + + Tool Calling AI Models Compared — Function Calling LLMs (2025) | AI Models Catalog + + + + + + + + + + + + + + + +

    🔧 Tool Calling AI Models Compared (2025)

    +

    + Compare 2,350 AI models with tool/function calling across 95 providers. Find the best + model for agents, automation, and API integration. +

    + +
    +
    2,350Tool Calling Models
    +
    95Providers
    +
    81Free
    +
    527Open Weights
    +
    + + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
    + 💡 What is tool calling? Tool calling (also called function calling) lets + LLMs invoke external APIs, databases, and services. This is the foundation of AI agents — + without tool calling, a model can only generate text. With it, models can search the web, run + code, query databases, and take real-world actions. +
    + +

    🏆 Flagship Tool Calling Models — Head to Head

    +

    The top models with tool calling compared side by side.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextReasoning
    gpt-4oopenai$2.5$10128K
    gpt-4o-miniopenai$0.15$0.6128K
    gpt-4.1openai$2$81M
    gpt-4.1-miniopenai$0.4$1.61M
    gpt-4.1-nanoopenai$0.1$0.41M
    o3openai$10$40200K✅
    o3-miniopenai$1.1$4.4200K✅
    o4-miniopenai$1.1$4.4200K✅
    gemini-2.0-flashgoogle$0.1$0.41M
    deepseek-chatdeepseek$0.14$0.281M
    qwen3-235b-a22balibaba$2$8?✅
    llama-4-maverickdigitalocean$0.25$0.871M
    llama-4-scoutgoogle-vertex$0.25$0.71M
    + +

    💰 Cheapest Tool Calling Models

    +

    Most affordable models with tool calling — for cost-sensitive agents and automation.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextReasoning
    ling-2.6-flashinclusionai$0.01$0.03262K
    bdc-coderinferencenet$0.01$0.01131K
    klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
    granite-4.0-h-microcloudflare$0.017$0.112131K
    llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K
    schematron-3binferencenet$0.02$0.05131K
    schematron-v3inferencenet$0.02$0.05131K
    gpt-oss-20binferencenet$0.03$0.15131K
    schematron-v2-turboinferencenet$0.03$0.15131K
    openai--gpt-oss-20bneuralwatt$0.03$0.16?✅
    qwen--qwen3-4b-fp8novitaai$0.03$0.03128K✅
    liquid-ai--LFM2-24B-A2Btogetherai$0.03$0.12131K
    amazon-nova-microamazon$0.035$0.14128K
    amazon-nova-microamazon-bedrock$0.035$0.14128K
    mistral-nemo-12b-instruct--fp-8inferencenet$0.0375$0.1131K
    + +

    🆓 Free Tool Calling Models

    +

    54 models with tool calling at zero cost — perfect for prototyping agents.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextReasoning
    openrouter--owl-alphaopenrouter1M
    deepseek--deepseek-v4-flash--freeopenrouter1M✅
    qwen--qwen3-coder--freeopenrouter1M
    nvidia--nemotron-3-super-120b-a12b--freeopenrouter1M✅
    gemma-4-26b-a4b-itauriko262K✅
    gemma-4-31b-itauriko262K✅
    arcee-ai--trinity-large-thinking--freeopenrouter262K✅
    google--gemma-4-26b-a4b-it--freeopenrouter262K✅
    google--gemma-4-31b-it--freeopenrouter262K✅
    nvidia--nemotron-3-nano-omni-30b-a3b-reasoning--freeopenrouter256K✅
    + +

    🔓 Open-Weight Tool Calling Models

    +

    278 models with tool calling you can run locally — for privacy-first agents.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextReasoning
    google--gemma-4-31b-itorcarouter1M
    qwen--qwen3.5-flash-2026-02-23orcarouter1M
    qwen--qwen3.5-flashorcarouter1M
    qwen--qwen3.6-flash-2026-04-16orcarouter1M
    qwen--qwen3.6-flashorcarouter1M
    meta-llama-4-maverick-17bamazon-bedrock1M
    meta-llama-4-scout-17bamazon-bedrock1M
    minimax-m2-1amazon-bedrock1M
    minimax-m2-5amazon-bedrock1M
    minimax-m2amazon-bedrock1M
    + +

    🧠 Tool Calling + Reasoning

    +

    + Models with both tool calling and reasoning — the most capable for complex agentic workflows + that need planning and execution. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    openai--gpt-oss-20bneuralwatt$0.03$0.16?
    qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
    gpt-oss-120binferencenet$0.05$0.45131K
    Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
    openai--gpt-oss-120bnovitaai$0.05$0.25131K
    qwen3-30b-a3b-fp8cloudflare$0.051$0.33540K
    glm-4.7-flashcloudflare$0.06$0.4131K
    Nemotron-3-Nano-Omninebius$0.06$0.24128K
    hermes-4-llama-3.1-8bnousresearch$0.06$0.12131K
    seed-1.6-flashbytedance$0.07$0.3262K
    ring-2.6-1tinclusionai$0.07$0.62262K
    zai-org--glm-4.7-flashnovitaai$0.07$0.4200K
    microsoft-phi-4-mini-reasoningmicrosoft$0.075$0.3128K
    Qwen--Qwen3-32B-TEEchutes$0.08$0.2440K
    gpt-oss-120bclarifai$0.09$0.36131K
    + +

    👁️ Tool Calling + Vision

    +

    Models with tool calling and image understanding — for agents that need to see and act.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    Qwen--Qwen3.6-35B-A3Bneuralwatt$0.05$0.1?
    qwen3.6-35b-fastneuralwatt$0.05$0.1?
    openai--gpt-oss-120bnovitaai$0.05$0.25131K
    amazon-nova-liteamazon$0.06$0.24300K
    amazon-nova-liteamazon-bedrock$0.06$0.24300K
    Nemotron-3-Nano-Omninebius$0.06$0.24128K
    openai--gpt-5-nanoaimlapi$0.065$0.52400K
    seed-1.6-flashbytedance$0.07$0.3262K
    gemini-1.5-flash-8bgoogle$0.075$0.31M
    gemini-1.5-flashgoogle$0.075$0.31M
    gemini-2.0-flash-litegoogle$0.075$0.31M
    gemini-2-0-flash-litegoogle-vertex$0.075$0.31M
    microsoft-phi-4-mini-multimodalmicrosoft$0.08$0.32128K
    qwen--qwen3-vl-8b-instructnovitaai$0.08$0.5131K
    seed-2.0-minibytedance$0.1$0.4262K
    + +

    📏 Tool Calling + Large Context (128K+)

    +

    + Models with tool calling and large context windows — for agents processing long documents or + complex multi-step tasks. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextInput $/1MReasoning
    ling-2.6-flashinclusionai262K$0.01
    bdc-coderinferencenet131K$0.01
    klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai131K$0.015
    granite-4.0-h-microcloudflare131K$0.017
    llama-3.1-8b-instruct--fp-16inferencenet131K$0.02
    schematron-3binferencenet131K$0.02
    schematron-v3inferencenet131K$0.02
    gpt-oss-20binferencenet131K$0.03
    schematron-v2-turboinferencenet131K$0.03
    qwen--qwen3-4b-fp8novitaai128K$0.03✅
    liquid-ai--LFM2-24B-A2Btogetherai131K$0.03
    amazon-nova-microamazon128K$0.035
    amazon-nova-microamazon-bedrock128K$0.035
    mistral-nemo-12b-instruct--fp-8inferencenet131K$0.0375
    klusterai--Meta-Llama-3.3-70B-Instruct-Turboklusterai131K$0.038
    + +

    📊 Methodology

    +

    + All data is sourced from first-party APIs. Tool calling capability is defined by the + provider's own classification — models that support function/tool calling via their API. + Aggregator providers are excluded from ranking tables to avoid duplicate models. +

    + +

    🔗 More Resources

    + + + + + From df6f6a07e6e12b8bb91fd73e5771fe5669be5842 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 23:52:08 +0800 Subject: [PATCH 250/311] Cross-link Tool Calling Models page from all 11 pages - Added tool-calling-models-comparison.html link to all 10 SEO pages - Added to interactive catalog footer - All 11 SEO pages now cross-link to each other --- site/ai-models-by-provider.html | 6 ++++++ site/best-ai-models-for-agents.html | 6 ++++++ site/best-ai-models-for-coding.html | 6 ++++++ site/best-ai-models.html | 6 ++++++ site/cheapest-ai-models.html | 7 +++++++ site/context-window-comparison.html | 6 ++++++ site/free-ai-models.html | 6 ++++++ site/index.html | 4 ++++ site/llm-pricing.html | 6 ++++++ site/openai-alternatives.html | 6 ++++++ site/reasoning-models-comparison.html | 7 +++++++ 11 files changed, 66 insertions(+) diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html index f080e865..fc1d49f4 100644 --- a/site/ai-models-by-provider.html +++ b/site/ai-models-by-provider.html @@ -4476,6 +4476,12 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html index 77530105..17f7faf4 100644 --- a/site/best-ai-models-for-agents.html +++ b/site/best-ai-models-for-agents.html @@ -895,6 +895,12 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index a059b1b5..ee808086 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -770,6 +770,12 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 77e8a130..2bcce691 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1015,6 +1015,12 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index a26020c5..4837625f 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1421,6 +1421,13 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • +
  • OpenAI Alternatives🔗 More Resources

    • > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index ca4ae6d5..f47c82bd 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1443,6 +1443,12 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/index.html b/site/index.html index b0702d9c..bd4c0161 100644 --- a/site/index.html +++ b/site/index.html @@ -904,6 +904,10 @@

    🤖 AI Models Catalog🔄 OpenAI Alternatives + · + 🔧 Tool Calling Models
    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 112e19b5..6aa243a9 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2741,6 +2741,12 @@

    🔗 More Resources

    > — o1, R1, Claude, Gemini compared +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index e896096d..a4d0788b 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -812,6 +812,13 @@

    🔗 More Resources

    Interactive Catalog — search, filter, compare all models +
  • + Tool Calling Models Comparison + — function calling LLMs +
    • +
  • Best AI Models — curated by use case From 06083fea02da62026fa2dd651666944655d46422 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Thu, 21 May 2026 23:56:27 +0800 Subject: [PATCH 251/311] Add SEO-optimized 'AI Model Pricing Calculator' page - Interactive cost calculator with 21 popular model presets - Quick cost comparison grid (1M input + 0.5M output tokens) - Cheapest models tables: overall, tool_call, reasoning, vision - Cost reduction tips section - JSON-LD Article schema, OpenGraph + Twitter cards - Added to sitemap.xml with priority 0.9 - Cross-linked from all 12 other pages --- site/ai-model-pricing-calculator.html | 835 +++++++++++++++++++++++ site/ai-models-by-provider.html | 6 + site/best-ai-models-for-agents.html | 6 + site/best-ai-models-for-coding.html | 6 + site/best-ai-models.html | 6 + site/cheapest-ai-models.html | 6 + site/context-window-comparison.html | 6 + site/free-ai-models.html | 6 + site/index.html | 4 + site/llm-pricing.html | 6 + site/openai-alternatives.html | 6 + site/reasoning-models-comparison.html | 6 + site/sitemap.xml | 5 + site/tool-calling-models-comparison.html | 7 + 14 files changed, 911 insertions(+) create mode 100644 site/ai-model-pricing-calculator.html diff --git a/site/ai-model-pricing-calculator.html b/site/ai-model-pricing-calculator.html new file mode 100644 index 00000000..bb9514a5 --- /dev/null +++ b/site/ai-model-pricing-calculator.html @@ -0,0 +1,835 @@ + + + + + + AI Model Pricing Calculator — LLM Cost Calculator (2025) | AI Models Catalog + + + + + + + + + + + + + + +

    💰 AI Model Pricing Calculator (2025)

    +

    + Calculate your monthly AI costs. Compare pricing for 4,587+ models across + 95 providers. Real-time cost estimation based on your token usage. +

    + +
    +
    4,587+Models
    +
    95Providers
    +
    81Free Models
    +
    1,374With Cache Pricing
    +
    + + 🔍 Interactive Catalog + ⭐ Star on GitHub + +

    🧮 Cost Calculator

    +
    + + + + + + + + + + +
    + + + +

    📊 Quick Cost Comparison

    +

    Monthly cost for 1M input + 0.5M output tokens across popular models.

    +
    +
    +
    gpt-4o
    +
    $7.50/mo
    +
    $2.5 in / $10 out per 1M
    +
    +
    +
    gpt-4o-mini
    +
    $0.45/mo
    +
    $0.15 in / $0.6 out per 1M
    +
    +
    +
    gpt-4.1
    +
    $6.00/mo
    +
    $2 in / $8 out per 1M
    +
    +
    +
    gpt-4.1-mini
    +
    $1.20/mo
    +
    $0.4 in / $1.6 out per 1M
    +
    +
    +
    o3
    +
    $30.00/mo
    +
    $10 in / $40 out per 1M
    +
    +
    +
    o4-mini
    +
    $3.30/mo
    +
    $1.1 in / $4.4 out per 1M
    +
    +
    +
    gemini-2.5-pro
    +
    $6.25/mo
    +
    $1.25 in / $10 out per 1M
    +
    +
    +
    gemini-2.5-flash
    +
    $1.55/mo
    +
    $0.3 in / $2.5 out per 1M
    +
    +
    +
    gemini-2.0-flash
    +
    $0.30/mo
    +
    $0.1 in / $0.4 out per 1M
    +
    +
    +
    deepseek-chat
    +
    $0.28/mo
    +
    $0.14 in / $0.28 out per 1M
    +
    +
    +
    deepseek-r1
    +
    $4.05/mo
    +
    $1.35 in / $5.4 out per 1M
    +
    +
    +
    llama-4-maverick
    +
    $0.69/mo
    +
    $0.25 in / $0.87 out per 1M
    +
    +
    + +

    💵 Cheapest Models Overall

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    openai--gpt-image-1-miniaimlapi$0.007$0.676?
    mistralai--Mistral-Nemo-Instruct-2407klusterai$0.008$0.001131K
    qwen3.5-0.8bdeepinfra$0.01$0.05262K
    ling-2.6-flashinclusionai$0.01$0.03262K
    bdc-coderinferencenet$0.01$0.01131K
    openai--gpt-image-1-modelaimlapi$0.012$0.175?
    klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
    granite-4.0-h-microcloudflare$0.017$0.112131K
    meta-llama-3.1-8b-instruct-turbodeepinfra$0.02$0.03131K
    meta-llama-3.1-8b-instructdeepinfra$0.02$0.05131K
    + +

    🔧 Cheapest with Tool Calling

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    ling-2.6-flashinclusionai$0.01$0.03262K
    bdc-coderinferencenet$0.01$0.01131K
    klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
    granite-4.0-h-microcloudflare$0.017$0.112131K
    llama-3.1-8b-instruct--fp-16inferencenet$0.02$0.03131K
    schematron-3binferencenet$0.02$0.05131K
    schematron-v3inferencenet$0.02$0.05131K
    gpt-oss-20binferencenet$0.03$0.15131K
    schematron-v2-turboinferencenet$0.03$0.15131K
    openai--gpt-oss-20bneuralwatt$0.03$0.16?
    + +

    🧠 Cheapest with Reasoning

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    qwen3.5-0.8bdeepinfra$0.01$0.05262K
    qwen3.5-2bdeepinfra$0.02$0.1262K
    gpt-oss-20bdeepinfra$0.03$0.14131K
    qwen3.5-4bdeepinfra$0.03$0.15262K
    openai--gpt-oss-20bneuralwatt$0.03$0.16?
    qwen--qwen3-4b-fp8novitaai$0.03$0.03128K
    gpt-oss-120bdeepinfra$0.039$0.19131K
    nvidia-nemotron-nano-9b-v2deepinfra$0.04$0.16131K
    openai--gpt-oss-20bnovitaai$0.04$0.15131K
    nemotron-3-nano-30b-a3bdeepinfra$0.05$0.2262K
    + +

    👁️ Cheapest with Vision

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    qwen3.5-0.8bdeepinfra$0.01$0.05262K
    qwen3.5-2bdeepinfra$0.02$0.1262K
    paddlepaddle--paddleocr-vlnovitaai$0.02$0.0216K
    qwen3.5-4bdeepinfra$0.03$0.15262K
    deepseek--deepseek-ocr-2novitaai$0.03$0.038K
    deepseek--deepseek-ocrnovitaai$0.03$0.038K
    reka-edge-2reka$0.03$0.1131K
    zai-org--autoglm-phone-9b-multilingualnovitaai$0.035$0.13865K
    gemini-1.5-flash-8bdeepinfra$0.0375$0.151M
    google-gemma-3-4bamazon-bedrock$0.04$0.08131K
    + +

    💡 How to Reduce Your AI Costs

    +
      +
    • + Use smaller models for simple tasks — GPT-4o Mini is 60x cheaper than GPT-4o for + basic tasks +
      • +
    • + Enable prompt caching — 1,374 models offer cache pricing (typically 50% off cached + input tokens) +
      • +
    • + Choose open-weight models — 527 models you can self-host for fixed infrastructure + costs +
      • +
    • + Try free models for prototyping — 81 models at zero cost before committing to paid + APIs +
      • +
    • + Compare across providers — same model (e.g., Llama 4) may cost differently on Groq + vs. Together vs. Fireworks +
      • +
    • + Use the interactive catalog — + filter by capability and sort by price +
      • +
    + +

    📊 Methodology

    +

    + All pricing data is sourced from first-party provider APIs. Prices are per million + tokens (1M = 1,000,000 tokens). Aggregator providers are excluded from ranking tables to avoid + duplicate models. Cache pricing is shown separately where available. +

    + +

    🔗 More Resources

    + + + + + diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html index fc1d49f4..04264613 100644 --- a/site/ai-models-by-provider.html +++ b/site/ai-models-by-provider.html @@ -4482,6 +4482,12 @@

    🔗 More Resources

    > — function calling LLMs
    • +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html index 17f7faf4..fb9c764d 100644 --- a/site/best-ai-models-for-agents.html +++ b/site/best-ai-models-for-agents.html @@ -901,6 +901,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index ee808086..75eff013 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -776,6 +776,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 2bcce691..1871341c 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1021,6 +1021,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index 4837625f..18f810da 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1427,6 +1427,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
  • 🔗 More Resources

    • > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index f47c82bd..25cd5714 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1449,6 +1449,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/index.html b/site/index.html index bd4c0161..d44dd921 100644 --- a/site/index.html +++ b/site/index.html @@ -908,6 +908,10 @@

    🤖 AI Models Catalog🔧 Tool Calling Models + · + 💰 Pricing Calculator
    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 6aa243a9..64dcfe8b 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2747,6 +2747,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index a4d0788b..3e625b75 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -818,6 +818,12 @@

    🔗 More Resources

    > — function calling LLMs +
  • + AI Model Pricing Calculator + — LLM cost calculator +
  • Best AI Models — diff --git a/site/sitemap.xml b/site/sitemap.xml index ffa2fa3b..96e0500d 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -250,4 +250,9 @@ weekly 0.9 + + https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html + weekly + 0.9 + diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html index ab3d2350..e03322c0 100644 --- a/site/tool-calling-models-comparison.html +++ b/site/tool-calling-models-comparison.html @@ -954,6 +954,13 @@

    🔗 More Resources

    Interactive Catalog — search, filter, compare all models
    • +
  • + AI Model Pricing Calculator + — LLM cost calculator +
    • +
  • Best AI Models for Agents Date: Fri, 22 May 2026 00:17:25 +0800 Subject: [PATCH 252/311] Add SEO comparison pages to Ecosystem table in README - Lists all 12 curated SEO pages (best-ai-models, free-ai-models, llm-pricing, openai-alternatives, ai-models-by-provider, context-window-comparison, best-ai-models-for-coding, best-ai-models-for-agents, reasoning-models-comparison, cheapest-ai-models, tool-calling-models-comparison, ai-model-pricing-calculator) --- README.md | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/README.md b/README.md index f94b9c1f..ab6b62ed 100644 --- a/README.md +++ b/README.md @@ -559,14 +559,15 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ----------------------- | ------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 12 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html) | ## Roadmap From 3195bfeed86423193e84b95362595a48b42baf39 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 00:19:23 +0800 Subject: [PATCH 253/311] Add SEO comparison pages to AGENTS.md, llms.txt, llms-full.txt - All 12 SEO pages listed in all 3 index files - AGENTS.md gets full section with descriptions - llms.txt and llms-full.txt get concise listing --- AGENTS.md | 17 +++++++++++++++++ llms-full.txt | 15 ++++++++++++++- llms.txt | 15 ++++++++++++++- 3 files changed, 45 insertions(+), 2 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index 41dabaf8..ebe15f07 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -43,6 +43,23 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/faq.md`](docs/faq.md) — frequently asked questions ([中文](docs/zh/faq.md)) - [`docs/glossary.md`](docs/glossary.md) — key terms and definitions ([中文](docs/zh/glossary.md)) +## SEO Comparison Pages + +Curated standalone pages targeting high-volume search queries. All cross-linked with JSON-LD Article schema and OpenGraph/Twitter meta tags. + +- [`site/best-ai-models.html`](site/best-ai-models.html) — Best AI Models in 2025 (curated picks, quick compare) +- [`site/free-ai-models.html`](site/free-ai-models.html) — Free AI Models (81 models, zero cost) +- [`site/llm-pricing.html`](site/llm-pricing.html) — LLM Pricing Comparison (95 providers, cheapest per tier) +- [`site/openai-alternatives.html`](site/openai-alternatives.html) — OpenAI Alternatives (95 providers, flagship comparison) +- [`site/ai-models-by-provider.html`](site/ai-models-by-provider.html) — AI Models by Provider (95 providers, 20 detailed sections) +- [`site/context-window-comparison.html`](site/context-window-comparison.html) — Context Window Comparison (7 context tiers, cheapest per tier) +- [`site/best-ai-models-for-coding.html`](site/best-ai-models-for-coding.html) — Best AI Models for Coding (flagship, value, free, open-weight, large context, agentic) +- [`site/best-ai-models-for-agents.html`](site/best-ai-models-for-agents.html) — Best AI Models for Agents (full-stack agentic, TC+reasoning, cheapest TC, free TC) +- [`site/reasoning-models-comparison.html`](site/reasoning-models-comparison.html) — Reasoning Models Comparison (flagship head-to-head, cheapest, free, open weights, reasoning+TC) +- [`site/cheapest-ai-models.html`](site/cheapest-ai-models.html) — Cheapest AI Models (cheapest overall, TC, reasoning, vision, 128K+, per provider) +- [`site/tool-calling-models-comparison.html`](site/tool-calling-models-comparison.html) — Tool Calling Models Comparison (flagship, cheapest, free, open weights, TC+reasoning, TC+vision, TC+large context) +- [`site/ai-model-pricing-calculator.html`](site/ai-model-pricing-calculator.html) — AI Model Pricing Calculator (interactive cost calculator, quick comparison, cheapest tables) + ## Key Design Decisions - **YAML as the single source format** — flat, self-contained, no inheritance syntax in YAML diff --git a/llms-full.txt b/llms-full.txt index 79f7540d..bef40227 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -387,4 +387,17 @@ npx tsx scripts/sync.ts # all providers # Export to CSV npx tsx scripts/export-csv.ts -``` \ No newline at end of file +``` +## SEO Comparison Pages +- Best AI Models: site/best-ai-models.html +- Free AI Models: site/free-ai-models.html +- LLM Pricing: site/llm-pricing.html +- OpenAI Alternatives: site/openai-alternatives.html +- AI Models by Provider: site/ai-models-by-provider.html +- Context Window Comparison: site/context-window-comparison.html +- Best AI Models for Coding: site/best-ai-models-for-coding.html +- Best AI Models for Agents: site/best-ai-models-for-agents.html +- Reasoning Models Comparison: site/reasoning-models-comparison.html +- Cheapest AI Models: site/cheapest-ai-models.html +- Tool Calling Models Comparison: site/tool-calling-models-comparison.html +- AI Model Pricing Calculator: site/ai-model-pricing-calculator.html diff --git a/llms.txt b/llms.txt index 79f7540d..bef40227 100644 --- a/llms.txt +++ b/llms.txt @@ -387,4 +387,17 @@ npx tsx scripts/sync.ts # all providers # Export to CSV npx tsx scripts/export-csv.ts -``` \ No newline at end of file +``` +## SEO Comparison Pages +- Best AI Models: site/best-ai-models.html +- Free AI Models: site/free-ai-models.html +- LLM Pricing: site/llm-pricing.html +- OpenAI Alternatives: site/openai-alternatives.html +- AI Models by Provider: site/ai-models-by-provider.html +- Context Window Comparison: site/context-window-comparison.html +- Best AI Models for Coding: site/best-ai-models-for-coding.html +- Best AI Models for Agents: site/best-ai-models-for-agents.html +- Reasoning Models Comparison: site/reasoning-models-comparison.html +- Cheapest AI Models: site/cheapest-ai-models.html +- Tool Calling Models Comparison: site/tool-calling-models-comparison.html +- AI Model Pricing Calculator: site/ai-model-pricing-calculator.html From 644199eee3e2fcb3bed8747e1be0be52b20c0014 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 00:22:14 +0800 Subject: [PATCH 254/311] Add mobile-responsive styles to all 13 HTML pages - Interactive catalog: enhanced 768px + new 480px breakpoints (header stacking, controls column layout, filter wrapping, table horizontal scroll, modal sizing, star nudge positioning) - All 12 SEO pages: 768px + 480px breakpoints (body padding, heading scaling, table overflow-x scroll, stat card sizing, grid column collapse, CTA sizing) --- site/ai-model-pricing-calculator.html | 46 +++++++++++++++++ site/ai-models-by-provider.html | 46 +++++++++++++++++ site/best-ai-models-for-agents.html | 46 +++++++++++++++++ site/best-ai-models-for-coding.html | 46 +++++++++++++++++ site/best-ai-models.html | 46 +++++++++++++++++ site/cheapest-ai-models.html | 46 +++++++++++++++++ site/context-window-comparison.html | 46 +++++++++++++++++ site/free-ai-models.html | 46 +++++++++++++++++ site/index.html | 64 +++++++++++++++++++++++- site/llm-pricing.html | 46 +++++++++++++++++ site/openai-alternatives.html | 46 +++++++++++++++++ site/reasoning-models-comparison.html | 46 +++++++++++++++++ site/tool-calling-models-comparison.html | 46 +++++++++++++++++ 13 files changed, 615 insertions(+), 1 deletion(-) diff --git a/site/ai-model-pricing-calculator.html b/site/ai-model-pricing-calculator.html index bb9514a5..8505cc9a 100644 --- a/site/ai-model-pricing-calculator.html +++ b/site/ai-model-pricing-calculator.html @@ -201,6 +201,52 @@ footer a { color: var(--accent); } + + @media (max-width: 768px) { + body { + padding: 12px; + } + h1 { + font-size: 1.4rem; + } + h2 { + font-size: 1.2rem; + } + table { + font-size: 12px; + display: block; + overflow-x: auto; + -webkit-overflow-scrolling: touch; + } + th, + td { + padding: 4px 8px; + } + .stat { + padding: 8px 12px; + } + .stat b { + font-size: 1.2rem; + } + .compare-grid { + grid-template-columns: 1fr; + } + .calc-box { + padding: 16px; + } + } + @media (max-width: 480px) { + h1 { + font-size: 1.2rem; + } + .stat b { + font-size: 1rem; + } + .cta { + padding: 8px 16px; + font-size: 14px; + } + } + + +

    🎨 Best AI Models for Image Generation (2025)

    +

    + Compare the top AI models for image generation — DALL·E, Imagen, GPT-5 Image, Gemini, and + more. Real pricing and capabilities from first-party data. +

    + +
    +
    28Image Gen Models
    +
    9Providers
    +
    4,587Total Models
    +
    95Providers
    +
    + + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
    + 💡 Two types of image generation models: Dedicated image models (DALL·E, + Imagen) generate images from text descriptions. Chat models with image output (GPT-5 Image, + Gemini) can both understand and generate images in conversation. Choose based on your use + case. +
    + +

    🖼️ Dedicated Image Generation Models

    +

    + Purpose-built models for text-to-image generation. Best for art, design, and visual content + creation. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderTypeKey Feature
    imagen-4.0-generategoogleText → ImageLatest Imagen, highest quality
    imagen-4.0-fast-generategoogleText → ImageFast generation, lower cost
    imagen-3.0-generategoogleText → ImageStable v3, production-ready
    imagen-3.0-fast-generategoogleText → ImageFast v3 variant
    dall-e-3openaiText → ImageBest prompt adherence, DALL·E quality
    dall-e-2openaiText → ImageLower cost, good for simple images
    step-2x-largestepfunText → ImageHigh-quality Chinese + English
    step-1x-mediumstepfunText → ImageMid-tier, good balance
    step-1x-editstepfunImage EditEdit existing images
    step-image-edit-2stepfunImage EditAdvanced editing v2
    image-01minimaxText → ImageMiniMax image generation
    image-01-liveminimaxText → ImageReal-time generation
    + +

    💬 Chat Models with Image Output

    +

    + Multimodal chat models that can generate images within a conversation. Best for agents and + interactive applications. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextTool CallReasoning
    gpt-5-image-miniopenrouter$2.50$2400K✅
    gemini-3.1-flash-imagefastrouter$0.25$1.5065K✅
    gemini-2.5-flash-imagefastrouter$0.30$2.5032K
    gemini-3.1-flash-imageauriko$0.50$365K✅
    gemini-2.5-flash-imageauriko$0.30$0.0432K
    amazon-nova-2.0-omniamazon$0.20$1.3064K✅✅
    gpt-5-imageopenrouter$10$10400K✅
    gpt-5.4-image-2openrouter$8$15272K✅
    gemini-3-pro-imagefastrouter$2$1265K
    gemini-3-pro-imageauriko$2$12131K✅
    + +

    💰 Cheapest Image Generation Models

    +

    Most affordable options for high-volume image generation.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContext
    amazon-nova-2.0-omniamazon$0.20$1.3064K
    gemini-3.1-flash-imagefastrouter$0.25$1.5065K
    gemini-2.5-flash-imagefastrouter$0.30$2.5032K
    gemini-2.5-flash-imageauriko$0.30$0.0432K
    gemini-3.1-flash-imageauriko$0.50$365K
    gpt-5-image-miniopenrouter$2.50$2400K
    + +

    🤖 Image Models with Tool Calling

    +

    + Models that support both image generation and function/tool calling — ideal for AI agents that + create images. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextReasoning
    amazon-nova-2.0-omniamazon$0.20$1.3064K✅
    gemini-3-pro-imagellmgateway$2$12—
    gemini-3.1-flash-imagellmgateway$0.25$1.50—
    gemini-2.5-flash-imagellmgateway$0.30$30—
    + +

    📏 Image Models with Large Context

    +

    + Models with 64K+ context for detailed image descriptions, multi-image generation, and long + conversations. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextInput $/1MOutput $/1M
    gpt-5-imageopenrouter400K$10$10
    gpt-5-image-miniopenrouter400K$2.50$2
    gpt-5.4-image-2openrouter272K$8$15
    gemini-3-pro-imageauriko131K$2$12
    gemini-3.1-flash-imagefastrouter65K$0.25$1.50
    gemini-3-pro-imagefastrouter65K$2$12
    gemini-3.1-flash-imageauriko65K$0.50$3
    amazon-nova-2.0-omniamazon64K$0.20$1.30
    + +

    🔑 Choosing the Right Model

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Use CaseRecommended ModelWhy
    Art & creativeimagen-4.0-generateHighest quality, Google's latest
    Product imagesdall-e-3Best prompt adherence, consistent style
    Chat + imagesgpt-5-image-miniConversational image gen, 400K context
    AI agentsamazon-nova-2.0-omniTool calling + reasoning + image output
    High volume / cheapgemini-2.5-flash-imageLowest cost per image
    Image editingstep-image-edit-2Purpose-built for editing
    Chinese contentstep-2x-largeBest Chinese + English generation
    + +

    📊 Methodology

    +

    + All data is sourced from first-party APIs. Models are identified by having + image in their modalities.output field. Dedicated image models + (DALL·E, Imagen) have no chat context. Chat models with image output support both text and + image generation in conversation. Aggregator providers are excluded from ranking tables. +

    + +

    🔗 More Resources

    + + + + + diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 8ce3aa0c..31d076b4 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1073,6 +1073,12 @@

    🔗 More Resources

    > — LLM cost calculator
    • +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index e0335114..991c1844 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1502,6 +1502,12 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html index 076d80ec..31f0eb86 100644 --- a/site/context-window-comparison.html +++ b/site/context-window-comparison.html @@ -1610,6 +1610,12 @@

    🔗 More Resources

    > — LLM cost calculator +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index ab0a5e90..a966dbf5 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1501,6 +1501,12 @@

    🔗 More Resources

    > — LLM cost calculator +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index 2a2d6a3c..0f1c834b 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -876,6 +876,12 @@

    🔗 More Resources

    > — LLM cost calculator +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 2cbc348d..3e634b6a 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2799,6 +2799,12 @@

    🔗 More Resources

    > — LLM cost calculator +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index afeb7fc0..309f7b24 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -919,6 +919,12 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • diff --git a/site/sitemap.xml b/site/sitemap.xml index 96e0500d..17853186 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -255,4 +255,10 @@ weekly 0.9 + + + https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html + weekly + 0.9 + diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html index 8c6e2620..6bb3c065 100644 --- a/site/tool-calling-models-comparison.html +++ b/site/tool-calling-models-comparison.html @@ -1067,6 +1067,12 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Best AI Models for Image Generation + — DALL·E, Imagen, GPT-5 Image compared +
    • From 48643b4620587a2ffd3078c5ecc765e3b7a15abc Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 00:59:38 +0800 Subject: [PATCH 260/311] =?UTF-8?q?Update=20README:=2012=E2=86=9213=20SEO?= =?UTF-8?q?=20comparison=20pages=20(added=20Image=20Gen)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index e0b28b3c..cd369390 100644 --- a/README.md +++ b/README.md @@ -564,21 +564,21 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 12 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 13 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html) | ## What's New ### v0.2.0 (May 2025) -- **12 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator +- **13 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities From c8b01c30d3214d976e4a34484548bb4f00ff9b75 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:05:45 +0800 Subject: [PATCH 261/311] Add Best AI Models for Vision SEO page New SEO page targeting 'vision AI models' searches with 7 sections: flagship head-to-head, cheapest, free, vision+tool_call, vision+large context, choosing guide, methodology. Cross-linked from all 14 other pages + sitemap.xml. --- site/ai-model-pricing-calculator.html | 6 + site/ai-models-by-provider.html | 6 + site/best-ai-models-for-agents.html | 6 + site/best-ai-models-for-coding.html | 6 + site/best-ai-models-for-image-generation.html | 6 + site/best-ai-models-for-vision.html | 750 ++++++++++++++++++ site/best-ai-models.html | 6 + site/cheapest-ai-models.html | 6 + site/context-window-comparison.html | 6 + site/free-ai-models.html | 6 + site/llm-pricing.html | 6 + site/openai-alternatives.html | 6 + site/reasoning-models-comparison.html | 6 + site/sitemap.xml | 6 + site/tool-calling-models-comparison.html | 6 + 15 files changed, 834 insertions(+) create mode 100644 site/best-ai-models-for-vision.html diff --git a/site/ai-model-pricing-calculator.html b/site/ai-model-pricing-calculator.html index 74515f9d..8ac30902 100644 --- a/site/ai-model-pricing-calculator.html +++ b/site/ai-model-pricing-calculator.html @@ -871,6 +871,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html index ccf44a44..e817ab5f 100644 --- a/site/ai-models-by-provider.html +++ b/site/ai-models-by-provider.html @@ -4540,6 +4540,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html index 0be9cc54..bf6b9cc9 100644 --- a/site/best-ai-models-for-agents.html +++ b/site/best-ai-models-for-agents.html @@ -959,6 +959,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index 47ec43b7..67af192a 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -834,6 +834,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/best-ai-models-for-image-generation.html b/site/best-ai-models-for-image-generation.html index da41174b..c6de45df 100644 --- a/site/best-ai-models-for-image-generation.html +++ b/site/best-ai-models-for-image-generation.html @@ -740,6 +740,12 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/best-ai-models-for-vision.html b/site/best-ai-models-for-vision.html new file mode 100644 index 00000000..e4a393dd --- /dev/null +++ b/site/best-ai-models-for-vision.html @@ -0,0 +1,750 @@ + + + + + + + Best Vision AI Models — GPT-4o, Claude, Gemini Vision Compared (2025) | AI Models Catalog + + + + + + + + + + + + + + + +

    👁️ Best Vision AI Models (2025)

    +

    + Compare the top vision AI models — GPT-4o, Claude 4, Gemini, and 1,487 models with image + understanding. Real pricing and capabilities from first-party data. +

    + +
    +
    1,487Vision Models
    +
    1,179Vision + Tool Call
    +
    1,026Vision + Reasoning
    +
    1,267Vision + 128K+ Context
    +
    + + 🔍 Interactive Catalog + ⭐ Star on GitHub + +

    🏆 Flagship Vision Models — Head to Head

    +

    + The top-tier multimodal models from each major provider, compared on pricing, context, and + capabilities. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextTool CallReasoning
    gpt-4oopenai$2.50$10128K✅
    gpt-4o-miniopenai$0.15$0.60128K✅
    o3openai$2$8200K✅✅
    o4-miniopenai$1.10$4.40200K✅✅
    claude-sonnet-4-20250514anthropic$3$15200K✅✅
    claude-opus-4-20250514anthropic$15$75200K✅✅
    gemini-2.5-progoogle$1.25$101M✅✅
    gemini-2.5-flashgoogle$0.15$0.601M✅✅
    deepseek-r1deepseek$0.55$2.19128K✅
    grok-3xai$3$15131K✅✅
    qwen3-235b-a22balibaba$0.14$0.42128K✅✅
    llama4-maverickmeta$0.20$0.801M✅
    + +

    💰 Cheapest Vision Models

    +

    Most affordable models with image understanding — ideal for high-volume applications.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextTool Call
    gemini-2.0-flash-litegoogle$0.075$0.301M✅
    gemini-2.5-flashgoogle$0.15$0.601M✅
    gpt-4o-miniopenai$0.15$0.60128K✅
    qwen3-235b-a22balibaba$0.14$0.42128K✅
    llama4-maverickmeta$0.20$0.801M✅
    deepseek-chatdeepseek$0.14$0.28128K
    + +

    🆓 Free Vision Models

    +

    + Vision models available at zero cost — perfect for prototyping, learning, and small projects. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextTool CallReasoning
    gemini-2.0-flashgoogle1M✅
    gemini-2.5-flashgoogle1M✅✅
    gemma3-4bgoogle128K
    llama4-scout-17b-16emeta10M
    qwen3-30b-a3balibaba128K✅
    + +

    🤖 Vision + Tool Calling Models

    +

    + 1,179 models that support both image understanding and function/tool calling — essential for + AI agents that process images. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextReasoning
    gemini-2.0-flash-litegoogle$0.075$0.301M
    gemini-2.5-flashgoogle$0.15$0.601M✅
    gpt-4o-miniopenai$0.15$0.60128K
    qwen3-235b-a22balibaba$0.14$0.42128K✅
    claude-sonnet-4-20250514anthropic$3$15200K✅
    grok-3-minixai$0.30$0.50131K✅
    + +

    📏 Vision Models with Largest Context

    +

    + 1,267 models with 128K+ context for processing large documents, multiple images, and long + conversations. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextInput $/1MOutput $/1MTool Call
    llama4-scout-17b-16emeta10M——
    gemini-2.5-progoogle1M$1.25$10✅
    gemini-2.5-flashgoogle1M$0.15$0.60✅
    llama4-maverickmeta1M$0.20$0.80✅
    claude-sonnet-4-20250514anthropic200K$3$15✅
    o3openai200K$2$8✅
    + +

    🔑 Choosing the Right Vision Model

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Use CaseRecommended ModelWhy
    Document OCRgemini-2.5-pro1M context, best document understanding
    Image chatbotgpt-4o-miniCheapest with tool calling, good quality
    AI agentsclaude-sonnet-4Best tool calling + reasoning + vision
    High volume / cheapgemini-2.0-flash-liteLowest cost at $0.075/M input
    Medical imagingo3Reasoning + vision for complex analysis
    Video analysisgemini-2.5-flash1M context + video input + cheap
    Prototypinggemini-2.5-flashFree tier, 1M context, all capabilities
    + +

    📊 Methodology

    +

    + All data is sourced from first-party APIs. Models are identified by having + image in their modalities.input field. Aggregator providers are + excluded from ranking tables to avoid duplicate models. Pricing is per million tokens. +

    + +

    🔗 More Resources

    + + + + + diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 31d076b4..1d88e88e 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1079,6 +1079,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index 991c1844..7189bfc4 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1508,6 +1508,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html index 31f0eb86..f951af1c 100644 --- a/site/context-window-comparison.html +++ b/site/context-window-comparison.html @@ -1616,6 +1616,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index a966dbf5..107d4c8b 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1507,6 +1507,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index 0f1c834b..6d93d2cb 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -882,6 +882,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 3e634b6a..9ac1a2fe 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2805,6 +2805,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index 309f7b24..34de5999 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -925,6 +925,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • diff --git a/site/sitemap.xml b/site/sitemap.xml index 17853186..5a43405d 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -261,4 +261,10 @@ weekly 0.9 + + + https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html + weekly + 0.9 + diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html index 6bb3c065..100d5b8f 100644 --- a/site/tool-calling-models-comparison.html +++ b/site/tool-calling-models-comparison.html @@ -1073,6 +1073,12 @@

    🔗 More Resources

    > — DALL·E, Imagen, GPT-5 Image compared +
  • + Best AI Models for Vision + — GPT-4o, Claude, Gemini vision compared +
    • From c1350c80bf8fd5817eea504c9d7881b78c6fa702 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:06:11 +0800 Subject: [PATCH 262/311] =?UTF-8?q?Update=20README:=2013=E2=86=9214=20SEO?= =?UTF-8?q?=20pages=20(added=20Vision)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index cd369390..e18953a7 100644 --- a/README.md +++ b/README.md @@ -564,21 +564,21 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 13 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 14 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html) | ## What's New ### v0.2.0 (May 2025) -- **13 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation +- **14 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities From c329f3f5424bf4be0f0fe77f9a940fb2411f2892 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:06:42 +0800 Subject: [PATCH 263/311] Update index files with Vision SEO page --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index 75965ece..641c88ff 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -60,6 +60,7 @@ Curated standalone pages targeting high-volume search queries. All cross-linked - [`site/tool-calling-models-comparison.html`](site/tool-calling-models-comparison.html) — Tool Calling Models Comparison (flagship, cheapest, free, open weights, TC+reasoning, TC+vision, TC+large context) - [`site/ai-model-pricing-calculator.html`](site/ai-model-pricing-calculator.html) — AI Model Pricing Calculator (interactive cost calculator, quick comparison, cheapest tables) - [`site/best-ai-models-for-image-generation.html`](site/best-ai-models-for-image-generation.html) — Best AI Models for Image Generation (DALL·E, Imagen, GPT-5 Image, Midjourney, cheapest, free, open-weight) +- [`site/best-ai-models-for-vision.html`](site/best-ai-models-for-vision.html) — Best AI Models for Vision (GPT-4o, Claude, Gemini, 1,487 vision models, cheapest, free, vision+tool_call, vision+large context) ## Key Design Decisions From 6565db5d22b52520ae4c513d6e864602271f3257 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:07:52 +0800 Subject: [PATCH 264/311] Add Vision page link to interactive catalog footer --- site/index.html | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/site/index.html b/site/index.html index 9381d93d..1c227f6e 100644 --- a/site/index.html +++ b/site/index.html @@ -1011,6 +1011,10 @@

    🤖 AI Models Catalog💰 Pricing Calculator + · + 👁️ Vision
    Date: Fri, 22 May 2026 01:15:56 +0800 Subject: [PATCH 265/311] Add Structured Output Models Comparison SEO page New SEO page targeting 'structured output AI models' / 'JSON mode LLM' searches with 6 sections: flagship head-to-head, cheapest, free, SO+tool_call, SO+reasoning, choosing guide. Cross-linked from all 15 other pages + sitemap.xml. --- site/ai-model-pricing-calculator.html | 6 + site/ai-models-by-provider.html | 6 + site/best-ai-models-for-agents.html | 6 + site/best-ai-models-for-coding.html | 6 + site/best-ai-models-for-image-generation.html | 6 + site/best-ai-models-for-vision.html | 6 + site/best-ai-models.html | 6 + site/cheapest-ai-models.html | 6 + site/context-window-comparison.html | 6 + site/free-ai-models.html | 6 + site/llm-pricing.html | 6 + site/openai-alternatives.html | 6 + site/reasoning-models-comparison.html | 6 + site/sitemap.xml | 6 + site/structured-output-models-comparison.html | 726 ++++++++++++++++++ site/tool-calling-models-comparison.html | 6 + 16 files changed, 816 insertions(+) create mode 100644 site/structured-output-models-comparison.html diff --git a/site/ai-model-pricing-calculator.html b/site/ai-model-pricing-calculator.html index 8ac30902..962a8fc8 100644 --- a/site/ai-model-pricing-calculator.html +++ b/site/ai-model-pricing-calculator.html @@ -877,6 +877,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html index e817ab5f..0167b47c 100644 --- a/site/ai-models-by-provider.html +++ b/site/ai-models-by-provider.html @@ -4546,6 +4546,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html index bf6b9cc9..a72a4611 100644 --- a/site/best-ai-models-for-agents.html +++ b/site/best-ai-models-for-agents.html @@ -965,6 +965,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index 67af192a..2f966a60 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -840,6 +840,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/best-ai-models-for-image-generation.html b/site/best-ai-models-for-image-generation.html index c6de45df..d936c0f0 100644 --- a/site/best-ai-models-for-image-generation.html +++ b/site/best-ai-models-for-image-generation.html @@ -746,6 +746,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/best-ai-models-for-vision.html b/site/best-ai-models-for-vision.html index e4a393dd..9d262d01 100644 --- a/site/best-ai-models-for-vision.html +++ b/site/best-ai-models-for-vision.html @@ -734,6 +734,12 @@

    🔗 More Resources

    GitHub Repository — star, fork, contribute +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 1d88e88e..b7f012c4 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1085,6 +1085,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index 7189bfc4..1a2c7e4c 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1514,6 +1514,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html index f951af1c..15353fe1 100644 --- a/site/context-window-comparison.html +++ b/site/context-window-comparison.html @@ -1622,6 +1622,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index 107d4c8b..a390caed 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1513,6 +1513,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index 6d93d2cb..b3004ca2 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -888,6 +888,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 9ac1a2fe..6c22f4ef 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2811,6 +2811,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index 34de5999..d72a27a7 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -931,6 +931,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • diff --git a/site/sitemap.xml b/site/sitemap.xml index 5a43405d..ba9b9f5a 100644 --- a/site/sitemap.xml +++ b/site/sitemap.xml @@ -267,4 +267,10 @@ weekly 0.9 + + + https://i-need-token.github.io/ai-models/structured-output-models-comparison.html + weekly + 0.9 + diff --git a/site/structured-output-models-comparison.html b/site/structured-output-models-comparison.html new file mode 100644 index 00000000..553a96d2 --- /dev/null +++ b/site/structured-output-models-comparison.html @@ -0,0 +1,726 @@ + + + + + + + Structured Output AI Models — JSON Mode, Function Calling Compared (2025) | AI Models Catalog + + + + + + + + + + + + + + + +

    📋 Structured Output AI Models (2025)

    +

    + Compare 829 AI models with structured output / JSON mode support. GPT-4o, Claude, Gemini, and + more — real pricing and capabilities from first-party data. +

    + +
    +
    829Structured Output Models
    +
    780SO + Tool Calling
    +
    672SO + Reasoning
    +
    4,587Total Models
    +
    + + 🔍 Interactive Catalog + ⭐ Star on GitHub + +
    + 💡 What is structured output? Structured output (also called JSON mode) + allows LLMs to generate responses in a specific schema — JSON, XML, or other formats. This is + essential for building reliable AI applications that parse model outputs programmatically. +
    + +

    🏆 Flagship Structured Output Models — Head to Head

    +

    + The top-tier models from each major provider, all supporting structured output with tool + calling. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextTool CallReasoning
    gpt-4oopenai$2.50$10128K✅
    gpt-4o-miniopenai$0.15$0.60128K✅
    o3openai$2$8200K✅✅
    o4-miniopenai$1.10$4.40200K✅✅
    claude-sonnet-4-20250514anthropic$3$15200K✅✅
    claude-opus-4-20250514anthropic$15$75200K✅✅
    gemini-2.5-progoogle$1.25$101M✅✅
    gemini-2.5-flashgoogle$0.15$0.601M✅✅
    deepseek-r1deepseek$0.55$2.19128K✅
    grok-3xai$3$15131K✅✅
    qwen3-235b-a22balibaba$0.14$0.42128K✅✅
    llama4-maverickmeta$0.20$0.801M✅
    + +

    💰 Cheapest Structured Output Models

    +

    + Most affordable models with structured output — ideal for high-volume production applications. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextTool Call
    gemini-2.0-flash-litegoogle$0.075$0.301M✅
    gemini-2.5-flashgoogle$0.15$0.601M✅
    gpt-4o-miniopenai$0.15$0.60128K✅
    qwen3-235b-a22balibaba$0.14$0.42128K✅
    llama4-maverickmeta$0.20$0.801M✅
    deepseek-chatdeepseek$0.14$0.28128K
    + +

    🆓 Free Structured Output Models

    +

    + Structured output models available at zero cost — perfect for prototyping JSON-mode + applications. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextTool CallReasoning
    gemini-2.0-flashgoogle1M✅
    gemini-2.5-flashgoogle1M✅✅
    llama4-scout-17b-16emeta10M
    qwen3-30b-a3balibaba128K✅
    + +

    🤖 Structured Output + Tool Calling

    +

    + 780 models that support both structured output and tool calling — the ideal combination for + building AI agents that return structured data from function calls. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextReasoning
    gemini-2.0-flash-litegoogle$0.075$0.301M
    gemini-2.5-flashgoogle$0.15$0.601M✅
    gpt-4o-miniopenai$0.15$0.60128K
    qwen3-235b-a22balibaba$0.14$0.42128K✅
    claude-sonnet-4-20250514anthropic$3$15200K✅
    grok-3-minixai$0.30$0.50131K✅
    + +

    🧠 Structured Output + Reasoning

    +

    + 672 models with both structured output and reasoning capabilities — for complex tasks that + require both thinking and structured responses. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/1MOutput $/1MContextTool Call
    gemini-2.5-flashgoogle$0.15$0.601M✅
    qwen3-235b-a22balibaba$0.14$0.42128K✅
    deepseek-chatdeepseek$0.14$0.28128K
    deepseek-r1deepseek$0.55$2.19128K
    o4-miniopenai$1.10$4.40200K✅
    o3openai$2$8200K✅
    claude-sonnet-4-20250514anthropic$3$15200K✅
    + +

    🔑 Choosing the Right Structured Output Model

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Use CaseRecommended ModelWhy
    API response parsinggpt-4o-miniCheapest with SO + tool calling
    Data extractiongemini-2.5-flash1M context + SO + reasoning + cheap
    AI agentsclaude-sonnet-4Best tool calling + SO + reasoning
    High volume / cheapgemini-2.0-flash-liteLowest cost at $0.075/M input
    Complex reasoningo3Best reasoning + SO + tool calling
    Prototypinggemini-2.5-flashFree tier, 1M context, all capabilities
    + +

    📊 Methodology

    +

    + All data is sourced from first-party APIs. Models are identified by having + structured_output: true in their metadata. Aggregator providers are excluded from + ranking tables to avoid duplicate models. Pricing is per million tokens. +

    + +

    🔗 More Resources

    + + + + + diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html index 100d5b8f..1c3fcf74 100644 --- a/site/tool-calling-models-comparison.html +++ b/site/tool-calling-models-comparison.html @@ -1079,6 +1079,12 @@

    🔗 More Resources

    > — GPT-4o, Claude, Gemini vision compared +
  • + Structured Output Models Comparison + — JSON mode, function calling compared +
    • From dbd8bba4a3b93b66090e619f485d8ec635f8f0ed Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:16:21 +0800 Subject: [PATCH 266/311] =?UTF-8?q?Update=20README:=2014=E2=86=9215=20SEO?= =?UTF-8?q?=20pages=20(added=20Structured=20Output)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index e18953a7..f44419cc 100644 --- a/README.md +++ b/README.md @@ -564,21 +564,21 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 14 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 15 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html) | ## What's New ### v0.2.0 (May 2025) -- **14 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation +- **15 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities From 14b9bdd8d71ad9df82b63ac079e64e05ec7b7e16 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:16:52 +0800 Subject: [PATCH 267/311] Update index files with Structured Output SEO page --- AGENTS.md | 1 + 1 file changed, 1 insertion(+) diff --git a/AGENTS.md b/AGENTS.md index 641c88ff..de57deb6 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -61,6 +61,7 @@ Curated standalone pages targeting high-volume search queries. All cross-linked - [`site/ai-model-pricing-calculator.html`](site/ai-model-pricing-calculator.html) — AI Model Pricing Calculator (interactive cost calculator, quick comparison, cheapest tables) - [`site/best-ai-models-for-image-generation.html`](site/best-ai-models-for-image-generation.html) — Best AI Models for Image Generation (DALL·E, Imagen, GPT-5 Image, Midjourney, cheapest, free, open-weight) - [`site/best-ai-models-for-vision.html`](site/best-ai-models-for-vision.html) — Best AI Models for Vision (GPT-4o, Claude, Gemini, 1,487 vision models, cheapest, free, vision+tool_call, vision+large context) +- [`site/structured-output-models-comparison.html`](site/structured-output-models-comparison.html) — Structured Output Models Comparison (829 structured output models, JSON mode, SO+tool_call, SO+reasoning, cheapest, free) ## Key Design Decisions From 010a7e824dfc9e73ef8282c1fc8402cafbe61499 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:48:55 +0800 Subject: [PATCH 268/311] Add interactive catalog screenshot to README Adds a screenshot of the live interactive catalog as the first visual element in the README, making the project's value immediately visible. The screenshot shows the dark-theme catalog with model table, filters, and star CTA button. --- README.md | 6 ++++++ docs/images/catalog-screenshot.png | Bin 0 -> 90288 bytes 2 files changed, 6 insertions(+) create mode 100644 docs/images/catalog-screenshot.png diff --git a/README.md b/README.md index f44419cc..4c43136e 100644 --- a/README.md +++ b/README.md @@ -31,6 +31,12 @@ Machine-readable YAML catalog of every major AI model provider and their models > curl -s https://github.com/i-need-token/ai-models/releases/latest/download/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d["models"])} models across {len(set(m["provider"] for m in d["models"]))} providers')" > ``` +## 🖥️ Interactive Catalog + +[![AI Models Catalog — Interactive model comparison tool](docs/images/catalog-screenshot.png)](https://i-need-token.github.io/ai-models/) + +**[Try it live →](https://i-need-token.github.io/ai-models/)** — Search, filter, compare 4,587+ models with 25+ features including dark/light theme, keyboard shortcuts, price calculator, and model picker wizard. + ## Why This Catalog? | | | diff --git a/docs/images/catalog-screenshot.png b/docs/images/catalog-screenshot.png new file mode 100644 index 0000000000000000000000000000000000000000..91d1e2df5b30d4193e871b8c772f95b223b85031 GIT binary patch literal 90288 zcmdSAWmFt%w>1g`cXta3ZVB!XG&sTC-QC??LU4!RZo%E%y>V#VrE$J~&)zxTIrqo? zeeW1Gs=G$_Q@U!cIp>;Bgo2y|3L+sQ1Ox<%3W=cqcLP3OLB`oGHzE;KAmsWPNtL~tXT zYLr;M=RI$5fOc0%i5HYn zy~(U^NpS z*Z@O)W~$@EyPfPZU}g;~>l3CVRv3wmM>Vs!N9??s@hWA~C3>IDgcLa_tO7hwf7zW_ zu}^298SL18>|pnh(9rU?LO$n%rV=v}GZ~B>jm~ntF-b3K$HjCJb>1YZ*>YZWZ1Nra z_agGsoG*S}Zjj6pN^}P(8HiEtpO%TzCBlRSj~drKVW!Fym5n0SkM#3!2Ku`SQ!gSy+3@x1+$&2!CqdkV%@7H<5abS&>UyE~xK|k*c3T zP1HrN$4w+6Vk{}2W@g5PBl{gDtiWvtOFmYUw)aDm2pMx#k`RpwipP$b?c}8cCp%dC$zQ*=6HAp#oy@!eUo8;>p z@GVP=>TirU_;VYzU3$r()e#oOxq3gWQ@P$j-VtFDsTIs9t1a;I{uIs=O79JPL9Gfn zbCnQl^w{VR$M9K}5L7Y6Gymwjru|x%jHFM>%tU##F+@hpVqHQZG-1Y<_g!7}N2k-- zWJTY~^wE*e&ro?%TtY%%%}N$iikw!x*}>%0#?u$x-S)A4YiCe6y5|CO_~F+Pk|Uce z7z$50VAN;1jKXvID-pZ9vdxpYy(=`=j61w8!3dmhL1d3t^KpKY8!s4ER0VnfN;K(y z;qyOJ@=QJf|3-5EK7~+WDhErp7^+!v8IeTx%PDnUCBYmTLgpksY5ny>;TgK&3rW9k zo-YbYCrQdP4s8_TwZ>3UP4i1)&hpO`s9DS)U3Up@y83*6$KsGOd7RvIza>VjBAH~# z1ty><8-CvUVQMfDo;U%p)oX!HnGQpKuZ~w#nAZ2_$OkI9to)^;4X=G!8(GAX=LOXzQGBC!=rs# zE}7*jPnD$P4a+>4J>AC~1n)%pZ%*RDPYfHl{=n-c9x>Ygk(O10VS-WxheXTAcT8o&Lw*=UnLS6r>{nsd3;M`N{^l5 zWNWmP{i*0Q^6+cNs99VbI=m*08c}Y@_29sPG>&6wJhMGl*W?ljr}`7^fD%QLA;zYu z3nyAQ3}4E^qe$tS|KzF;Z5kOK4Hn{OsgQDnrQArVBTbY-sUQ z6LZBFg+d==w~{KWzb64_j3;0n_}YHZs5p*=)>ZW7@NccfwC^1r&yBrJ#tbBteZs)H zmz%>`bum7UKKO-^)ebjnOvV-ZA+2E?>uc0xlDS=D7@tR zK3*;;DtYcaAH=i!TzHb{k+_}|3Er9fQ;zyw0<@`O5R@b-DH7tT6E96zB#pxzclse^ z5+Ec`P>)KL4#vMvF$7{v29oJ#69$>B!+8NtSDcg{1f%rnhc z=TZ9n*3zS7uFq0rIyE=ns0Hc~3CrS{nSN>3QEa0sbYAT+;yrOUK2TY}7tX}^7*%|@ zbb%m{u`oE7(I^TLHL!CDfs@lRjym)cvk$R@_Q&Y6UE^bW*1J&tPlR<0?y$Zq+w3>9 zj?&`v_r{(I`md8{`!|`B_G#_vpX)gAMdS&+n49#=m`2wgRX~sL{#yaDkpyglc0(x@ zF3#!K!57||@ZI8fSu;XU;Pg!*_~Lq@mM0475H2q|CF!-1dZvRmBXC=qq@AJ<#BfaL zD28(*+;Pk4xbWsR-O|2Lt7Uov_VqosAn{W@}K1(;wgTFUL&a6w{J{A8HTPe1hU9llEYL_5o%B`GmuB$%>WS zBng97_il{O5nC^07KtS9lc(Id+tg6!m6o>t_k+B?oK5?^dN_;_)4OiGNO8WVZx{*J z&!J@slPqSwZ$pnRc3w-MHjt9J*M*vezViEXX9-vR0;uAjY8-wN6qiHFRadNto6UfY zxYfxTkx7kY*u#%6AE*${8zMxKB=8vHSJYeqYs$_NwqQ|zbQ4##Vb%;1Z0mS-;1ak%R4H2Hz>urpL_t&K{)V83tQQM}AwGOA< zr$)e=L>x)ZMm;C4tMyu)%N*?Dr26>z=`KQ$2zIf=t*p5dB;Yi)&VQ&)a?5+fgjLz< zUd>!h`8YdTscVh)>5XfjnXs>xC1%jN8xE<<5*b8Avnd_nOU1NxK=;~-m{nW%bOALa zE1Z(sV_W{ihuK+}X0;3X$08Uv0hWRjkTj%Ax$0xI!cW-=gvZ(hY=_+IAB=9z`mC z1XFq}FyF^D*Me@*QO-Tqh}5T3GV9I9yxjt;pP~@kLA`G6$9S^%8z8kFfsS+csD%$q z_J@h|;A20tE&X!$JM-+z!4zlMOXijJ1*pC9oN-K|=IKa${fDmXv=%7Z{h^0a-%;01 z(P*D!b&pU|@BdDP7^->)uMZs{D?{8PEPk0!b%|+u`SW18AS^W`KVAAj1$Z^)-Y%74 z0;x=IjMAh?&PRpa@}W|AM!a(i8CP^iFPNffyk4s~6{VUbo3u4$YU)q=MQtkSE{7wJ z;`m?}H!;81PcrC0qYCc+{=qE4lo_EwA1}9@F)lC_5O{z1kR`|KaNW>wn_cd!Q@c}K z7%@Uy58T+x;QGwwygWQZ@2By$B5_f^5Kth#JG*@5#;6Jd?fra*M=ejB!EGgHu8yxd znbwt?Oa5LyCBQ^WJIHc#YmbVGT3S)@a$i0bBfgvEwuyz0fBVX*cedJz^L(1Qn#3#W zZ8+IlER-abaq&!dw--7BDPO$PULx^K)gK3~mw0O8k|Ol?S^U%HWT?<(zb6oF|+(^>y9ftd-PuHso8dZcBaq*`ugLqC)2>HPdgcSnZs7hf)qmEZbq zXmOm{5NI!XOwN|?@{OdKID*G;Tn|K^IjIqf6z3I2F9JIEyWmL0P(AY)v+Dfj@;79+ z)7;{)Pu!!R>7jl~NPaNFSTo>DogzoK!HfKZAa+s{og@0QP|gfGu$5xOj4#^^Ksm*f@T$e|ikP4e zj74T{TpN@&u6Ag{gqIspZqog}V*FXLREBD7WYUs6Mi*{H%LoOD{&pNGD|y4{J$qdO zwWXjrzWI|x+?^~#N}G?}q2lnW-}}uwA2n5Ybfj=@^A|y~99^F$7g_1eT9=zHUDr5o zZ0XN%_h$hR;w(Vl$N1MUyDro7rAtH%xVNVp`MOCn3NCF`Rm*8Te3QquNRt(fgOAy5 zMJ{o%5!^W*v5nLtSI1lW5q&1x2>S-3gT?4pOcS`>G^!raZu3AInrcKe#HcUkog3}% z-j=<`BE`Lo;<_Fd^M7pkQcm`kuP{{mmWV*ADkq@qF@z5A}>>f-D@ zIyb+TC9Wl`tSVrBd#C*B6Rv5MM^_iAkaog0(i^WnSSNKn7L?sQH|xx=2e;ha zhj7|vlt-!04QqqC@;{8#=+jJX&oet^oHNQnhy;%X?ndKq-`P}+@H8L9XpFBLD-_ZT zj8%@W-@^p+nAJW(+nXPMw}ph`f7lj2cVu8M3={0Aet*^TtP-O@imgp^)(!BnU9eK< z8aKNCq$fZOMyw26$^UYzB|NJ7yZyL(Aht!ehC1Rw+H6kScP9~THt}WAm zDP{I3+_lB`Vx1P-&b;aghL$!98Mb4=(&mXuDyi)EULdQUqBbI0@SnIzW8swknj z3M$4sox(nne>Y_l3!_2Ho?1^zAfUr5c9o4M;h`BFbD0`&z0}wC%&9lK(S5clkfJ!v z!L<0IL1*P$?shs(NADEz+t&^C3@@1@R*OmCb}Zx*kM+U<=-Y-qx9@T#BW|}CoZ4;s zU6(Bh4rXzrR0!V*+hZ+~Ac^{U7TSum`qB&l>J~@2%8ft2dO9(45%GwHm`<6*%Yd*c z{l%#B<_8YhOCdP^WJw&Zq(XLedHRnbl1gGPV6L>SeD+t)HdOtqIN!KkfHz!)qtQC> zai8iYj(copqC(EezZBnO1eUUpk!6bX4!1#Jc}DDg6Hb#N#a!y2z<8?SYa$wqIOS)& z3l)@7&J)&v_*`!HZr=s%*7-$OQ_UUm3}fTYq?p4iMWikE`ph#%``O1UrTp~aDfoxm z-+Jr0myu*vA0@-gRCO%)kjYiYj8#G)?`$AiJT{`u0WABYlDXPuaIq;!$Cq4MX(!0% zTL?yH0!(VT6D(7Pp1sEIN~NSqoQ1_XH-aSho~G8oBlR1}qxlwc=4x764?MUS!kEz9 zEu>yvuFvzP^({;!A6~|;b9|po#4ML=Imgqzb^(v!uZH_xb{#)-RgYz{FbR3|r{me_ z{rGGA?5wQ;y#tMUu}T_00=f(gwqYL6=lU@6VzQX1t#}{!d<%GGVo|cbP$^RYH5$NJ zQn?}~n9j^bN*cPR$C7b;!tD@QvGN-ox(lKjmkv|^HHx^>%4230EY@}Tt#_3`i&gPq zE>|LfhgnT3qpL*HWbO!aH}DQ&Yj{JcVgjN#Bpz!A-I+b^QI zHIL)OFhQ?r-sK0t@%y4IS?Y9#)+LS(@nEF39E6ik9>$;j#ph$=DU37MMAgGI!{Z%+ zXZMs2R-GkblBN18h)^oP^Ww`lG@l`t;6XhVXHIW07mt}+1A~9fFklTq*+91L^lbO{ zMYdqYh3>FU6shAWP1v{;j)MzHqjL+O>G(5M*(!(gGO;Sw5)z?VA@juTn=#GQy=*DH zjX4d>KT4`YW&N}LhBTRZ@*sA^XB-*QK6g4&y+EPxF%!c{!tBO*WyM+lr;r^N?t+IG zhC!z=eh1xx>^AFK zW{i2Vn8<}ta5QyP(h^;Khifp3nHtQ(0Ns;xO6HEcnXBM`74= zh6Kh}MuRJ?6=#PDglSo7mF(%q(&*T^+HA~4(sjMsI(@{Rq7P`1l8gpcPxjkL)i` z1vskh5G#t{sef}n_K|5s4QoGBQTyjL?1pU)M%ef`gzAae&`cpplElI!+;ih-zgWN>|VDJ=A@cU2NCbZlp z^i*_mmuALJim#_AK1NZD%x*5V{^HAqG)!$Fk(c+L4Oy;BYxyiEQTDxE`}5mEqTE&u znn9(=XH43pvZClbe$T_L8jJ!`-mCUJvnvZBvY>YpG?Je8^-R9kx!+D2UQK3Vmdneb0sJl?^YpQ)xllpZHg+?}_iY3& z>fC3mwm8S>;{G(PKmv8lxK7g|9W+=heYC>wJy_z8*2T;)_kfT+6b1+2X9nJTY^!z* zm@FdbuzMsK@S>&E{V96H&?oJp`=U;auWr~F#lEWV;@@{vcT}d5jv3OYTQ~y^4NH;s zCznQ@%p3!LERZ5ZDha}d3f3rl#u^FUShyCcwQG{z1AQOq7J+YZPe;+YgZ1+Oa^{*L zq;(nHG;X%=GHzha^Gixg(gA1tE~mrekqy8qI9Rdpd(odX%jP@2P9a>(DMq#Iwf)ZJ zUZu=il_5~C&u&LUcfUYm>ziu!I;r3(2k=PM=Vh%}gSA!*S%YYEXsF}nx9|GF8yU4~ z-kwM+@-A%u)6e=d){A+CAX44dt2m!iptO|MfrY$NkBUO`J9Vm<&)4%9^T!!xvh8HB zI$)a*5!;Ec%&7vjdpkj|GU7&^DS=otCfkCx7m{(?yn8rqGMecqaI;L)yO zxJ`FAGbgzM(7#eAOz=$D6454mqkxDJ9s)l-SaI;vW4<$B+w3Dvu2$4T_Zso}hJfqI z(f7TgG0U$Jk|U<|E52*coH&b2jVl)!XM*?>K8mR7#6)B&fnt| z$3>g33Y(T%e*M`tV8{H+}*u4~vsl6lUUOV09%{}Eic1+l&NC#gfT<7`h*B6Po zo#2|!t&Fn|KzT*14|v^3s4Dz6u#v5YshQj7Dd)T#z|0;Y)UUhz%n`IcGB%OTMb1tu zU^;fa{~FecGarCLBrZ>wkfru;_Eul@%NFsYycNqrY6VY57D2Qx2svgiiDykOEq9`8 zZ+&_?pC9)n%46f{gF)#C_qb;NW8?ct{V*eWi_y1m-1FxdJE?G$i)(OISbMyn1vpNv z>4NU(m;^gQJfJE8Q6BcTM>o0xwevbsxlgcs53-d$s+sFOjO1l2CQqycU9VBki2Hn} zX$t__Z$*nfCO>k-$V5R-R??zUoUf&V!G%=MTW=uSoXcZLrT{lf8ohW}VPzme4-$dk zZRuvL_Zyr|55oCU73a%W55Oy+E{_3;%tHScClY%5K5lMgn~YE0`DBM(u%ejDUDfS9 ztOVN{Z$;}^!(gHBz{=Uz4+meT+WgvNyy%Sf{rJLgZ@Vs>F1DLzCSB{ZGpO5d=BX_m z(NeMiH0WSy?MW{kGU~rtmpf15l)Io1crXQQNXwG3vbT`+rrIk1#nd=Uxy|dLs=N6c zA|c+a{uxDuXTJ=_v`?@CI1sIaZ$+Ez%aG8~!-ypI9GoBHdV(F+5vRy6UgW-N{E)M9 z=!;jX&@T1)and$xTk`PwaSeUC6)~h&znMH^cLxhSv@!$P;-$>6Ee_DQc2?mVKdUQ2 za->TJ!gOk+y?j8R=BN(XbJ(!l5qpk`3qLoQkNKU7XNmL#`@{ml|A*MG` zpMvzbSluLEwUTgQsWQAG$e5-Y1@{WnzbC95w02n~5sykDBW~fXesTU`(jr^Je;j+H zwiaA|PdvArweT@PuEFVhzh@#rh|1x{{JYC%>qzs9C4H>vX>qkbH-q)3bJ~I@nS)={_eY?Rj>ooS zdnom?SmLXfr@#(D^5herFsy!qkr{U1YQs%lB=I*7q9!L5<>T-JPV#8$ndA|X9dG%I z)9QEQY&?=(H92Xwj;pC}Bd?6B*I=aeG5X|urGj=l^ebHd-62@PoW5^a6Ohn?z@s1B z?1_6(PxNDGo%iY;KZt$$nCH?$QX2%-q_8+`yO*@nUXJB7Q$7u(kZE zJ?{^f(rm&Jw|V$k>)*uHDA%O!P`7GW`G=6sou~NHJ?*xmQ4YTN+4MUv9AV%CM@zg5 zO6})h1eU1RF)>p4Pf7xN=NqRoLNBRP%=+VMNXvk(I;2{r&aF%e{m3yGxCP&j_w?@> zX4E_?3Qbg*)x4NQSSO39nXnrT20H6(3SGzD-I|}gs?}mM>AZGc<$EOmY5^uof~1MX zAe!$N#G~$XAmt5IAlm|dS+r|Z_CZCh#)Pz6NM~b%a=}8qYc3@e{&mtt1N~!rS(&An zb3^x$V|c$rv}fl^4+U@A>3KIfbKQ+a{TwYi*N0zFhpe@~%M?F3wM?^YtG!Ao8;q7P z0USrZ3TmSEBIk|yFFP|x+-LT>^DVMySM~Y`+WTarnQd8pumH7lyFECaG^bVdkD@b; zj~OheSbKReP-n}kz`zN%w!Qo4*$aLfa1cga9XF_-kDNI7HM$=bv{e;%Hq7p>ZdlmK;d40k?fy`pXV&*G*psKuf2xWX;5HaN#G?thG}CXuc;gn!Db?cxIV9a*pq9 zO%8e1@lCHt=PKRb$reDpG=3Ls%ALQ6A1BtB`RhQpm*-gI{`DN47iSD4KY22VE~7=} zZIDV}ZeRYSh{rvRp-yT?hGToTcO~ZloNvpKUTdT7nZa$k+rg?1uffMHHvE5qFo{8v z96JG1yCVh;XP(~eT~BNfj@$d!tyqGOP0>t3hZ`0L zYs7pR6Bceef@;2n*K>LyR`&LY%+j?D5 zh9$M|eA#bm?$^TyXpe;3x7S^~6@$MJ(7cMXg`=NNtE-WHu9u=z9ktnZ-;H}5z*XTX zTmG?B_IabTWhx)7nF_+S)TO#`suBGHz|VO)>kEKd1%cN+^nGsvhj28 z;oLtm@nhuedit7rZ1jPxjb2}?#|{4LBMgiEEerlr-KVQV+Ve-7MD zr(`wFvk91_Dd75((JwbK(ZX<|}UK)#uQd?_fxpbFR zTa(v7NygYbuYV+$bHTC8o9mFz^i>Qy9Co{Mw6(rQ3I06F& z+4co>Za)D8$e-J`ZBQlu>|vWi!3PsyURBI9 zy!x;Kezrz`{V&=~Ji!pA*$OpnE`{5WEll$5&#TBUm}*1;5k6r8cUnI6x&-u&Zf{nO zW4)7q2Q@@ni9s~Eo@+N%3=FMo4ee9hNjKo6XNqX)5|4Laj^L%t>3ScgF!~XRAE*4a zNOd)EK*^x1AqJ@9qOXbIWF0qXm9c=pJO4MCd_z|YY<_^QRJCfBnPB1XvPm1Sk{yhE{Dt|JO=HLJa>SPbgQJ?jNU-W^NM< z=86fA*gy9jyx#E~FlM8mqfh>xDF+(!SAQnQjZ%Sxe9%(V4*pPUDy)^VXO0>e&#h$tm+9+&d7S?LVF>$wbFB$4;)eGd7An#v()SzI3X$uJN;_~7 ztClYqvmM@fW-N@Z!i`oZe`}f39yTP&98-e$r>U8aDvyKRqY0@!wH1$^CS%KzR)HHe zVHz~l7oH3S()o-TL5Z0rW6DBMPDj7CX!ZNW66F`~Lg@|8ayEX2`Y0)s0f)6O;H8= z8lErkj8CSC5g2Fw3zqB|69M0Pyal@M#EV`mwP@VZku)lT=KjF=|JS&DlLiHdPxrkZ zIj(J47-yk#M-9z7gXAe=CbjD7(s?hIXn#kj{v9#UV2o>?piYcSnpav?wC$P#iBC`z z+GF6lNC7@Ew7)0T3Gt7St|2rz8ZtIK(ve3;me`FN8{DXIu=MWt|0Pr#-*h80a8-d> zzYp?IaF1@mpMjT3ikvz~!o{lx&h@`2((HM6`nT}4^z3R>sv^#?CAdyx)Ku;f2ZQVS&W-v0--<&4!FPMa!YHqRoGDEFZ2Xsi}OdfUMlx-2piUCUtisgx~ zvce*cbLfSw5F!4nXp35z-&@5`TQet)NJ~o*$_?%!K_%6l>nJ_lAFeXe@=a@6GK@5& zp=80rWSvW)CBjpQI@~fxko7)B-Jr~!qVZ8_@ zd~kl>klahVW9693S3khJn_o}0Tv(s}?~mSHbG|KD*chz$gS>;$gqS#l!}v~*ZCdVp zpv@h`31FoTN999Z>n-7Q_nBsP2Xu)tCxd$GJU9P4z(C1F)n1{0?=gy}kVO0a#;rB# zDeXVk6NU8e_0U`Vfl#GODuO*=PLM2|`8e8dZ}Kje_v+~yBY8fWLU$VbS;SdUwnxLq z<6AYl$Z<_LR-pm!q{@D>g$ob=r*w#M)Lk_-6?Us~@f)OQ;tj#CEP%3VrdgF0659kM zVU<~;;@F@)N3P6=$mLHtQ_y_b>Yuri(z=|cAjKc1^I+j?A45c;#izvOp(v$BhFAjRii zb(JEz0qHIL=`v?h8{F#2((A<-qjY%m!MMAH6`NPeQ|j)(UR5 zn$PL>c){xM@+;h*O6@M+jG5E(alOKs>_@VAS6%E&&U?A_u7QTAYptO~FF~EUaM@7>KtL^BT@92ay|D2@f^>YfnZ7RP6 zRifY1%yZ)(On7LJ;4M6+(5vs&kwgWFdDr#$Or?Z8WIZXLGuSmj3|sJc*&`Aa8ZCc5 zUnl^{=khcTuDus(m4AS1)m?kZaA+ipP45t?F!W1RnrRt{K4~Y(_xivJZW_|+j;Fd! z3?T8HiUuU4CD`_;qA`*#KlV4h{O5HHH3MAkm3tP5*||hMn3gJ&;uBlWFX7*MMs>DQ z?8evTe<#VY#Dn{#n(D~@C$J7tV9KKMK#WRfuGQ@%dy?;G(V~3M$B&Sf?4q%5F+KWQ zxaaB(>XoQ`HFnh#re}iom+K(@OQ6+zojQZPN=F37xlp`Kt+VAX@c0AqkI~^Na zSV(%t={g-SKLE>_5s|JjLS3GS2?<6Inv55(OauI9|4d{|X7f)lF`xb7bY!GWxlorMGhb0kHevBO9WYmEHhp*57rixA|1HUylJzk51^DYE zCXeGLoCFPqE~3g)waK$iCSu&|o^xqIaM5A%qRu4NZeZZpAqssVp?@b6qJmT_%3`t_ zeRD8L`I&Oz73Ua_)N&m?M}`Us5pkp4fgj-GErt&K!;42rSxnA#n&IAK|G&>asgLMMyLGeF!0Qkf^C}sf(QhsRVPUA*~i#7ysS zrXCOaGVjB4hXs1W#Y{P9QI^Pxgm9Fbo|W}88yg1`6DKvb;6sNA0OVT`vALSVTC@w&mk!zHCR3DoVc=g2h2!f=1R+~s3=KI-P+pH z;r(MID|IkAGxK#}VIdx*wwF|oS}bD67FIgIZ}dyNykPJ!Tp=pP)JiUdmQJ03W_Ky| zR241?89j%NL)pBtvESf>TR%$u!4CZ@68da~%k*V9yfN;L%tDkp4NGB_3Hsyde4(_5 z27&7izj;C7tK&}}0TGIA?$gGFI@NM|My5Zh6L&V@qX*q@9nYZi(#B4;i%!ac0eZhz zg0X|iY>&%?p*C0PwKg-g0vDv4@%b1UD{XJSNP1h`DsM*@xo_VDjtPD-{rXxUe{tT5 z3$#U?|6USUQ`cNlviM`#P+j9k&w}(YJTV(JZ`7K~cXj;xIFk^?h&l1sG>7}>mAOI! zxB9F6VK}a-{@=gz^Lwvn>-OZb6tFxmr~rn;fSvr_ikmcm+Ak)6HVtVr&9B1pJdS_N zt>&qJ@Va5^QCK9$-L5>qe7{P??P%5LH*g7rcK!&XhZB%jbgWqa2cWJ4&Y8GC)S^(SBY{rvmY!)*LrNo3(w`GqT$sF=E^cMa(t1e(>}a7f)}8suxc* z@XktYY1M4g_rsPz??O;DZNm4b+cP{&!q95_3-tTPt@S4Kq0sXVt1BOS`fUMJGHyaT z!^b%3G_5b98xBkB>fJL5%jIzKj#sZ}6mO=Dnqe{c2xM2!cAfY9G?~i%2{oL==KUoz zS{ybdC$4(1@d%KP2=ee(CzHs=}Pelj_q5p6&)(W4Z!cF z;k9rcdJ5XjP8kM8xl1^>+n+iC)YMr?qxMA3(alrn6%`4M5n7HpvE}jIF^*7<1j6DU zjE^LOjZ<@`00{cL?&yp!7VF0<5^==GOMVbgwij;~D7I?glI9G&rG*cYR4TI3?!vup zYk3Sz>>(t^E{0BXJ0U^BfM2j^oDMJ=ly$$TnHKXA`3gbv?FrNW_4VZbRvMgAl^W5fH4jw}3wY$_zkJnDa($+gmz0I_uhQm$> z*Urn~Kg^Au!0Fx_jm`MbzI@^O4ll>RV-fH;oUru%ikLhkml|Fm0wMB5aT-Si`9~<* z_jRB+{M!j3T_HB^{4utqg$FGY+};{h>cogszR1J&uViF&i{prcBo&4Cc|9jFaD9_JTlLRom1witu?7+3u^`HDmjcXIf1{ z%}=(x2G`2J1sxCd*VlU!pSRoWwwguGZppad(;@R%xf+r1HlvouX8OW_8x0T=FI!pf zJ)eSa;LunatvE%!OsBHqYS5a%&FAguK>+IZ48`Z`$00I`nrQk~a1XiR{HRzi`;ye* za@t^d&+q+}Gg%^*K&z=9v=0fHEmfH}c#3|F0cq-rC*QA)3x|c^!BY^zyp>|%EYm%A zD9h1?Ww5P@v8B@W2GvEt!RfKnuyQb!!{RJ-D}V=sy^DTYrKwDT2;_8bD^x>;cpCTxiq|*( zHXjfm(EDpzH25uW*<*qJFwZYeg>O;8hytGo$#HEbaBpJW!ORSBS>jfYn$C8cEu1>@0;9JNd-ND@gqj}vAu=DY8 zWr#*FFVCPvep>uyjhVh#DWtKVtLkbA1_W27qi=TL=gi?8@dBy8;ax zG<`y{BYh6o3)MO$kZuV}gucY6jnAA#IN4q&1$3ooq3NyA3%5GEgv}W4?mpk1pqYg4 z|K5|PIvS9qD1j>V_g|Z|(b~SQqYt82r$tkv4^){IYWe(~NoAz}8)Aee6{^Yzw&6E( ze;HbgFld_IpbzMfRgDgwQ`t5Ca>bPu75*4ISXg=PK3FoDd?>vTxHn#>MiQt_{DOJJ zHQeofw){}qfIXhe8iPIAlO=}}&MK`jz8;zWimg}?j$Nw@fmd{O5mOJhwL0%_fMpDB4k#e}yN9hO@{tPAK+SM}WqHD{B z=VnXfWW*R!#g*+6k^Xz&aK?Nuvn<*JT479Z9CWH)l$C#d%8#Y+*;dwJvc(!c>S%Mz4}oH7 z6hI0Bvqf9fFIBFQptK4ka?Pj-5n;*pNN{2X0@@r!2QLqEAEIzu-wJ0&7o)E~avLhxo8&0Z#8^;&B8AG5i#lJF12f zFZ8vVGpq-N)d)#r`}WVynw4%S-lD`%wdM@ZlQJhRQ8JO1R~Fqw_i|F>fJ$mG@R(p+tH84! z2>IFGSCiM!l0>5q#&<`H9lTI?2oaKvMI|;Go$hnwdY!LNQDCPt8qrrrFfOq_?|Qpt zb;{_3Euw7(WL3?G?=dh`kI+>h$4fg&mpMw8|D*`g9w(?u?uU->U@A(E+7-rGp=6Jh z=C$DTlwc#NHVF(8NsYU+=XJ^oCSnAKSauuvih};FCJEF*iGhNxELGxwv1#_(AeU9! zf@;$Rg1^Aa6g{&?SCw8pq9f{=_698o9ZUQLt1v6$7$E(@1u2qZHMCvNg@=TW&2|OI zYC)i;3YPz#7?5){TVT#(M9+2C1%z~r!b)U#+>X-I5_)7^V0rcPnoGj1=%mgTV>0N7 z%|h+~HoXCtb#^4`^v^O>yMMYizh~jaY4Vp zP0XPp_LW9)iFifhdR|e!fH{|l$>>OjwotIy5>(|NFN#fc!aH-=K)~4@0r;~oTmxeO zS5@y^BSE`=pp)9!C|(f2i;s`p?-LUn@eN>lcAQiphK~&U5OoRG;Gg(~)B4;*Q&)4B zb`>#nwiiDwQzS{X$Vx>|3hka-5!>r@(n;Yw3&J8*fgrL~I z?0N0fYNXO*B5i^Q&|`3xzxWE)BVM1-04+ZBZS7oXO0*c+W3-dwjNf82N9p(yb@uFA zgQ09S%|q7?lyxa{rkuDkv%k)!gtvxXgeU({Ru2b;qHTJ$sRsSPs`T^_dL|^S(MMe# zdUfDo>Led!)%sKHQ+0l*2ehb8J=8FA zkybG=`EFhUvT_#<7FU{JEKv7s`&v-ZYdGC{r_oGH-X4v}%En+lQ3jq*1t-S^#xLin zJ4KYwE{}CkQ);TzafsJ9x%V@My%YZHZSq}3#5450)jwDN@oeQAtbT;Db~~S%4K5rB z+K0!!<`2*h@am)o#W9fvjSVD9J}Mp>X6w+XYA~eqSE#<+2HC0a))r^;dp@9N(lN-V zaqOO^&1oF4VnhcT!BMlY?lR!*?f@SGh)Gzi`vbKu)Vo}9cEUtLQQSbUZOci7QpmMY z;Brx)3r==Kmf3txmst;yF&{p z2a0&MS8xx&j8fipjgCCG;W(WV|xZ@lL`NxZhxZR0wiIXVhW>zDYsPmKN zV44zjZO(t-1rNH75UXvfj;`v{gqj}PKv@b3wzdGF7_@LeH;JHItNTOX?0!NAlN-t2 z&NG6cC!*H(?}zOfwvS-KA)I$P*m9FA+>)F^%I_6J$Vw@BH~v=(K%JnNE*Z%J`22Ly zWSp%ZJWk#niUpT#7eK${mmf_Hw>35PRl>W?V2`<#+B+^j{=DB7c~y?I53Z@~*6ERw zQqhzsFZSxERBGF}V~4&~c6BCt#vUK%zDy7B^b`Y*jG~F=uR@hT(~%t^*(gt(A+b>i zm)w#!PBD*wKNs^YNXD$`-M9ePp#_D%#Y?6h<_qY05LS1 zHJ~2aH9SG9LniC@p^P~Oh<77n_X#@^hN3pHC6@rP{0>ge6ppUYGf{;WOKtIxcR>7i z$j~KVF1@8N7G5A}hy$0i?NoGjZrE_%C zC$p1Iz<^JS)U;YFh{Gj4~6->~j^Ek~X{hKgFIH-T~l~l5o1J)Hy7UZ`bfz&6dT> zhaz))jb-c0n89)!f9^U2BP$1ptZgoq1G>$NiJ{6jfJ7lYt*cz*U^nhC zU)o(CXtVB#@l@J};FrTaGIR}ya`3U`-2V$AGw7~-O0eZIdF^1`8!7o?sNcgm5fCBV ztNQgOPVjW0SkGHnSZ>ll)`eZIXHkId`E^$t!JY`yvu2Fg_uXGYOt*ssB%EM z95PMkM{{Ny4`5yOI#vukPsPB%ARKbhhdV}_DJ*^$HtG*|_@@zLmY}H&R_4q&_3_=C z*G@e)a-qtlljD7mYp_h z>&H0HWMfg7X-dNpka%+RSZwX58ss91oshPls0N0r-u6>=hvXJl5qeF)?Z4wPx<%q_ z-#2+HwQA8rnLe{vwYNrewi+ z{*6-6xb)%5Zu$KCws(81sjT19;c>McZ8wo$>#TawSx}_J2ND?y?PT+ztI(mZ>!|h+ zJU5v}7?vz^Q~wb*h``wKkPO4>K7%*qmYOU^*ShakbKJ!s>#G*68e)`&6Zx55)$M9V z{hoOs_Jl)mMM=NDh1z7q#>RY*v9hYEumQU`IWE=PW08{A!9U}({KD}Y6ZM5M&yOvE zayJ$L1xl~ zUJ0CjEE&PvB6a4vIm3ZUVB1aKOj+LTheY#iC;5G-zNxdum#OF1cD^&%PFFBX2fa%d zx^e0P;mqD>{2s%4b+qStp4HuHn|f{?*?A~1E9Wm%<9du{(c$3^b8|BupVE?_JP_th zH&qR9 zTQ*>s#rZKSjUTH%5VPvqQQForq`Usm4-I%je~*tF6X6DC{PNt3wmE0X+o~ls;Vga73$9FMjO3-ZZ>w zzhqU-c^b*+fa9Pnq!=fk=iH#5RKmXVP@Q+nwyA@rA9cUxQEPh6zlnwh-M;B*0?iTI zlwqdvUsvwv88{MIZf?3PPA+6*lQy*q4S$x}5}D7Bt6%LoE9MMF)yT5{QE@kF_mosL z@K8}FesBqgv4j0bjmG*%jjs7eO@#gn+A`XAm?=9;jEiaYC9QJf7e5ayxZg)=&Zg3u zo|!3w^dEYsFuSsjq!G?`ntK0;2qIFDK*hM!vwsZ{;BTdRoJb_$>J=almoApkP$-^f zQ0{y6i!b}%=ca}A;qR6J_Ovh<;>t96v$nw|T1ArVxD;!oVgdYJpyX1Y1|a_b?*w@x zbUHmzLwmcgq0*EbDblJD{U)_sgx2h0b})U{4Wx1EWO2WZC=;qnYo_K3$M|;DEdTjf z5YE3pc5g}(Yy^x$xkUNT4t%vbVoV4cJ=7Ubj|OF)Q007MR?*kTS>}H$`;PAa=vC9^ z1plBx{?hK%xai%H(Ep2=Bc$#4|0&n!|77F%pWs7&vGqT?vOgx$2-3wd{o{g3_5GKh zCH3ftvXCmd^)v4UK8)1UqeQa4tp32K1Wbq zQQN3?5&!Eq23L@XyM5HuP2_ciAemB)2|5OWDSKMIW;uWZ`p2}xiu5pUm;dO=-xIc! znhLELKfJmoudAKM%D4S8Wd#=53mYhGB=8S%C-?DVICaw3Kw+>Kw@$dB-43gN9h+)@ zJ9+Z0O1bny;M;$l5Iqzz+so0D_ow7%2jRcY!){QnK-;D6Z=Rq3nrPc=+6uxwRFPilWE$0 z=F~9EH7`bo{i&~C*|K8cc+9@tugYH$j9i_L^_2gwlh)2E1_v+wMf3VL1Ef>2g@lA4 zppk(*7IaqQJTe}4->#EEmbtmP!RF@f?rs@hSY-Io>@PKMZ}0M3KG3aViL#1{$_Y&# z-xXlRLLugLJE|+8Q4!S8NEw|QQ2g}ybBV+_u|rB5An>?)W$}9f+j4CM-N`<3LxIm3 z+*mkRv&mhyDR790qF;WMmepvM^LgBSJxxuGjYTENLdD02hgmaYtr-@Ez*8d`^+vIq zhC_nOWO8Xzr`YNcezjR?`BhaMSPu~Za}y`|sk3~sZz=0CFGNMfyy0v`OV1xqX@H2& z`{^KT`Xn4YZOC{^PtEp5=kvz3y$8=d;mF7z18bmeNdxs_)_MFKO*&709nchY5MBW^ zvCfA*DraB#q85u)6Eiouvel;Y{5mg73r2|N+ky46bWfL{gY_bbC{;BD4K~9eUlJeP zRd0HUf2}8NpSG(b+5l*jv5}E<9#FLFZtHn_qrvM_QI^M3XXU#ZHD(kE2N2Yaz+uef zv9&qRY_}-)Id8v*kAUAF&ls$59f%=$|I?|91Xqwe= z`9z@AXj9HOn)UWlcZ|6&~`086A%p`u2G_H7ZdBe@-!6@{h|o&rT~el*xTZ zFB@C(XJP_u{CHfF>)vpc6L>HJak1bmPUnWEx;9WF17E9a`&QKFzkey(i^rEY2QeR{xwum5LnGz|s`|E4Pdh5!D=@%ipFh}r57 zaZ6DV)hDV5hG2Ap5B7TsdE)NQS3NHP~L#_R54DL{oOv* z`o~DIToz?9c=(2o?2g^yAzRSdchn|?$mNu8Mo<7Uow#Q^RyQ=ShZ2NBcnLb?Ybh!$ za7p^w$GNrk8d8mJ2xs5poYrW`tK*U=**^o9XE4?DliEngk7{SFNbP>?cNke56FJKu zNVKzA{y0?2^#qU4)6Pf~cZyuX|H}k2)I1kUhP^7mal*tH}`J1RNd<&1HmV%zxW9@lCqv z8@tFg+YZh?t$SY|d_O<$_W7q55c7H&gp*n`Ta@KPi6$j1EJ8*`R>#Jc<^6PsljY-g zyX;~m;J&rfB?K{QaxHVcCm|~ScAe-G930G@F*aQ!_rB=|V_!CFIsbAqE%)wytG80o zV6_Z{@^T_0BNrV_-j8S1s`omc&s%RBmVZbZ3kmrNKD3g*o;Z$=rt(RV#U0!}sl7dp zCw7*Xf%=Lv-8}%dSS&PU!r|-v+VfQeqsDdfYxJ^OrRM3f9Bwj``PI%RJ)hAf7+8#h zsUClSlB74^%BDJqXB6J|o@*hYZrJGH;Gnnj@2VY{zpI+RazzKCfEk5TxX&S{` z8Ok)f9b}@~w8M3pjYeGxZ&&j;ajV=J$V;=Wy`Pl|dGC9!!N9<uvgAkndji>oJrc<`kDcZ!cM*)BVKdrs8MtxQqPO4k{FyRWDOL zARAjHVb2%9w9*xt44cNYyEm044mkwWNN1Jdr0n2h*gBZ%kbR@3Z!*$h_bqx+o{$BEnwC3FcCl0M!56A2`Oz&t`O zZ~QdDYzH1`Aiu+0g;H_QwfUVa$L{JuL4K%1BJe3uk^*!`!Vh?ZC`d##A+UQ7g;ww* zDG$10W3z!?ogm$8%N?mQ)sAwVb21az}TuL|I#3Wq=PrC<{Hx@c1ZZ28+wTpkwoz zTJiZ#bXZKo9g1V4d1#se1CRL6zb(IA^o?tZ8X9my&U+&%a2Qmq_P?qoCe995XDx2; z*S$hMu5HWle$mfvRV$Y)z-M;)s$q_|-FShBalET2akxjQvDq!-X7noZwL+6yT>3PJ}f-^3cgl?1mpxH&Psoks?q6zQN0u- z>&eQ=RXXGdBjHr?IbgP?+1T~3An@PY157VL8fShDu@nlgCZYh1I|Rvet1oYd8E8@4)v z0|RG9nfDI$UiY-#d!Onp=gAbYu!#B#B(PtuXVqefhLagXm@+wh{`#H9H2d~s&T>jJ zcw9~B3tpKkh*|}0mzz;8zZ`%;-dywr+m@cg}Fqc5u`C(u6? z?8ZiDj=y|EwM7FA50?q1q&C$|jzn>O z%*HK&Gg9xe%}#f43;i{(cnG$=h;IONim*2tXy0TR0){h^G5JfPk$(Ap2M4o=;u9gK zC*g6_D2CH!Mjn>;K)rS`clOBNAN-id z-E1iya2fSAjg%3}wh3(s7q`0VYOQh#BWKPxyKr@t6*MANz>HAQ|m&2Hc2=$w!(X{zWt;U11RxAfGs9*Y{2GfNy2so=) zI3(|jrdg}?7Yz3)K6iK>Tat~l0&NYqbWMkQuIoOPKJNe`Ai5vM#Qn9i(&u$ntxDa< zh{6gqzSaNLWxpo~HKl55?3Cxid3nw{hWCXDJ2(dSi^U9}6(>11GnsXDJs|@v(RH!` z9)m5>p9g3CuDxSCt|x-MVBRgDOT$hLc9Rja;+57jZMTODgR=DwN0@o0%aiNz&2yP+ z+pU2!o244}HyMu2I*n_;jloZ#0%ruX&PYoMFvRAj^9=>j7GV+p?eBRl1up)@0Vaak z#GT?2!6NNUuBPU$FW7$dD=)&uu+-F{+_2=-a(Y;5y6y%wocX*iSBQ;>-PgmVNGr^B zWE?Ck^BA~>C1;;ilvN!OG&5vNVntjH@Dw3^(0Y&%kOqXmEe53|p+W~Pb#J1YihoT= z3xz!r+s2CwccFu%u@4583#9u_Lb9Vqb7;q!%XE}je?hFAd(`bPeG{jMdbMdoQS62w zK(UyQW+y$kcEovxCWnq@*#uuTWvNBA8g%C-l!auvcf-Vo$r=ygrDgcM+u?vFwx?vhq}Og)lt|1*-V~W)N6d z{qer|@Q8oBKmwyxE%W^#o@5XH(B=19mQtou3r$Zn1P;4#i9t!oAoe}`F4gJS%vuKeRW8+@&u$&g7cJ+$N@kmZvvO)c%)?|!^EZmDR~G| z;O}gLVOknAxEF&U5VO{SRs|z%ooJM51j<$&UW$fFD!y5cdu<`|$vS}oGh}!>ctEeA zeFSPIOh_jOqDC6iC`Tg(a%Ib%Y!yOK1j|ha$w5y0ymoy+PA=)OPDk%*m1NhB@l8 zveI9_g@@RC7A@_KDLGY3RjCEdB_kvg#(fEOflMDa-6d0kzB%Or4N!n}els?ae_u-& zh&oA6H>+MkCgItAR_bk@_;8z1yQZ&wx!i(L#1hMPxY{z3`sH(RLAZjv=@L`z_@b~r z2+6FOvoc^piM^B!CmgJ2m6H~-4yYg$K)|xi#eE3Qj_a4H#5}2rkW(>-xS1Rq3)?;`Pa^tkwY6~$Z9JXJad;HL6%nP^b#fJR=P6wWs#>~NgU^r@HTtKi^(X>26E`K(`t3kL z@2J`POh%nBdBA=wrq|R~w#dlWO&U!iFxDRWwOrM)TQ~ZtMw%?KD_^%U$!8KQDfH7q z8i$fBp#T^cJ$8h~^(J@Scd9b1I;Da7AlE4iHX(^eLAo_+5g(fZ2;i+>l~ z<(C!NBXSz%mmcTe#$SKSob5t|BP5d$IBps$7@=wG?r9Bphzc2;?)16!VR@~mLg2*| zJ4??5mAB=6J01?wEDD5m$=0m@KJI7h{jhs6nGpIb#A9m*DZo*Lla;D+qczKWwauD? zqg%aqehm|N;$nX2J3=Ame*?M&vB{kBxpLi>*y*Fq)zgI3R%=UO!)q6}=cc=ObD0@w z;~BHf*UiLfS}*@vGlQc0IJ#4j{s|>2=zo`-N6g+YZhm1!Eu0TFBSXxX|U$aU?v7ZAX>y0VaPwbiZKKtbWV34tLd+hd>L0n#<(W8Dl6ZaRqU6f#g0ga92F7aNSF{D;3FZ z&dH^-qRm5<9J&g$6^aRUJl|=lKV-doLitA!cn&M*B@qd@5FQUk6AhQ2XSx{V=Kd*_ z+<_|G3r_O$y5pJ&=PZ{Xou=3FNMtfdbX#by4?c-+N$@BDn<^T;u_ z3=5@kp}f)(6V&o%Hz7VveC9$MdmYMwAWiy+iInMG1Z+hOt6@9E#RK79+WYslO=N*K zMNFMC>}2;Vh`dKbdy0L!mE3=L1>^lL3Km7?o8(ui)4C& zesUHb&abPjrm2!8UF*7stAuy!BGaMP1O1{%Br5_GoL|EZ8nmGUvR6VNVUP({3PMKmquf^o?pE~;Dl~~A(05T`;l7mnMx@|XP2F~INOe4Rr#gXZnQ+SwC2;Xl2Adei%g6=u~fD$zYl zB7aNy4e>+}9p}+ARz+p*#%q+x_i4T8rmv)cjy1(J)fPul~?D^0zl1pYxd#rSmnfi(74RXv@zN zvayQksfMQ`PtM0^$bO-o=dbs+&9_Nnu8f+E-h2)Qi*f?i%Pm+gy|-_hIRYa;xvo=8 zvx2CaZBCaoL4x=%8SGw~$qYIhg5IyA@mjX6ET;d6PT4J{@IE)Y47!m_Bd)`y#=}&5oNrtvKsMgez_b5h6MA8>lh7PX_{ekNYi1 z$2_iQ>2Y!P2NR@o7E8Q9hX9eus>N{EgF36OA|qS+r|Hc}(kV!iH&2VhW|7Ne2bCHX zp!T$Q&3h+kneG)*^v4i!1-iFjXYnNmRjr*@H$@1Ik`KxW+1;IHM574%FT0#w6VTrI zs3vC`l#@h>Ul&@p0-AXj(igD>>)xacb!uulezzUbbK=$PRy!;kqT@vjvRLMlA8A_% z9_M`}h?P#8=qD?VsGu~Uh;n^gVO5?uT|mh%dz(s0;#9J;MeiYhR^_2f}PsACY<|_1oToI`{`nj|%w{*_to`knF_IDGXnj7@JL(x75PJU*R zz&Tn>CMOXwHsE=-8#8RZCvqr?leB|lO19OrbSbe$>m^yJoR_IxRRYiC$qKN*?A}i( zVp0FzEN`lo?0E|s3o1k>%+)(5HnZMD4soNqW~JWX1sZ9iY}z?aId8kru@R!eFZt7W zP4iCf)r(e4Ji^Diy=1178~%N;51@6Y{!z~HPfTg_)Kq!FD5;z-kk^<5J0P#Z>((id z6Kk~K*%SLgAT*4JW7Oa4n;-T0bgACb#)g4oXl^+8_Sm-Ra$woU)skw={nDxNmgNDl z!?@9Ab!jMJ?FRkYX{&v^VYxYL!xGpy zqbuG8+zS-fF9WTvjHD!X<0H(I#ceZa%k4pdrwj5qcz6T^An?te{>7Tx=zW7UfRr+2@VDo66vrf?uW?`9y2tY= z=&U{2eE_=SV&&R&EPXTHz4Ep1a@(v8m(PQhg4!xEGgDVbMd$ujKsV;*BBvKnOCDaX z$GyBuIfk?S-rpKK+b;$M-u62y4N&l-psy0$DP@n>_`CpJWa3mkD=YcFZ#$;e;}0Uo z)B?{MpjQeqITi&*0GpR13G6$Y<+!t|&KI1!>95~A{T zxr$00OWFvc9caR~9tdI>m!L7!cQd@nL!@$E!c-OyC%M?Q+q<>*`meDkTioQ_lzT5+@1&!9Ei2_a^XuL! z)QN!8N~8_VuQj;+bKiHMJ_cOvb4D9Q&Sz7PCn>e{bHlq=eSLkvW0I|k{s$HX>(s*- z`Sf0$lI}*&H0Xw!{&NXIdt_*+ubh^={6yDYoPq+@?ry#EH-uuw;n-C!F0OiRu2vWp zms9x5m71~f

    P;3jQT~Y%^ptfz^Tx-u4CF2d&@u{Wh)(r>D8ziv6L1|NDBe2?JrKC)_@k~eUvfdoB=g_0{9S%k6)Kv(I4aQx^e#1onbF%LQjW@4 z0*JQvi7zg@3$^dC|Bhk#oOPai1IIylbW;C6>Yx9UY?{3YzF<)u+Hd&x7Z*YmmGYCY zE`RTD*m2O4rDm*9z#nURGHduScUb2Ttg=%cJ$yKth^Qvv$vlG- zGR%Le%5cEhs0e;|OTxtIPPX~@sd?oTX}6zEa++m~5wJ`_TB1=y0@<-YVE_3Gb$aX! z+2xxmEZ|s}zm9ceslcYk`45Z&(jG~uQi$yD>J)hW|7z-wb+~ZJk6|qI5@ibX%g1CD z9i9x?v{92UX;I<23a}COytyr{V}rqcset2FW>-kDLNelG zG=kR~!j}oLn)O&(2_^$$WK&KjBx7IHW-|#H*y;L%f6ZnhQa`hLghq`~TmYUDnyg~Z zr)5yjrCqi?2@S4}m5q#qz|y>|j@Xp9pz;B}T{tK9_bza&;I>AxA|THpC;9+&zxIk`to&!MH2ilP0PA{Pt+EI0y2zDrRRHp zQo050v3{NEhmatNug1&(i()BZK?OS16MV*^_P4UG>QWg}Qd0xc;16sZjxVol_6lnBMyE+aF24{KcT%%Oj4={_&k4?53m0All}$Pg zsg@TtvlC)cA;iK*W9jX0Rhj-!wsbie<)}su_{)!h?Wg)FN8asrUCpFcgCRK&Y{ZO- zHT6NW`Tzc}rz0S{BcQ<1y+6YFaddhqWLW=0FZY-_V@HeHR|U+XKvD(BeOExquRzRL z0OX8%b^Na`);cpP2pX-%v8zYI?9G%1@{6LzHEglBZLyy{NiA84YE+KRbCECt`VQmb zUx_HG;gQlY>0DZV4xH!{`6i9{jqP8H^!C)T(c?*Y7j!Ldr4#(Te%1VcSf|BEV&J;t z)Za1-zqeCUH!g+$8mO;D7L*$J0230}QiNIgJ{SJ=Uk>TZV4 zWD1)17mP*ARkh0JfH6p&o$MY89ZoXGi?#2l1>9eiblkVP&LUZ;+5~3Xx1zrhXts(> zdbH-q#z;wWd7Z#6aUM%KS|?7;Ed>tZM2L+X>N8L!wca0anNJ1Q|C$G?n{=(-2jP>f zeTS)xWqj@#`t{TZufyuPk3uOCa4Y=5!~Jf&RNxOlPXk8sW(8WfsBG)%VGDF=BYR`8B+<`>%F3J; z=Jy+ow8xgC{-uTn^oz&M?wT5!ggo+g#wr7XLW!?ChTb;LN>FxH%IrL!Ev$vUb&L&F z7<-6+i>PnFD8P1UCpQ|EAi`Bw(A8b9M1#{YS!r>*nRs&u8?sFt-u{x;kDZ);bsQff z<;b$I+oR7UNHC)|N6 zaJKoxK)UcvHHZa?&?Kzb57JL6C?Kb?(U(hom7l1jo&Pza)`~JuTzK(`lardAotvC| z+H*F|zdnvlSVXv&oSfc739#6xiv~{X+cmqD1VCeDa4D%QwKe-xnk+T!M4sSW#ef#| z_~UI?=1Z?!tS76G5=8ny*Qt|_izE9n+bS{92`?qS$$8eEPM0bO{#`n~pD27^v*l@; zIn3shv+qN{t+qG^^p$^^Sze<=hmWQ7>C3NAR`{@;ZokLr6VTz(YnO+!TOil@hH&0& zyQoHDwN|%iRXv=a8n%xT>Uv%_Sl#X4c$^Z7DTo?O{J}{O$4V`Tb%^ z%BNO`RDmk-_X)dJTAM=^8@rGS3#-Z^hlXy!5rwiilbQCf!m@11uI-{i3YLy8V>*Ij6LYz`saLc(aE`Ce>EvWCWG^*9 zUC)JHg6TtVjU#CJD|c(SO6YYk1`mS}+x=#u<7Fj!nTPzJUckyl)dumQ`7<)1i5kbL z0ep0Jc9*`hpqaXvw3D)n%NZM~ua<4p!omU`#|W2;n_kf9Os5{nH+i$fens~;I~5@V zD@}axo@PALMXZ8y?QV8DE-=YmmPQ2b+Y#@l===cFANZ&BH*J4RG%r<<-wcF!-Z!vD4jM&<^blV<~E2AD+Y%oq!J->r4V}IU5 zI(|YV;_#eVm>!IIYTxmOfQ}~QQ>AuF(=3{`S!}x;DM=`ca@z?j8qeUX_1LXckCyw| zc9^v6P*dv-IMF?k5D@g59@ZEIo}23O^XF8jZGU2+qd)Z*k<6}=j$H%2_~)Z*IZ4mG z&=4s;kCP@nRl4t^Ic#&X5mKJ!bJh?Zw_9A=FHCF#6cef%^olS{d7qfMF4xV?C*#m; z>0jr(OXS)%8h8x1x7ziHkO$XRd0cJQ2bQv+rI)ldtvFac$J1*!rtjC0m488b>Sk+x zLWRS}>hf{Bu|H{^`=bff#fTmT*?$gr<>?q1uXP06hN{2m1)Gbgl6pfzQK^BhIrNvDR?O&!$==^rewdUomqR0t58WJ?!gx_n+ zuZW2ZPi67((`GO|3r%-C-irc8H}qfm3I*8XWN~?gvt`X$9ZSXNwhzjszdUv7UFv*B zg)NlMTCCT?Fa50EQB$+EZ!9p!0RaylGi|cgWZv6SX!%^I3|?NQ?APLc7nqU#-TX`1GgKOTcDE#AR}#W!q-b(WD#6{P>`h#!*4SJr#*vy_0N4}u2Yo}KC;oH z_<=)j`Ia?@1qOQqoK;63nW(*@g7xZQsBjHU50sTd^@c?&z-Iqfiro@HQ%~75EEN{G{di9ssycY&$Pgf6wH1Mc<2C8s4teAX-#^9*}J( zRW9FQ328kI{56R?-oL%OYjo_yW>KJ>6AJ?ct*op}z2jgTVy8Wn3p#&4O~&fbaDOKu z_UrHI>6IfCXz&^~Q5^&?d)y$FK3_Zrg5|EY4Hm}TfL2gR!(?_^iUk;hIXR7nMjk)^ z{w>U3$qFpYiEi0;5Y@H%oQ42?Q?v1RAinQ5;Prf)jKce3MxO_mLbbaKdFhr8D7V&` z+43Eo#{6EdW`?%3iYVKzGbZWxj+mStLJi2p?T6j^%<9>WRV)2_e>3pNc*X-W5<`eV zR|zQVmim-cPgb}Um@}-#WYolrl$=d<;u!hL^zfLjF(zeNW=DQKRMICJ3EGyWtNGr8 zz2m=+nzV*H+&bIxW9u)}--Ua~2W7LKCp-0w%Ax0x5n?}7AmIlnT;YH9 z9W;vSK62u0|DO6&g1d#66wc;?bbQ;7)VSPwJd?lPb)4Z`tzC9-36u2>aCaV6*xuxF zijtwNZ1uw2n|ro!D61a-iMZbSo(Cmtuu9q3EfRu6!sRj8P!3S1lxcnx3z@7iAtu1>XxT)AWEOYSo~XsW&X18*foF$L`s@pARLR3eoOz&kWp%&Y5r zXZ?KAm|e{*TpQUE6*GAd^ReJX$7qPg59?ITDUN&xZQBZRb1i-svdLz}rSTGzkR27B zufBx{mmC1d9)JN+TRVn>)zIc(P8O9~cUZSzV`mnONc7m}=(Bh2CgkVmzArmOVOZN7 zm$YZw@#Jsm)&fM1nx4lzh-qYw`1)Hn93PntZ*XD%LaY!8cwlOLAUP=1vhg`T<#^?f zakqK!zE_HvTWO{yQ8c#mbZQ&6>8AIzFBou&TK`!cmk#7H)~admGH*()^RjF%CbitIo4H~VpIVl~CDr5ttsb-nIThDXCT zJwuNk@-E-NNO^m4I82VDW;H@$|4@FELUYnw~|s+NP9+M~MjykDfes zCILjC{F3RZmiw;Ln!0R}LS+cO@Q9#P0?X-aZB0@tsh+KfOF%Y|G`GVppRP zn!R`MnZ4eo-i(Pw(9+UOJo&QF4~++eU5tE{mds?fhIR>pM&M2Nd9e8gF&>G%JZe$R z@VK?j<+vXbB0G}QO*gIvWqt-SF+ws)Dpc}&*<-owe5CWd z8!?|V(sbN!AreXMwW(d8qm1YuX*SkIG0@HyrQB#vcg~8taFvRXlK)WEW;04NBS8HD z7BXic5}zg79a_WvSK-O^KIvyP`~D|@>eK!HVEiZtgo|QP?J)HX0>?b=Pxn``jcaT? z4peFvF>$g!3_5KGJ3H>J24-w$!`z>5donl42LvOET>hdxm5=;x_}$=ih44Q8TQRY~ z`;G{!H#wi=b5H-hFa|Ndo3FdS^3cbx9yX81;{qAn-wP+rcJRldiMZPAUzZq4bmA<{ z8oE+vDpq6d%|2b0L#p!_VT>E#2{^pu#Rz10St;#zd=8)-e}3_K9~&KQJh~@Z?n|q? zU=$DNC~nHnFI`GgX-bY0f>m^fVuHt|^W{M( zWkP$Y^-@8agdE&#zJ3A|jZPOfy{>&Hea9>%YqR#2xs#O4)b#i^=V4VZnJ8u^P60-k zMNM^W1x=SQIy@#jCL6*X+bnO0CIa~T?SL5GXmK)a&deU5CRxDA@(g5kwq>=wjb3&f zh^j?p2Bw&4wbyh=BC7T$5TU64ghfd{Px1^JVzRVUa*~NMr3}y}3zjN(45?bN`9Kmg z@Bs`sbH!h|Pyl2DEL$(R950{Y4DLfkF@ptPRnrk> z1-J#V=(Ym8T|ox7tDo&>^xnjTo`$9o&0CZtAQo$>?8HhO)arW1>7?(ox?}}lq;$7G z-s#>89abC{jtiDA{1}q(Vpk4NV|}4eAQjNgOAasC4$@4*@142mJ&`^`y&6I9$hCP& zEnOcZIJPLIm5f;7Ib5;6`1wsZ4%g)Iveq?s3q2${&A(Dp=JtrBqpYMNz4k&##@JoB zNV$w|Y+U23UWX%8FVDB&(n~f$Fy9^C$-WE_P`&so$tAu zRgM_+b~4i^w4%5BR(5nry*Belh`k<^p9aK%u!;f@myrx5Ugt6J>>Ts6P>Gdf_m<}< zkfsr;EhxXEPmTQ%i0yS2umoxOLlVm-9vJBoL&67MRBl?yA8XW&J=L$0v`| zok+(>8x$~x4{3VlI(041_y=j6kNf!beto$el&nsH8WD(oZyg!K$?(OG3iWDLE>mj7 ztf^T&!G16D{11Ttxxnr6C{TpUVte+aB!s2jESWZaY9|a+_m^O_QD_>HH?}kMymw*P zrgOm}!ayLwZ{>4;!6jyLliV!22>m+rz4JNnqn|iD(ry7t3O|M?qTK{1nCcB0neCP{4XA?mN1t03KFWLYq^9MAtCYZ!L8XgaUZ^}tuzdIzUg=n~4gAj}B>B~}q^gDm32w6T{$cNcFJ@Ra_ z;L5w@@nQQ%dGd$w#BC~>0moy2#n$kX9@un+vRyn+@*=_o9ud@30pUv~&YC<&3$?O9 zsX&OFyZcnk<_Wg)=ORqJ%zt_T!2VMKClB#EUd;I7wT-`J)75-dHEoG94QBXKugwVk z^R1re-QjWNtL0%$Rx~m94^(1qP`5R>!SzzJo#STdCQA9!t!etpXydA1A}|~EY^r?XNJAL?N;BO@jN3$pyw@mfG z=|czIzRiFl+yKac*jcI|W9TwE;COh6pQrrn(RPk=uw#4S;gR0Ds(~1O6EC%kG;=a)J%2(@?s;YEGYLMrCm=B3^>&byC zo;79a#To+u$mKc%qTJV}f&3qsGxQ^?qRfOo3Z0vp8;N9c3QSG>BhJ4*Dn9^(mqBaf zL|XK4V5HU|C^Rro^d*SMQ>{X$9^JSMYP^KKe(e(l)j%{co8jF^%Inh|pZ$}xm@!ut zBd5p7PW85_MU;OCkt`uk@H0X(3QkIMQ3zm2Y-T9&9A|C=bi~v~k1X%rC^OT*{!$v+ z6>oR_?Dg6n`Y3#qhMsFEt!R9fr22r*(Pf%0XAe0w@h|PI4fL;1wTZ)>di92fe|DvO zwC+}I1-yKm-pATTrr}dq>$O~KU*lgtW=EmNujS#~W1)Ozb!3ri78cu_2X}5~+&iZ< zR2#!5gbES^_>V2mXJW4~NJO0a7rwe)5TnqhHiex0{FGllQ&Rx>MAaWj6h#G(HnV*- zV4avpIC6q@3O%Yg6|&AbF4oyrtb6ct6mBFS?0jtKT(d7I2 zMEcZ5ws9edy`xIWiIu}U-=Gh@gqyn0Xd;WpS@|PoRI0;wZ&!afMDUbY5@)Ut93nGNp(U-C|8k68PyL`wv|C%XI-Nk&OF1nxMa z9*Rj5t~rVL-rLHhj@Y2vi23zy7@ltBc|#;rRA~E)D}ocN#c55ABBtxiP-ye!qUXIA zir`dsP+C0%di4#;ePmda>d`IbeOSka$P!4d0%eZ_=_AAg&cvGuB<`<2=QSRRZv;Dir^AGrb^dv^&ut ze{TFU9*Zm%Q$S#sQ$D!x%Qm^Z@h5qIm>zd`jkuZcyLd3T?#kBxNi+*-h%ou(3-PwZ z?*2YmpOJaiJ-j+?X|1C}B6A7@j}}Ejt~DD4Cj~QuY;Z_y?k z<*hJ(%|y$viMbh10FTI2vwT_p9a38#7%k^Uv+WQoabefQmn}TFx6bew585LIZ*W{z z6&<+f&a0MN_twt(^c9w_3rtL&_2Wi-5me;wp!{G&0zWDeW}Ih@yKTk5#d6i*lV*cx8BYVe>Dgbk^GBlwg@9m}(cvFx)oY&L+dV{;tJ6M1`3>j}wWXA=|xYBS?h)8=L z7s_JtR5QZ0vX@L-6odiE2piVf>{wpzNs|)pFBA#c%06Ig2GeK9uw*TQ~d} zY%VPfolMm}m&U)yv{&d43ZjPNX|NHxpkZb%1)<#QG`r5r{@`A*&vYT7`V+iPQ+)0V zwzA=pJRQ`jWV+IIOR;qq0|TGbT;d_W$k6tzYr~hxI2c1@e&$%35gHLCDR-Xg&i=hp z5nxYi6i9}3nyJkbSnT81;O={#43W;rsw5@?$fQc!6Ed2XCmUAYw``Rr(jYT@x04dn z9Jof?+jW@_F(x)P$SZ9}T_5L=3Aydbn3=VyH?VN_aCUb4sVGS?>AGC^L&AtAFU}l6Ri_=&A0*FdaH-VactudE9fDrItPV3sM5F`siJHq6-pxd2400t}3Flzoy+uk{${c>IT16b^vuG zw%4?}Z-1S$e-wn{Z8M5(eiB1`K^_s62%F|(-AvWehIMs4_Kwx3SVjh{m|1SGiFL6j zZ|Ud zTXJ3^7--|2D*c31RJ%)`pg`++qfbLo^4!wo`3U*ot%BuPnmDMd0-F8phJF-KIn5HKl$KQt;J}RrNFOn3C zG_z}eY%tS%g8$)%H5NxFt?>TxvtR;tBV~eTq{Hbl>4fbUg;&yDv|f641Ge=o!Gq#{ zgQJ5RgwZiE)DRBQR?BVmm|7;eiv901h|vPnDwW^SL+=p>`0@|svnK6HdHM)aaHzkB zrC;&U$XlFVT`Brv-#Pl;T$Nl=iqZtBp1K6HayVyy|E?w%M-qWO^iDvS`m&9s74g{Wl-tB75AM-o{gvQ3+?2qN9r zxygU%9&8`gImmkjLB;<#LN3iy`_(TNwe6t|vl5-k3agFt}ETQKZrQJE~a* z*}W)b@(oDV}t!6uiz?fm-amrP#iS)upRsnbX5gxD8% z(RcY<)8Pg-`9wiP!>I2SH*h{YHe?#7nv+-VnW{VFB#6RF5UqAEYmk5P)_~nZ<`p__ z?W@gomNC0X^um%4F-@xd25|}q7_&?g5_;g)959-c`}ascKA$~akEcvio$+Te`Tj!LaXmrtd8Wj z43i4}i2q~F7p4fEq$Hp)Fff30RyyzU2O>KF^%@MHH>CM>x)psPcXd7HygHu=Z)IxP zDMbiwQKA+6Wn&^`N} z8B_6i=5<%A(=5o8t>NaO6fQA|DR%BP#CM3UrZk_{@dk0%d4uG*M%dgoZPcW!q*P49 zU=xI`g1#2;v?GfBXaarA06%Z?faRJ_##;Xf;4_G95v_P46H$JJ^WGA#qP34NBNH&0 zf}K7lc$XS0O>QJsRWx%p%D|_o9WmS`_sGS{NONSrR?TKF!JHmaJ49EeUI0DBZo%ky z0OT}inhpH)SPMX%R)__a@6=Yf<>@c$X4dzC#O8{;%&V35h^+b@-fTC!3E)qI)yahX-q~SQ)S-?e?`}<;iFE%wrCfC$45dssTqat5TKqt@ zJ}2RrFWI7XDrRDWTkm;Pk**>bMg&V^Ns(_=(V=SnCl}D_>MVdk;~vgXtxZe9}{%gDdq~6~Zx^I4;J%P*GI(S@h>g)`K-MZMVFLkf4 zXz7QXneu2>q}^~Fb8dHI^>laRmfXa?Yg^|kgHtabtP{*Q#=9;#9-&;+Ie2>TAQK8D zsKTP^7S>@sUeOS-k1UhmF(4sM4ZQg}+Q*GCFXeAJ>4^sYMv?f7#iLw2BvbSRy z)9eI(2GY4??|oRU7#+iZ?yk-|cvLG9rAOVXX$>}8`C z&VnK3w8fJvor4Sh843Psd6XFeJ*+eDmmBP*Lh<1^7>}*je*^>c>>Kz)$`C)_%8H2oT{EKcAZqRC!a~)qg;M`?tM!-UN-j-th%^P|k`^hdX$;?u5 zJo%OlNY4s@bASCZKh4hi6orZ!#RQj!=J=Caf`XWwG&3oNF+XOu0R!;oK&Ks%fph7E zV^xTUK1TXbSbNl=D@l3R%F~S*I(p3J<`%vh7kLoPTe%E_shW_?@8z|+F<;a*G*6t7 zN@n+zGp`yZdCTl2^uxl8ywHBn26teJVy4E08+6GX?cMy*_^8s^{v}$5n~mAX()1H< zCHAxOPhv-@0@A}bc_TjK$=%nm8G@ASjtInYz7-gNHogLmy{S&w4;6635?!tp@T=cd z8lRai+X_jrjT~9WBPY#VYsTg)M%(U0*$)|=lJ63oCH*6ephWqn~J^lID5EjnnU#5yB^*Svj@!iWG9Oa)*lLZ&~T|y zpxm{P%aHnC1?wOny(X2h*fLAB`LRg>w^X*(c)kv5tzr+|ady6ZAN+Y79ZLOKrA(@9 z9d(ai_X*}6u^AH+fkpN@iMDfALIsOSRhnrf^v6~%+4=9#ck85lWVwBFm9E<0vJpe8 zDrW9|w2KRcVde|;7^R$H{dV8s8n!ZJ%bewUCnsogIlRUF5!z9f6`~@6FX%L+F{13o z4&j#d105ZTWw++X@D{mugl}(zz4rfZwL%7*dy>dXEk8H)9+o!Q2W0gpd9fcZ$k{d&ePz=P@QS}le z*)$61zezd#mYbTIwsU}^B4Yajh(fFvErsw@J5P#o-Z_qEvJl1q6n(uW<+bO`8V_&X zCMhZD=Gab)eEWhTDHXVQYJI)(0WJJjr%hr`bQ3`(dYOu{I|SktYj?AtH7e0R9OXxm zU!wvOl%X~2NE$DOsv2{MxCV~%$=iDb_{Zn=JR&-+?Lxa1@Hn4HhlO{;b3lhMgXOlp>|JX?P>mxgtd=5=nm zOZ4$BBT9P9t!9ayPciDGD~Icpw`Wwg)FaB5dJ2R3#u4uSky1_>ALwJrh2dLz-QQB#2$6+ zT@AkJla+0U-RudOdnzbI#I-59Vy;*(E<#CPd3KV}ijH3bLqoQ@_R+B;tXW-*gi$Wf zDFhZkL@eySk0VtBZJs8FO6{J1S?LjdJRGkrR>mX|3u(5(Yx%YI_22H>(l&jmf*qQi znyG23qm7Qv;&1Hp{b!GRWvhf^mj#N&y;|jyHG^cAB0=#1d*1%OK5qdV&Bedt!_pNf z5w9I$>U%W`EDBz0{c(I6IrN^W%c_uRjNS53W|C0bmB=coD4#hIcI+}@ZJpO1vH679 z7{51*=Oid9Qk(00F4MYI+3J!OIsM>o>~a9ZCdV*>yXI7Thz{JXdfq5{djcuW>E^K} z7nYfJhAfTPk&Apkq$KkGYJ}U_ITqm)4eFV(>ak6Ft3q$>r57rpUh;(#fndr`n@AX4 zqgBeM*%VLb`%k*u$w7S^E$d`OQpFQG;jLtCTE$ZV;yGfz`|vb$!uu@lB0&za)c0Oj z@7_TL+}#UWWd{e`oJVFBaVK4*2L#06(JLtkjpv0&<6n_YuHxX298#qP{f>k-NdxY$o&pG3RpM1WbSEFXQJ_ZImyX^Cz3B&RVzO-#RVZf1JCHGZU))-oBR+Y60@~JPB%SJdQXu$w#L2N%nq#cbg`wdB~`P934H@ z95xzF^g16=fHb^>+J#B|1f1-ScZn#CoLL3L-RI*S+kHlJcT$m>VQO4?>Xz#xYB_$N;SgM^bp~c$~gv?YRLM|@#9CvY>h}cQx=iNA1`tvD?FD9Ct4oSmtfs23J6kM)5AwT zkwyVGT)NV`06F(^+onHldIEmLjbr`LS*$~2`q|%iYRdO>Lgh)sM4g4l8}3J!^n?mW$-fP;~S? z`by8&q2b}C?0}9G34g>UzD4tT9j)?Z6^(*4J`dN_$~SOd!u2pyuo#TG%n=uL7?iRX z0_iW^cfG;4ki;Hjpe5rAs6~B&jfUwUZy?(|y9Km&s0pHO+$4%C!V?JFZR&b&@Gf(1geI$2IG(|^?DUWt0Ex6xa$p4Sp8)}>Lv_w~Lk zWGquGWu=5C8eI$;wW#%E+3a3UDv;_66s&!`4cWsK<0(YjY@UQf8g$S8H3QLQLa_jD zRA&@&-IpkogU_?#)pfvZz?aaS@sSu+{LqTLu^ zDTbY{k!@qTG^Yvqo&r*2+-xq-o_}ijnVj{0+|VzpP>u_Pc<>&Pa;?SL@4P>hi-9{( z96_CgwNoSwy~^XDQK4z9--Aj!^ZURT%2^IgQ_PWl;f{W%r&`lRY?W@Wq0Vri)q;hc z`ARRtJ8tV%GRh#YQo5E;V1H@Wc(H6e+`{LB?pFj#75HVL+N;g$zPi&hXoF5 zP^+|gTwf$6Iek8v9B|&?_P^0*5M5}yVE6?-zws?&S#gpn6GoUNs#kX|ty?@b1MjYv zgrq|}Gm@uJaq}LEbf`E@Jl$Qjz{=J+JuOGj{Bj=HJcrpR!j1+b&J3wf^llCk{oI~} z9;sC7;@;B$2GIQlDQ}>n%kyscjF$+&QdjbhyhhQ{+eStJTAa-PY3S6rLaZ< z(Dofp(plbH_K}rU;5+<^iuwlIp&ec#poskUy_d20HdXneZ6 ze|=y^faZX=Sn_4S4GmC6P2iDapGPZkx!mNpu%xbY+p;k{QlH!%<*Wa*J3*QeiEzA{ z2BCTye%Ypp0>D;W8jSdNVt8SGMNFbElShF8u!zuIA<1`Ng>lx-D4P02Zn+)5wrB;{ z^j=flR7&coAtKxna;`ZX6;&b1W06|;abOm8LgqU=eRrA4tZ~ace4IZx)6_?|$T0Uk z=k+8|h{8D$U;lLaiJ#{*J}kz5@L{h4;c!Y*R)AZ?ee*BwU=%9Od&1h1n*7Jlnqz>5 z$xlW^0%7YnCgVqc&ov{I8whBTBo=d@05Fi;<5mT@`m8^$ z_7~xawx1U_Bp|M56J7b6-$~m)71%JwJY{Pepx<40=w3RP`vTP@ns%R}Vvmy=pX; zo3PIA1JUOzptxxd-_ZO|F2F-5tMyHQXu^B&wQ|J=?TV$?o_RqJs^k?VTW<~zn}u+m zX?3e1R!nSoELG%};v|c^CJfI@uMhh<_m%1l$^(Fs*7D^|HB|`Z3aMJD&7fHf2hLvt z`SylM-yg*)B^WLJq9)C1eC{hVY-K<4WLf!3rt9uO0oP;TSRVub1LVHpfcJ%F^TKRt zh@_1Ieg)~IFWyOR7rESwG9_4+m?EinB3VUfOgfj6eX1aH=`WN$P`663sqJ0<&kwce z5GM@T7@p_6aOakS6)aG#s8qKI$gRB~I3TOYhc|Rh2WLtDX|7ROouzRlfFXUFwxB>Y zBe%LDSk#Ncs0?|I`d?ok&p0HxP~a!a&$HG`5{9xe8Nb!R^mMmyOJIa`9A!ypFDSj` z$GQp89=sjN@1`w0JfgK1YRZo*ygZnnPZl&i%M=Q`+HS9y!qhDO2*X*#Vas=I{eo%; z_^auIRpl!h=&V*rGf=+L_I~8|yEF=whDQDI7Royl2@UYpXDN{wo8J=KJ`XD&`}iIu zyd^BJ!*|h1r?h?wq~~CEuicorAj;d8vQ~rkArHFY*BUT;g0$PnHm>&a^@oFoyTTB2 z$AgV*kUZ0ym_%{H42=SZ*`%i6htj@V4?A{E_B}cfW~K`CKNB$HEXY#;#QRXF=6YWrC6F7BlggT{x42i1NjLv=oul|hoQ(<;K~ zc07G+eR&@0gN#ng_^=0b>uRvxdv4;Jwz;OsP_Hw*^Vp0c016-9(H!JFRZUzD*1O;59!mS?Tj)(L?efVIBBu;EK5fxSgj<=#>MDt+E@_=on}b2 z#ywC|-msb?+?;Qb3mgJ!xAMVB%1Q@cm34Fu4!$rfA7)WiG?~>A zs^~mHKJAET#=EtGPzaAii?(r4`3P=}i<4zfR!F6m&(*Ppq2aFZYc+sQruWSJD}@L) zu$=#07*M&+8{Tjg=+npV)GzG!sQ9pj5itH2r|rUl#11`O8_@ts*GBaAZKs9XyoaCa z(d=nTVLb>lfe8P%=HEhSQ4%`cL3a@Acomc`YqfmZfGjbSzeIYXDJ>bFis}uZ)c%?N zC@t2Fj;?TZlqS}p;VAFBQ(?uOU!u1D-m1%eb56OYjF;{dMno*Ji2FjKG_vB!8GSQ< zo+?>1LK{=bkX&e|Nbn!K-^Vv4HNi<{_jw7)+m(=d%h@Wjiu>8QB8V~g1rW4$?w-|%Ys0%-YWd$sSY^m;N3iOXbMRT& zXF-vxCZYwu!uT3^cIIEMw8+`;J4Y^ev76OsQirQ9TG0_98}80krFO}7JOBF=*V=36 z*G4cQoJ-Dwk-{Hl(uu=6FV1bLuaH-DQlu&h};{*8le--SNMT+ z-<;PlS6H!>Q<#kHwbaGEy2)2z-f<0+VJ3)@`S&QfsYC4|t7R?jiNQ74svc3l>wllO zKC$Zg`AW4(4d>k(#y-F5Gg4oYj9KWEUe+xAhqP7I-wqpPyH(gr+OOXH8{7CVLPPH? zs;umPr}-nXUZWcS!^=m>m;QeY-2eaJHWVGbH^lhA)DR=-@I3uZ>b2d4DvDat{{bgR z;>=S!Fa^ods+z0UK{h6A>ov<#42%HeL9^6r2LQ1L(E5r03vzH#?D~Qz`-h*FU98KV z|LEGs^C=%-(yQ477_!^Z&)XTdJy|RM%_@h4tE?ga=LKxG+(Ty{C$iw7v- zYGtQ8C-uleaMt#vPv}!R$H{snwtFgead0pX|?ua zLaJeDeNBG;rcqVL&=P)$dZ`K{GYgAd#op!pTe;Yuf;MXIpRbnq>;M=qRx?zz5k}>e zt5ub)RD85Q`Iai+EN|Af{UsQ-z$nvj_=?{GMr=!vI zV8hC-SX9$wpI5JaLb-$l`+v>f(L*7{x5YXBNG>P|;(UA_Grv`n7?eAFCn0w>hjlgp z^LIYtl_Q#Kv)|^gDE^Gm+X`{Og>hgB%5!ZzS$N((a$?n~`8B7U6L>E>o-GQ1|1`_| zq4#^)0lNn!NyPwL7%i+jE<*t5A>Z6woH?-+{Lzk&fIToqo=SdQP} zU~DbPv2LytI6yEkX>glGaq{zI102v{(sdem_WXxN8bE2RA-L_1iGrZy2+r`gaHP~7 zuxHqIvblM%VKI1b6@VO_u8Mb@bq}p+`rL4SpiSJ~-E}tVTJd&c(`5PQK+AvYOo!+! zRm={Lj3LX-cP+n}EAo>tETHN(u)jDX4f){XN7D&ctaj~!1?;hqTWgt$01 zMn+?b@8bYSSs;0yQbKXo;QsXGk?fFYV}%2J_uV_>2e?8Pb%+0`U&lHs=;3J^=rvrf z0&f)S)LOOQ&4bySEV=+-E#hnMQ~%qi>pr69tj{@I((mxvU!HxBmrPbOn@@BXHn+B# z)&mQ8ojD%wIRe8U`~i-WIPB`n<5&794Li-3t-@4*_mn^`*6k4gd9lTUMXKwcZ$Su8 zYnSh1DZs0YIc>dN_5F-A+RRhDxn)Y}_m~ze{lRSvEF^dru5$^Oaxs#xwpL-zefw9t zP;Jov31;-tV0w>My&WQz!R>MQ9EyxZC0b(%Spp<0%0Hc?x42%HN}oL>P(W&(POnI& ztMM+@0n8Dk9W>t66CoE%9-JZ!I+cR6hXFgb+K&9I#=O(YAD>auj)%3j0a(TaJ;2&NaAkb!dD% z@V+qBaxBw(`J~ax!h!>Kz~C9jX)&}tmL=>Bc3k*lvg5?^a(NmEey9orHYgDyoQafF zw^@SNuDcLm5&>Rqj^AT-&qmuV9V~YCpIktP@0#~11O{!sZQjeK5VgEY5PwwB)C?GJ zzxBK6+j|?1d0RhTV=)Y1$102ho_AoU9ha1JoSZm^c55B}9ac5lM^(*X1S~%a47Tox zvIDm#KPY93GzUOhULIlM6UKK3t*0Gf-vGL~xXUKOkVLUkRwd0Bx5xWFFfk_5={#sP zfPsg{ZTJTpa8Ob5+2U`re5D&Qe|h~xhxbXoSSb*8WCTq9iUEl75x2d6k*EWWUN}fm zR{gf(l9Q9u1Xj+Vt*5Kr16f^^{7y^i1+z=#4c?b~?PrhZR^!B zaZ#lO7{8ZQuQ$w_fR4f`qI z7cACz^V}Ytn+;Ge{NZ$+`P<~W zP3ZZemgSY@7rydH?Iu&bpRh$AC-%9_9&S&?o!2`DD19HtgrPwy`WM&VKT!H#Wy7HU zR~sLohR4t@7*zPQ#xD8_df*uWf$@omkk4OtyNz#guidR>H03kocy$Y#5T}#jr1T+6 z0v{~^W;ozUOiX<7yU)?LlS-YL$!_;?VPY9#p$t5BVU>8=JRPGDJ|u%bF6{6PX4$qV zw5Xpds9<0g8iHK|IxrL>=6_=XpN@Wv>iy&Uxf27+6b=@1?TxfB@^68ism9D|YG@GP z6I;6p)}@n*)m|TG`#euTbNDF*vkZ*6)9f73(^R#!?^bxmxyM;m#>H+~%~yfF`HxbiF>D#pxUzgxY*-zd0XHn=Y>2tAj%PwLA}|3PFJ7dwJwB9s4#H z`LGH9)Mnfj!jKbi2o3AY38SbZRa&S1VDcoYJ zI0Y-AIUL|Nzr~9dy$$BW_}AeT6E$ah;75`;Rx$zmtI7}RKlB)xSpg&+v4Dxl53OfU zn#~s9d8i|#1rY(kxVf$_nGf@HaSp_hE3*EW^gYSWff$eUduqm>@0fM?-PPrITUF5i z0K+2VJ9jAIMLZtdlO$Vj&u~L0MR!slP{CAhvA4nl*7Yc9m-`JVYcE$@dDDCD&qaIQ zFbD{zcIO8!t?xMphY)%EOl8RmY4P+%2Js<5g>0Vzi2ZoJ__PZixoxwn3`*UZ zODLB>3Mtld^2-LK5%zL?vkhpn{U6ef%UKgYZnxqYRbK#(irMJK323#SZ*yuDw(fsA z1U(>nXE$yr>cFq(smcEXc6~HlLC$B(`%Ee98GFAlz;Q<~N(ewc&#@#Bot{qv$HiVU z)#+g{@#oVmW%2VO{TlO6j*gDz<_L|uS)kRhi?*#XVJRsovG$+G(XSH0chJ7OVa$jd zxKNI#XJ7DPqu4vVK|ra&hCf4ETKcs0MeOtefZw{It4>aSYs!2k=W>15h$rclY*?Wg zV-caO-pu|lhv6+ePa{VLzLPjHD|o4*c_8|_zH+;V`0V!mlT$%gp)TYAG7*t$vUD{vY5V5ztSvpvlG^m4l!o?oW3P|HD4{jCCr%U!{kU;Yy?G*coN{I3JE z{-r}xP5%)r@S=6(9O$=^pt8R3?SH~tnSpgzP%!)0aGW9;F%Dkktd*M`Kwf%2+ux$C zeI6hfX5n2zYVbdv{9qJ#8Hf8;4&$PCj(TC$8F1Uo@rwhCtML zpZe=S)Ybrv$kf!-Pvshl<;J%%-otE-j}5kT{K2e#_f#C!EXZS*KQf5??yt1~!4LCn zuVacv2vh(_7E6oXRf-2ZBIG9m9Lk)4O-d|Lx7}Fays_;G#~h)sbazf7Ai~FA0iCvM zH@$Sgat&SpO0QKm<7R#}|>iJjHH*mATx2%T82F-zM&ww}(T7Fib_c_O% zEhn+@;q96mc)}WB^#b8n1ZBWJ;Lq?}n)hZyT4{c^FF&`L_GO_HId%wg-w%b*2OUfo zJv|2=7 zeP<=YKQcD6>@p5TVce-9YU97a$Zk>3WGReLd_LjDl%KOeF z((1ooAUr@q4XH*~6KlSxm%S&QO^)H^bH8tCE>bRbJm3J9Y1r|)v z($~-!O@9?)Z<(|RvMhR#=w$_>Sr(SEV8`~zbkD;XCLW&JjY@Eb%s_9Zc1}5e; z%vRr?0G}K|=kHPoJ!zZ8oF+07yI#j}s}nf%m8C=1ufZh4r+ILZQZ_AK6wtR`kvg^$ zJ{7a>R`Iu%C|T}jU3vKpt7e`_9)AV}*NyHGSrd`U-PW7U+A*u@hJpjSoPVQqL#QjI z*g*(hw;iFhlfFFzpPNqeJ{D*t1DbNrp{&6iYS`eN_)0>e zm8!fJJ~dASPEwkk)m2nB=FfO4HF7MIyY)+XrSaZzM0*Zvd+uO+gp{S7F)rXF+_--2 z3iB8<4D2w=@sDGEr0@9JzHzC+VeN~$Iyu>!6W8W)OE*u)fKNJwYp>o;YLW~4B$Elt z>myxUUMOb?72XOYlKqoA()tFRoM{Z49=pfhe#9=#O47^l4j)Aqy zpGAAcmFRNFZXN4CNKxQ7pCEE>q#o59RfiXNH9l%NvYTWhV1@38TbD;WpuMtf;I(MrpU4s0H#>-xgTZoFdU89Me!{@8jykf zqUrkW_G2B!e*`gU6Pq>=4n%zcIZ-=a(g+CDfP{n~WvT>S{*O|bg`(M{=WN&{% z`kGGLd6}XoJVB%>km@b_+q^uJg8uk~_;^(y2N~LBd^@=R0~eqCu~Rys$o&wQ5Zlld z=hMZAJbE>Q@%QOh zFfT;Dj=;6KP4D)q;|aMm8A#*U)#~by(GbR*2JR>B|6BmN^=2OqHU@G?McnUJA+6ul z)Et|ilz?2Cy7s<7R8>b5*%@V2d^p6?PSc0R&JLJ04h9EHN-py|T~ucByjV8tCnjZG z_kh-TUBnlpd&V#iT}b%c3_Yno03hTmgIc?Ro2=!=8ck!a48SSVD9ribRL~dHG^bjs zek*qNO@ey8Bc7Pg2R1hf^m3VxS^$iu;`DSNVl!2)4+`8=d@200Q zS;4}KnwoL(`-#?#pDtTpS_>tKvP9j|ap&KHRy;s5=J<^miKrU)8?8=@q*pz<`igul z?+MiNF(os`$^mT$3om~&hRw!C=N2=O^mc`~FTS^iq4 z_fTId37VI?P>@dO>RUo?6ip} zWJrh+ocRT6MjKI4ab*0a_;s&;Y}jn*GypidHgNtS&%cF;VFZ{$E8Ku6&+ksS#*>#? z5Ybu0jBd|Axd6q?7phKU@swddfmsKI=r(D2@-6$v;o)I;-N_Ucp}CbsT$|lzfJ_n>r)Xl*)qj070TSHmif}P~-W?uyrteRl zwi5f6k#xDtAwKH0zOg-=8fb7|pfYaCd6f5|klRMiM~Kk$6q?0K*7#%nn@Z-0GZxI` zS-<+eP*aNrb*1{t9$#62tg(JOiiuBhUEK^m{yW~S39dGW%QZpK1hB8G@i%t+Y!Q#8 z`&z5H$7ix38nJkr8CyQF$XY?6tqBoXVUN{JC;O9Ql|#-aD)0-x?Xj0+Ak)^HaS?%+ zRwt+MBjOTvRsy6RujUoXRsudGX5{mLvi6UBD|MSAe_CwXo|KxnYl)D)x6epWYkS$Fw{Vqy#I zH%Oa#e0Fi1_EiU4oDiS^u|GBg7;lK)<{WeEr*l}I8)g0N z{^pQ)p2Xo#4AEGX;Wdk4wm9ep%5)2KTvu%afEPvse;Bq58KsrO`AT$B`GTwT!N~*{ z)(#sB&t+=PkAja|TR(|qwOHc-M$AIBN6a0Yb=yX7obhYSkABX#V<*H>nzSvyz|W6g zwsyE;{f+9=| zVX`6c+&s1|WLu*o@QVXkQ3Jn<{k)}#NKu;{ufMUtJ?eC|UJsjGmqD5Q1BYdm*po9M z?v++O-#kc~-ax50&$Q_sUURN6lr-*UbR!g*nT6nvm2?0OLf)GzOFMaDt*#6xiM#sx z9!t*nY@KG~&DJrM=J6E2S%HZ5D`sw=oFt`X$IJhbjrnn>K0H@(jysdh&kfna%rtOs zESPU3&du65KE}f(3A^7mw?cPUK*dDw0bAiIk$lGna6k`-waWbEMlmL}vL9J(HgFSv zQuDlp-vXEk?EB*rDG!hRI|`|I90oF68AATyRXJ;eTLTjvGHhR0kLqUBuO_L(@YdP~ za=WBmp2eD+oFpa1iUyF+T{0;Au{3ZV>;hgq=!S~b0y)SZp53XM{XS*lS+zg5iOlG34##J;8gqw9 z(M+76P_W}g{rZzD5gUN$&n?d?0*psBG_)k8CrzhVl~Y5aQ}Ie>e*LP>of<_ALwTW< zv>?8mzMx}N-S>2H^`s0K$YT74Vlk-`JEt|i>dzHiTgxNo2_?-*%+@s242SLl$XK)o zWca>*4>fglW1qI;sd_S6LznE|)&71@>_F@O9{OfTn)Tc|`d z-qey;ySp7_1_vp6yl#BXhmCms(vJ3fdSvq-vj1czS(p*m8WEApGX6fC%N7zk?T#&4 z!N~D)GYp6rUDQ7L-b6O*bLkRpe2(j1$gBcOR7IyY3gid5St*%4AYOI|H#r$?_Mvyr z#yd?|^R5M6?f|s`C+*vaT<~<$9HO)#k(FEB8|NCiUk;C&u=o}g)06bINqLb!PS{Gf z6}pWC%E+XO1WH#5L>!NYwJ#e6!TbFNaB;j?YV0UUIEA^WzW1*A8*J`j`FB5(_bHLt zJhP&;E(=1DxB8|(X<8O-;31ythm6rYMw#jB8z4R?=dD#*Q5ob=c|!ZI?cChlL_|c- zr_S^>x&91}4-?#)*~+6mW1^s}0NvRwgQ^BggFC;;UD!ET*!Xz0!;lC@L+%7p#tS>8% z&6aiFyqkOuMS6?;z066Fa-mhAh^~sGT{O>4abhjePGoBN!>vkAdHLr&F>RFt%P-n0 z>%h;>$rsI&4yI;mL(@kslsy~A<*gNvyE%~f%d2xK>tD9Ldg*SiHMmY&zKAtG0?ekJ4>$8jhajH7kBOusOQW_&paOI%HGt~ilv7Kjnka6+I;=esUiWrHDyE8?B@XbUx z)?UUBPFGgr0_+M8%+9NhNvnbJ%3Qc}CY|Q@Er}?`&Z|JunK(RwT*#W+iV(i+yWe|U zQBp1dk_Vv|6RMXUXaeBFbeSUO^WoPaVYI)jTuz70ThR5KvwOh*vP(4w92PyvbMEI1k2Li^NQ{my*&*%xaxnwpRAcpP#`McumpS>houI9xOj1O!Or^kgHzVVg%)nQKcK?yKnNjc6Hk`#9D(%X z2$_Nk_K0A*BQniRO>#1FCe-IsvReKx<+VpC$cpa;)1OVlT>kwnDeBsywIsCmf_Ox3 zyhoa%7*$&$?3$W^K!RohQg^QG>l-onrI;jeLYx}sBv3L{nlOpBxnMw10(Gt0VHij> zErB9WmMS7lOc|`K(msb@a zmdg)5hgGrsXW_yNBCz{F^8{D$up8n8anSB(*h2j6XYfmJwb_Z8hgNmAk|5pw>AgMf zwChr=jujrBX>9h8^I`A#E}N;`%6nVD(0^QCuhmH-5oi6$FcFAMpr{ikJ{R3d{4{D! zo;x)%^7(Xz&T!&nKPlun-3p;|lFd{_AdhNkJCqekla7xMF$zD^+6tIWIrviSEM%Vt zM&IvzZs)_NQHVxE4+#lH#vrY6gYnl%x)9p0<-LtAmHRSVphcD3s4V8^)VWTxZ@Ad( zZ>OGQdAHhdA_SxwwA-i=7s<@SZ=p*i_H9POzh1sU#m-=szlw{Z=aZuwKT^LBXFdP= z=?e~CWMgJjEc!UzlZQ_}Af5}@rfyVpOuiZMeL7V}{6(F#O$gO&Ainu2)L?Q{Z((VP zvaMk+)lt0F(_XqSBzC&F)@p)mrA<}=AMc&Pu)IQ_Fq5C3Ujzmien(|QF)CIe-K*ja zIy1x`IC3FH2CzS8kxaL%5!wRpczGEG5wd3X#xmt#iW38sc_tg1f2vpgVZLuqj`k7| zTy09&yI5|-3OJA$Zkxso;uz#~vOHf{M~qnSBe4!^?4zU6$Vmlju#O>Sk4a{x&Ri}R z>Y3Ae4YqOQ*MKzwirEnW?*>{*q1TANf64gy<@-L8r3e<~1Vl7}5i#+*OAS_*-0wAD zj@;+!Fm9D}8@`VLT=8PYm)=brijQOK&{GA@k6ydDHR9*z-*1rI7#kmk?GzJi>W#c! zl_@=mrW0ocQnIqCG(M1tp4+`8z5|;UM^|2~-r-7TdBQ5#6sdpH>%JgIo8eSedBsot zl&1?NJWAhC#I&B9^hdOQa!6TX`#gQjtUY%Te+g1DXxSuhHa)<$hRe@Bm@TGKN)oTq z-lf6`>iy{Ly$Jk8^W_f&1Rq}jQg(L#r-$uxGiz>7D4vto83v@Ta^_JtFT9(2-|0}! z>Vw|X7JNLpkeA%r^uu0@xJp^B!%U^LeS=qtZGE?yNUZ`9d+v55_JTC(dtV_Svh2XG z6x05T6HKfm_mhECo3RqLSP4z!6J!oYIzKGVKQ0=b0A2TTf7GZT(3V}v0d{i;{p&E> zf|(!rebcn3xHOfQP`zbwo%O!wu~}9&OwfiFJ9eL_2fZUFl-v7mLNZy~9g0CNqCT=n zno8l0MC00D%o#zTK#2a(>Zxjqn@Gz~I9%-L^JL|$zy4b?=Ko;tEu-RSqdngwgpfe+ z1eYLz;K3b&6EwKHySoK<2<{$&ySux)LvU*v8mFOQiuXNdW)x|wtves4K9O3T zqPnPOKYRcFTTrB16acT5)AQuBFV-+Qw4w@!qk592ekC z=1wS38bnIbVKK_Fqs25m;|%_56Zk8tx38x{o7mR0Rp(OVy(X+1pWuyr%IKMeQ~%_y zKSXMLRsBcH(#pz0I(>OmU*~f6Zm@ze0P?Tam!McAERXMD6rUOzlD-=Og=c?j9K+Ag zE?a%(g-nlW)oNdO(Ezrc;>3fU-Q_o|+&3Wv%c0ACW>C^AE zw6*EiKl;2F>KuC4*V1iC^5H)FSc9+6qF}6D305co*E2VU3}K7*TTvBp@c{l z$JCPek@c}OYJqMw?1#`CJM^RK`%+Yy{4F7);u0HlCeEI@c2mizJk*w{pYAgeE&1_pOb^sUtduCd$OeS zMe_gs;PbQiKjiTY?x2jnLPuAN3>hpOETnkpD-hICIw}PWsqq@VH|E{hJWO#t)ZZ^( z+w!=S;Bae=y2#>iI|Al-1|EMk2fdbMsEFSo&&jd0v?Ou|eIL!K{oOmw0_5FiM2E1fci+yKvKZy(3mRA~HIF4jT-c+{+#U|lb8>+HSOh8^&FvQ4j)4n) zf4o$%0g#|TQG3{Zd`QRvbm^_Iu%R>*6_2#vAc(O*hM=Ofw3IJjiXu+J2^;YadKVZF zgy^)rb#%0Stm3r;0~$scZa$=gN}ET?Iam})(aRAMS7%06hKl~kXg~e@_Vn{79z`z(Q8%dfAmyXpE?1{T(OgbE8z zJMQEOT#uTt8FZR2k0)pX@Ozf};PsAzoP0h>GO)0XDmU^q9LP|T=C0sBFCL6n{(gW? zv$D39S6qCwnleM4)PD* z7kHtJb9DffvFX$mqbzu6U7apjAm`@bit6w{+zmIHP;++G*T1zh38M;Tf$hi9I&puC{#WN3ZYV z1ym1Li3790_~|v$(;d|+f8@UJwGj)$=N4=LC(+aEREsMp8968E`9)u2>3SecW%#X@H5}*@Q}fyK|`kd60~i-_k!FB@^!9VngK0z z>h)nKi|@$LML4*)@O&m0CVW7>Lx*NX$xOVw$N~)A%=q|eMK@!y-;F|~>aAcc(k<5% zIskILTanN*ZFr2H1IA1M^CfDC+!g5?bgB%|1&gp4>BOQ5ZOr6=FynN+<_4%>MiU_U z)_k97_BIp!;S*o0?IS4HZ9|L0<Z1`w}zNDlqFdowYt<@l{41+SeSuiPO>FDV0pchY8YY!^>x7KeJ z`YL+-V2{TW$MX;1!(G;gsFvO)68GMp8AQ1dBYxU%FQR$3c26Q93I!8q7pR@hpZX%p zBI__mVwW=Lql9hSZBEOxk{G5;S>NiwBYhrrb@#0+4BB^m`k7h znszU!u*FSE((dCHm&?gi$pjc+Ka5tER>nMMmEVMg=fiM;#a-e*c(jzH{V~{~x~Ai9 z3zP9<=!2?&z~j#@A%7vk(e>NqaxY+QYTXQ+`tgCy+3o|Am96As>B)rez$FtGS8A*} zIzEO#B$cmR#Kg?>^;bqhs2N)R-|#(8SYUGC#n~kberee9KtD-QL8~5P0-)fh-x>ESYb7Xj6w(WWi24#mchDj?qJ_%e7V&@f7?PONzWr=3U&jeB%G9(%&1k0>G2fYLszu{X+>s34xq~%ISQEu z(LxhaJA1u?3GLicjfoV6XzFEBR=FnM+HbYXaVw2jdY$gC?bsUciHD?Q7i{Ke4Y~V< z#xsGN(46cG?kPT{Y7PAK(U%?n6q-YX>T9EWDA*(&K3xlED@uWm&6nKzPH`+6qM_@Z&YD7Y;4awY=LBc^k|4 za}l3YyQ`aCXg|EV8l%ZbAd3bCIsDhKBOxrK*|ATC$Nu0&(89wt64pmwb}Xq+q@>3^ zfV(nyWN2hedQXi)qa^XL(z5aM7R-w-^9e8JAO)uDY8bCb?PY(Bv?VFEcXE;^PX(b8 z`sM$vkNgE2I@*VTL*#{QAi~`AKo_J*7(T*m@o?2_cGV`BR|N996G%UX9>%B9qga{E--PvTw+rKKFcS7Io z7|;7(iq3>g8LzR+t(~m3VWK;gu$^eGtva79YWKv8Kr7v@Y-o^|3}V{a{PUxe*AxI_ znDW&26ctzF0C+>qH=sz`o5!}6Kqw6@?Y9M?k$Ecyt=3Nm3KfF>s=%`e33Iq&WRt*_ zK$;#gc2RRTmv9yS9T&&UC4ojB92l6Cc(~;(6)JxNkn(5K&$|vMC2V+7{-7SM?4O*Y zvhg@rtd@E{8@i1W73Cis_3hhYu2BiLwj12g`p)l>kU)wxX2PXRoAYVVCFw=8`e%ozGHcxvlRV@eVDX0{93mpX~gU zTX#~jqAhSNB=)OSKirq9=ckKl{Oqe~$?0uE^2C~ZT?UAQXog4nO6(#lUSkEL7g?*` zAP9!YXgRr@Wb!$kzqOB5KLK{20h#1WV7Ld_Cjmnx-oiq~FXQCt=EV5HArL@P)Hs7R zxB%o9XqM4!IWr@#LJP}*%7^N-yUmEA$}$9$z7>dZ*SZ>!1)^DT9hZYev#od!c$ij2 zP~W{?*Uim@JB}Q35YTIX^^X=pTzE{t-57^X2zK@YScYaFX8P4%~rX-tcQqWV<%xmw6aOW9gz%U&{~qV zmK05#lHdY?g$|$x^10q$9a-pM$)^r(rCJ83rmQL8s3|M|u{zKX4~lM;R09V2i5Iza zyyW$`fgQ9p6f>j^$y6-ZBSK%F6S_pYZ9i#uh8Yl3MY1c_X$6V50KO40%1ls{aEAY) z`w8gUBjZVt$kArqk(7w{rz6u#Ojx+#%kv8;E>@;caG)}&)2`uVsAfh-T+z(d3gARR zLbJL#vJwd-;tH>N^|F-R1#`);@DkaQDa_z8>)bbDKTvtj8!NK-V>_Jy4GxxYz_(%N$0XFjhZpgOBG?ixrU65jI=p!daiL}hlC7D{7x_0-V!#-@a%gx&pJWo2bV;7wv@ zeBRWA9peNq@d@yh6;)N0l_^$N)6>&eSJxEc;&9+F(Wl0wC|qo@fa@IWZ6ELJXlr|4 zh;?Mk7qQ?)*3?LveGdgZXUCc?-sd}KrvhS44N~YwJvp04$-2;x!C}V#w_xea1z3Rx z&hx+3982b{;^HtE)8jHSVq$l(%MclC$P@ z9865q_S3(==jZ2-B+`!z4`Tp#Tl&W5`P*BYMDxN*TUx}FloEhQS)4opu`emhOH)~$ zdKED@l}%DsvShm#5QnrHP4z@Y`+>5$KE|i}ow#-w6blOr5Onhk3c|O8qj4VRS8QlL zAf`zDnKM7db2CxT=*U$DTy5-$tY^w*t^5yRvo-;u%_6EHn~v-Aw2Jo2?dm>f`#UXz z{&x)-*gEcZQ>fu+JNW%97W40?PICjp-ETlwi+Ry+;Ng++`QC3oohGP^XYk^0MGpMk1#(W|Nl9}x5Ey&`V0WB7m&u#)3*bLLe!SIg@^F;` z!h`6Ftm^7>K#56vTK*XWMCgcb-`CdFU0g%dI9z1xr;Q+aZ_~2e9(MuZsMi@}Xkh}( z9c%F%%JqBTfr)|5XUhG7gdLz`!tlqDSo-7gIut|{Tvltp;NSwF0Q9zs4DjhZPsdak z+ug#GyL!O3P7X9w^p6fFM>Fq$ZvPIR4b!pHKzh)-h0Sugyn3q@EFvOn3k#a4!}UDw zM;NrE;_Ni)g!|VyA)jaOI2+FdZ^{YR zJg%c6k0V?N2?^m|`ByW|F!C`o_nc9M3j3ioI4oG>KWRR@cFCaj6jXiB@We78BC}er z)jQ0+OF)DtHIJyRsY!F+ka`cK7}y>sO4|XMJMXoj(7Wgi4o+S=Cx~Cu;h~uUg_K}& zBm{#F6Y3t>Fb9B{f5N%P+62vGcCk>!zBcuFV495DT1{J9hocF-J2@s^D#P~WD zjXbd1tfyyYM%iKuR5WDb!2Xa^+HBNI13^u*$k*p}*}(j|JjAT!>L8HT8rbd!hXF50 z%+@xf&A%sB zrg-d{&c_GRfjXyTz6p&_NKA=8#ERc9t;-@4;wM1{6J`G zYMS}8&h7HZ^SUh@>**(>fPerM6;*S9t#$dckqG~Xfj2OxM1rLj^3%ZXf4lwsfDA59 z`-AM;^L91V(3wwNz!xPbFZ|NLQgS)_!R`|s-Nz>=!fPcZGm15bih>1xmwtU^ZD^AI zC6b%UcF)=1mw;i}aqq6L`3(*9KoPWN<4eHNwX}TX99daep#->D~3+9)N=qE z1ZUcx_C)hXCIMIvNapc>Sq4^{QkTr)^0^+}@i@h-NYZ9G<(7gjWLXpdg(bj9ut?d@ z+RZq)bHX=vWzJnU8;((@QAtUd48X)mpUKJ1Ti5B`UgYD)18&_P1?3P|H|XpTSum>J zr)D=PfRfISr3HurJy;%rN*vw`5D2J>DVaOEyV%R(b->y-;lCJNy}M|K?pe?t0rgS) z;~6xw3S^JgfIXS&klNbXeAa{1J_Gq|k;)~l&_YAWo$Kp~w#%c~zzZPbsoKT%3q#9Z zr~-fr--aT~rdL>Y!HgN66VolYNcJ~%=iIG=5sao<0-tpY0g6ECv=S;CUl zrDydHy|m94v|C7q2nDU_Q6`bShQSsR`Z41Z<8eA>%~`sU{o%`rV2Due0AQE+7mkD! zVoBr!RWAHSzY7-Zpjn9w?|_Uq6l4+O+vz|gK=lOTFZ8w7{|HcXuf8v14n)l!lTi(Q zzkS};4vf41AAQ6r>EhPX(ILRC5Yd$@;U@tq;STH|Q+3NY{n>45Vhi=9C z2BB#Tl;V}kwQjTxDj%0rZQE@X1*-Ji@tYKW>{o5_o;tm=lv9h%qozT7RA9Rve@Q4R zdfC3wMI$VioTD5cw%szOeR@Ec=;9nabSYq$2TYHK=-=0aOpbSaC)A-~+7%$$?~;Xs zvLEq=huDYN|JZzonnChCT< zW3W}|8Hzuw6<2#YiG)t#=!8>wj<&=H}d&tX3s z!bl01DKyv6Z(SdlEvyJMpVUBhr}M3g{YTX~ueL?6 zyqy|0$=I!haz4~$oT~-jw^ukxY>Nn{#ETnsm!Zb+M}0zw{=abw7-6927& zs3ZjMgy$b<$-2qglO-mc4)_F5=+6#jW|PF)`?4tzf?Qf9s?=9*E8)yd=QlciK^|kms^HR+z7%( z{W;q{-$~7wZbL^6j#0eWwI=r!w~5s8&^*MTzB((bF!XcF0WAK^<{Pork2u`buuIO9 z0&ge7O`$EZDxe*~L*;%Pcc6Rmc96X-s_o@FZ0pv+_L|<86BJtK44@|Iy#GE&piB>V zQo}s;L_V+P1iFWxFw4_K%haD-LxAVO{_;5`0Qu%NnZoP74y)T8QJj7l1W!dO^l-=! z;`A1W!@K`mj^L#7JN10q>zg`!89ao}@t0Fd3A=CeUwhrVKB-`?2`=m6EQth|pu>pX zhGx~;IoE2Kt}Nc~bGZiUrh?+|Q6|#z=M+hD0z!@K#dhA^JlGgM_IU*@ak&S(q}RP* zUMQrlk%f5~w(CLX-{NWssZ${`Pn3JJTRlzm(_pL``3X}9eGAD(VSt>rGYq(-H+Ay1 zdad8_^6D(ddY#(@q-te75S}`Foj0^PwA^?xmIu+zY_^|PVj;ArkkZRnBsEz16rc)v zwO(3JmCOa)pg$O_p-yDTrMt_UqH$Bo{@yzuKf%z^H$wGfZlcv_;PvN^bsy2C*~gxoiL#;xL?4^1Y?&i3Qb1%;3z z)BrmKx6Yl=d^*tCIK`ylCexirhD5Ba^1;BmT!4=CNFusWMrk<$rg zgMExeLA6vD*T>R%OuldODrH}Id3=n=i^jUI4sWhwB-7fzubvYI^|-IEI1VbLX-^XB z>+yoE*gWeqYOw1azW3ucg-?K^oj5&rreg||mWOL(t-JLNiL}B$iZ#T0xCUzxqg2&O zS01m(-p@~QODSmwc%iR4bq5vXnrpFfQ${bQAn}qK#OTqF@pPQlFmAElO_61y9&;jd zw^ei8ALNSJPgo9!Kh+EA_qU(?MKAm^%5iluqWTqZIPk)rvshSH!!M$q&Trg|YNwqochB@^Nw3uqycF7CM`gvmUYW<$ zWmNRiVat;izANtP4MG1;hSU?%VYn9*`r?+j_)VBwiPfNRl+3PVNmVOLzdy6mCDVEHE?k>l7)<$}Vtb2n54+1%;?M7~jJQDBf zk0_^ne$c|zM}dR)Rz2M3@V%{qYw@1ihe(G&1n5pk3v#-napF}EzEv(nT4#R`H`XuB zmKaL!6Sl%c`d02c(tkCZh|G412{p@n^+eh-bo`@R*1}Vd(bGI0c-~zf$N%wYwBhQ- zwv`1=bfoUFs}c7KS0kww!-2x?;wgL*%HL$c(@>w49B2CG6JoCcq{YA3Pf!dbv(vrc zTtBN44RLISChu1@E3fm*k1J6L(zPSFRyY7I#G zzl0=0T0iE4(2;M+&}Nn47t)6NxFg%Ei9qkd(>+Ot zFJk*PC&4!?xVMEIo_gE+L)_TmJxbIfcE$9IqEJOCO{C^UgKSwT{vOXGj4@J-mR2YF zH)z+F&<;Yd9$^IYB2a8xljQ+kTB-pQVvCHA^R>H+4G$}*Vp}LvTJsQoz?2;lOun}v zDKX|TGD`&KJAAvCE=Af(7TSkA}6LM1qY?_}oE@;I$V1yIWS5qj@7Ck_1ZY zr|uE4_r#W$$nb;ET%1qjC11oXxGqz&8|B2^kB5KyACIF3Wiz~($zFNIFh(D1O9(GY zpX38OX+YW=TWNAA7}C$uzEYkKjl}sDk3Xrm)7R0Cak;<&f?2X4vOx)3^E?82H!`YC z0<%y1IAAOIe)N(Z)P*yP3j7@bo4-$xY=x6Y=RsgJ5{u$F)4*oHxHPjv8)`;?IxgevN=f(1SQw?Dm3*yJ0d(vZxpR!IA|9y)& zkmhAr>!HdmMe&CD=j2=B4+`=LpPRM@nY>~>0sf|+&vKgJ|Pg2PFdG|M0TK=khD9G z{18kNUV{b;>*+fzh!MvQR@xQW`j${UeMWBLSoaoVhGupYj4~DF0!=V{xOj1{_ktWK%8xJ34vP1!4r3nT=M=Pgru}j{1-q=^e98J;ssb{1=-gYUaA4 zHio+)^(=FB?q+zm4l0I4HgC?gL-qS5;EiB^FYqPKEOYxFHiI$AC#o!w9$Mq0V7PuJ zki+J-?@SsdME)butLW@p@1f~UpYD57QinK6^9&L{T%l~!eUQ;Z;wH&7VD0_*2$;NY zP8!zL+dc2{!xvs-ly{YZ)?m+jZ{_Neq=wSx1EJNB_L?XL5B0oaVdZOJAAL|qbwoA~ z{jHx+Oj}X&F=thh z*YM~y8^oS##ev7ZA^UlFIVryr(wcS?fp_cr7|83S_K1YfC$u;NTjKZ9hc0t9-qqpG zULCRyjvHky8}g(izeGCZ(r(SDe)IXTfUvFSVFlUc?7g6_&4aS*={=g6?z?Rw&4_%u zR>#Sm8j7ikY}uqf-p1Yn5u4t3i~2^K%_hQ1AAXbVt{isdFUG277QRF!Gacf(!d{oS z-Bp^P$XJoh@<%?#SbWtLq~drp)6u5w#VPxE%kzq)ro%x&O)Vucb-5Bb@GquszE6wZ zR#`o}zL3aeMyL%Z)01ya$wt@4HM_^Bb)-|)w!hOaFW)-48^Y(Rl{%fOII|N%Yx70S zZGC-5LPJ@fM(Y^K6IW!Putd0WC!~H?Ge?!89}R)Pg(;-`$ZM~`nWUpRLv zbv@73^H9NFg}>K$OY*gx2Y;I+LHp%`3LOw1??=LpiQY-AJTafA-wM6Ro7K85%|hM# z!|E_&0EA)p+nX6g+VjT=8Kzy>p9#5dPk1S2Ec!U#E_f1jmJe3I)660LY z(x`AMo-Pw~ziG*(yIE?9Kak&?p{#6x$&w4wF7EjG-yPVBf!?ptm5SHmb!2#*pGKk5 z3XX>U^m$sqbXsVc-<#gp2yn~@$(Cs-EFR87BF4IGh$hNp&s&?V@aS;|-ux^(eWpJs_BdDNK06jg2^7mv$r zu%0ii6Moak1`ZM|c<0>JR7Z^4uS-^pxM9mF`HksI zNkb-G!DO^9!FrhTE!fWVRE4NTlNLqsz>N0ndG~?kHE; z>LkZt+q#hE!qK)wo~D;n@tBmgv2zw4hIXRpud+&Vp8oB97)@ze)#MXD*e2~J zx5`xjS5hn0ZCw=1@)s5WRO_IOH?fxvwDHm16Q!_Yt{{9j2p9|DS19M}UBMveutTL6 z5KaYlu~0>rnG)@`YB23kKr=+CR91vYr_FK&2t{04E1t2x+YpsUKdabNmI~M-cemcdn~9XlRH`1n%i>Ea(alh+<&H$p7^7uy`0;mgW~Zd zIKfC~x-g-C8@BwuTf;~V`n!L3tgR4a>C=q(Tpl*lX3 z+M%%0{>+F&hzhSHQn1S48Fq-1uJ_MR!5P+QSULta#Pe}58eT$Vo22c&-Bv0W0!z-@0Mln`-j+b|89A2a&g>==Bgu^@(Y?9g!?yuiA$Xz5wHd&7eZFthX9>p-zaU8 z@-9ZDVf&i>I(3V^RN0D!clU!8SJ@%ns2q2~fTLML6Bg)hZkrObJnRA6i1L%XhqML7 z6l$nTyKYQzud-GK#`$2j8r73tSui(_81NUI=$I9yw&f%Ut(ENhAj=3TB|LrjN=I{yW6ICHEXdO9ebx#JEeWRa`AQ)h%@xrAgWq?9 zNh1z%io1b9F^MQA*^N_QP@T-Wj#AlAg7Db!6E+_DWW{$$iiP%v^Wz~H}__@ z{%las$1?xLWdG?`q^;yH@n-&(mtuLMN}(!%KO}Ne13OvwQ?|GxlP5??T#s*o{uS>h zPei0Us>o!ps&jG_Vjdq~YFP@|j~dNNuyaLLXI{(H>Uau=sJ_Ut2H-0zK6G?`Vf&cF ztEEV1tF-d5>PMNV70!}9DDCkiD33ydzqXj3S-`~C!gd6Y;+wR|>7ET!HQeV4Am<8f?keK zZISL@bvRxIoBS#d-tt$+fU_XJuGscO<5a}bh(I%^%VM=)1#5*$Q5?w1e38_j(@gYY zJblRI8te_WSG}lHrlL#VxZ$)4`-&)U zJnc;g@5RiucnqHO=CoV@PBzeiS0}+Eg-B1>IS+LAmZB)hH!g=Q5xHfcT^D{anGoE&5 zWs%w!s^DkO=jGXv5Ls}7wW9qy2sIv(^$bBUhXf;dQ@PnzA)6m7q90b>JPn_MHU0Fj zeK@J_yBP*$XhgWv0rct}&68{)zT|Gb_6f11mAzJ+?p82h#+}>w2VFh)#eAkSgl=FV z$e(AskgbYYT}dL-Hjuf!76O-`lZ#hb8BZDwqAynWz|m+Ez?QhSGa0Bg0ge%OC-<{^EB{p8&1AHdHX9`)1JN@hjmENc7 z4Y+q5t<)?;j_cg+fnhyPI`AwJlP|)EAl;7@r)&|uk`F6iGNdC(zh27-cO*584dA}C z5-=3kH=(oi9+0H;aThSPd{Ud!A~szME7MCKA-(61keR5=oT0bLu;7s{W4&<7)TR@7 zv1E14MZ3ADvPE+)UwPHcSJ%8&TTvrpS$IE3!#;T3;GSbGV1cjqGa1JAz1xIHBf0}` z&(@zhShOfng#P7NAKo8tB;B0AI2Uhrru;42Z^1FmcL;sD*xO$(EtlcNB>HnUY)tY0 zLiOaPNY8EfN6}swSZtqyh>72_f+z9Yv=0V*I@ZTLa-GaWVvg)h7Z4pZo?OMRAcK%F zxJ>D0&8FAiNxP<~6P7}Q-*B-``udF;>P^Qqv*6*Etj@1eh*5tWD9!0uq-pkmlemp^ z&OlnKoMC_0An}4=)=U%Hwb2XYU7B0~K_DWS3FT#_zlkjnESf#`6GCbOlNx6u$n;$Yai#&LfPYD0}$ie{mg*g?T+SIjCq%3#8DPk0?J5lQZvCg zd)!F1lg0}%l>r2-!}`$N%k%3%6O|5wbF~^Bz1w&!{QcQKWJp9Ic(=)r`Nqb>v@T+u z#iy~~tu@?oFegB*!6X}st`Kb})L_c)31N%l?<*HetCmnoskRups=bd%?v{xkB8(AS zy6A|3PeRSRnc{AyTp#9c+3rj)TBcqLeZ5w=+4d)oyN33n+t=olYNE}OiMT4|7;}c5 zQm+PFC+s&aStf^&4aB+4=X>ZO_9Zii8$(~oQ`yxd(Qyq?MdU`QZ%9i@RzCCuc z0YKHtB6rDfxD2h&o3^ilLLwUwebwlqFNt9fwU{F&Kx7gYW8BcX6n#AX3FB~mqQCG8 z+d+SX;C^%J0DxS7g**MDqzJ@Rxi9yKQaw8IM;txcYd_!bY?yfaAO(N7JEE%iqXpoO za0MBCw@K7ujSZq3Hhew}+WybDCcm^cfaoTNuktIUk2m_uuTcW8JB=kPmdF-(NwLW6 zsJs6(wt96P>gd$edeJ6~;x{|Tz=(Sx2_DX3wn^`T;JN@M>)vbwP{MLtf1l{!{;RIM zMhgDDl#h1_fv+VYh4TK?Zv`|py^pV9RX-`>!Gg(KpuBLRA>Nub!|eaah0)dtvcD9W2MyO>Q4! z;O9zPdi(2Hwn2{-;QnqLnF^>x9cwY;!Y}2p?ooh7NUuEA=r*a&kFtJwQ{n(%_7HC; z$wGNuOee1g6@V8lvU1D=BDQZLJtDqLX%d?k9nR5&1-9&DJ{Wf2?xd>nF6kpds)G$6 z5++`MDxQ}?WfM5kx>$dXlX^{m`Zcr3A>8jbe0SAe+bK>H;-MylSs`U(}x@s z)|)?z&79dOi;&2$9Rv-Q<=)F$6otEDouvCyPEVEnu z`$rua$&9L#(Q|(KP=lvGpUwFTHODuUiL5#IZUpiDt>130kE7`d&GjWq!(H-Tru9SQ zt6SE-!MZuGv)?cd=^GJfN17tXL?{}E>r>aogm3Sb^A&!9*RmLr1aC#YT==U@Xgn%3 z7ivIewJ`yN^?z52JM{R<`rE?g!_N%~V1h@PyI&AECY~iesERG0=yo1C-R1$Q*7)5{ zG()B)Bt5$W=Z(|Y`outXXHEfs?aBuMAGHvhW8YAX_8xYM(_0dw1^M_r-mLW$&9cmD za<_ZZbn5p>>(8hxY#L49$=$R`#W5O_#+VrT=#4Ouoyl;3A{Q?+8^ocfs|R>uXO`1y z={WhsB|7n*AGZHPWU^$b(FrqnVFFKpN(jD3IrbB!JIVm0yTaiX{->nTaY*Mv9C|ey z?xiZ==0@{wFbR{+GC;nuOy;IA-1!_wCq+^f$KEmi#C90zUX` zk7#wdYH@Bm(E#_I?8Z8JLl{?&V!cI7{TAj8nO%FI?tTvZlCT+>zA^OxW6UGX2R#qx zbWWW~R@DL$&wc#7I9^!Uo8(If?tDOIZ7(UmdhjxX;#x-5!gb^h za`jOBCkqhlrIJ9i(LQ41w7W=$(^-pT&R3P_0_o#qQ%#ybcnq>-$FZJ1lsn&es$D2K zvv$t+w70Zc5ryUFeqjc08WqCmtlVmN$hfHxr{!ec8~lPS0&bSqy|Jr!!*`Q~)|RJ{ zcR+C5h~Qnxf?3*_Q4`GZQ?&E$>PaYSd5Dt7`107V-37O8@Jr{S`$0IwhI>}V=feUv zUx*JQf@J;zCw%nJbb-=Q zC72Y!!}=unOB)WED&Nr<{oK0jX*U(JiG`G9E8@Y(K-I6-3j~k?c?%bBb@wC-ifF$)D;qT4T#?*h8vYlw)S_ZQL^Lqe^Ok8in0et z0c^Bc@)(mXuLQ4b8q1}P)a9b4*46J(RcJ;ByV$H)|0+pMec9jiq$#kTyM5^^QzTEekTVUU-KbEKau0 z^g}!k#5#r z8xs8?7@vjS%$PkqKd%Q=oRbCzV=APsLaOAQy6#yb#?13+`ULr5yI+;O2V^_Q1%GNb z<<3W0dMnVWVz3g1t0LfX0#S9$4fXO2A~fIYwHre>+$-$JtVcf503oF#O}60!ysu^k zhj{tQ$dm(jx>%Kr;3vn`Z?M|lN-;4TgQz>}G5rDJ3hRwQ-qe9+!^fhW)i8!OjZ40> zvEL&rSlRfWe~-9!t*W)Sr>zQ4Uk7$&41||al)mC(`XLf-qsG;wVO~79%F&wnpBAsu z5sDRUm!2(`Wk8`tq!Fs8m|6U<@=e>INH-U3oF>)Ktnmphf}Z!mVp>Hd)k*tF`69Mf zfz^lLBbK|%wHGJR8ttHNc9^hIpg#El*!AvH;A_204nzr4TzI8~#h?$6lfj zj(a)57^ho2Box8y9l0%t#5wjN*7o6Y70qIf;>%S(CRJdjJfSrh9-+lPB2rSLp#4k_ z=W9^9<&2~WdgdlifDJ#-jd%1Q2P8@XGV045WKwg>T(T~)c8{q|NIUi{J{QZBQb_E3 zvw&*VX$w?lXp&-d;ip(#&c)WQyeXA=Rw}O0M+Yh`DJHeTJR5!PH(vs_f(qHf&J`>b zeTHIc8*U{`tNNp0s*rG!_R$0>?H)0`Th0Ugp!-(PbPY>VvU6lOmSDS#WW9*=z?wXy zkFB>M{B&!_#Pe@XT^Oo<(5?AH#*B^^dF;rpj?evxpWB5qc-5?$KEw&Z+i^GZ7eG5! z)~EATV~{zRo?hDhmK|xwX7fyR5pdC>%d@H%HYK zPnvU?H&&5e0{U+f3&Z`LP>k?Qlv*t`sua35<7mxVl{3mZhsN6lm0Qfqvgx3NQBj?i z&L|iC3(~fqzh2DADLw^HHVzQjI`^s}BWDokF+;b>E~`wD#k{Wa1&a|*I7I~1?S@&S zS37a(pSs(bJ31ou;t9whhW27{7HLxm5!I0lZD;!h9O+ZuB6A<6t*IZa5=I{U7G6 zpEg71N7$u{a0zG)4(^Lb9hL(av@e`5H#9QDt&^Q zvC{{B&&da()jAxn??D-UHm9aJ7bu~y=~0IeRY3a`45GXNITw}DuS(zEeTJ+cuVFUN zsMR@S4Xxr;{fAw{7(HQ~g;NE3xGgP#iqK9#qlI0Ta-W6Q_+UC#>*{dav43BMdR@V@wWy5}6YeP^L*!SevfNn1#r_*@XB8FKx~*Gc zKyY^n65QPh?(Po3-Q5xhZXvk4yF+2Y-L=r*?(TQ8)?Rz>b6dOT?bcHsXsHG@=d6E> z(Z4HX<< zW9>j~RO5ZLnb&*ZNY}%TJ6r!TqMe(3&33w#!_>gL$0b;PBTMHduICpnj;z`H^;)`J z8<0B^)Thw`^bY`x!iS2+4ztFtxjp}ZjJ+s+OIr{;&g<2^%awT_uQjoDaw24d9pQ4Q z5!EltF*uxFC$q3ghQjU}gr=t7r$BN2+U!dj=BUTWD%EcSn->+CTn96FOk|Z>H~4sBm+05DMb zaF((*kQbRwbc@kNviFk^>GH`D7}oTbgb-+xuZj!I-GK2+#!rr@$;rKD6(QBf^|(E( z&NglxFJXF$0mc)Rcb&5;tbDQN0NxQvE(X%1J-fIXZn3*^$in9l1r)?swdR2`T{E8> znEPG5DF>uRJt*xU{~w}(U*c1q@9847HxR+!TK~4D;JHqV8aySx96Us`o6#TSg$xOI zK)gFBH?;qT;r`piw_w40v!R>9_Y>1X?YA>~6_mCB&DU~e%>+QGsmm2KQPqN8>e+cc z3|`3fA92PN-Vt=P{WcMUy4Wha2v0D&_r75@pWN)($*3qS+#=viwRI@Z?;P(gr5WfE z$SqTxtK5=_HPNZOgh0t1<>GCTrg0$ds*?9VUm~^tqSjKOeTw?ki<*^;A-$k5P(@Vg zEn#1MC#7%;wy1QN3O(?^S?anl`LsB`ao29AD93+$aF!>>+GxafzYv1Rv3FireikZ< zHJBoN^9@GFp4I6Bmq1XrOF_@Oh(_SM0JR>TXoFuKIT4AWKN2#$60mbP1Zf!wB8A=o zC!*Y>aY0?+H>GG|{B>+c6(K>Gr=9(h5Xec%V7)_Vb(!48*WZ4l4-*|diDl0&6PsO0 zTD1}xbL@{^38(~9ZNqZXf}@?o?Y-!=_m>=noHHk%ryOK5QF^kejm85G!|$%rwlq=m^`g&pfjv{!imi%I4TelL zp5qhL(hW+Akzpj3ZblF|PsP=Oc+g3GFVeSn#>+W@mv0!dxnn4MBy+YM^+;3^Se1w3 zWkvm(T20P1qzVM_0^jlTRg_Z*@W~g#pMUqUhN2#VRlpkmDGq)hsXp%iXqZBBq~U$P zpVBiqVc;?|?*EPRlR+>J)5#&i;p%S$lB3ktvl(>=no3jUf<#8!%ml)%_l)5K-dw;# zLVSq}libY7>r-MM&qTvlCV?xGiPQMJ99=K)w|aQZPI0#T?#ly!y9s%=7ZZoZwdt#L zGE*}(H@3IK_JESd2J#gb6RSVoc;#@k0aFE_RhDO0i*xLY+yvRstIOc-VBtMf23=Ll z@ohP%+_}TSvl((fl(@DTAn5R_J`d^+mVfr!9=zg+mpwK~vG5BZ={!X{J}@tysr)(D zWfxdHd8I^Afk53&GIWTnEaa_z6QOq_y61b00sT$kUIC`dLssZ8)Pwf9Xl8d-Q04~+ zGA`NQv@)}Ktq<`dA>2GiV-img{=*DyY1-Q7YtULYtP^A(ZzYG%^@6qSHG_L5%f0xx zi$KQb$j#a(>WStli%w4mwN=0fSFLq@{6;2-$Jza031}RMpfj9y-Saw~{Tq+j_T?aV zznKL7pv29=bVoqjgE3qzTI5b>HN{OVw5I)7llzi!gN4imCOEaFtLM94xBoSPfG_W& zU?Du!?B5g(PAlJ*`@pTDmC;;g{A;Dy?8}SkKh+LKOn!oK8s32JSV7J}^Q+H~NXyCF zQ6yIH>F&o3cJ1Cfhp>nPA++N-JkD1Myn`eq|E%I$BB?>g*E&v9a^M588kbV=mdNG#yWg zu{SEa&3vYIWKi3z6bolBTC*^y@f_R?t8ZV1ngG0dWxegIF8i3ha$f>~3?>i=PWCG1 zIAVdJt*<}st_T7x50I~)wj@O~yFtEa&gir@;;=UX#}7EULFD}}+9q?qv2vl`Kh;I6 zKkvW!_ED;5G7;Q^n(+mB!8+7!68SQS%7Yh0^*O`wOD%Ak3~EaWbYI{CA-cUNdt;6r z5};)mM{H=O`a5R!ko$xo&ZF*UQd}7=k9kAFm*ZI!#?z*S?{`iXF?eiBY1Y9sIg#SV zxYCuolkElhiV+XBFOLuBfMAUQt|H6F3>ZA6vcx$*=txM31?8!9-Ay#iK(L_60sTk4 z34riX$KTx)^uu4n!@0L6^F_J&^9x>1Zz{*A{w?z5+?t`%kDKIUf)xR2h~z>a zQk-)5DyQk@fO=#Lqatfm&OZi^o_idegxDykP zbYn8TUw*bScUoDkX+IJ9KY7}3B=9kMYstA0Hlen&`2&;4#NnE@Emeks`s?~0>d~KS zC(lcPujG-ZyFu4ZAl0&WrijHqy@1JqWuOaGX(x=eD(}uK!zu+^G@*x}B=pVE`RL$@ z!lXv*r@FC91>ynv;dHCQeMa+#<2Qk?47Jr{6Oc@phrfSNUj5?be9yj2OFVB-sFw`q zj`DyQA*j>TR8NtM|A?N(1!P6=$a7CuZf@KRV%Df|@Vqi0lWpD{NWOI6taAQ& zM@~isrI!^ZD^X<%#+=pD;Bd!Nxc&2a$2a&4_ zhy*oX-mrQXVh8#3xmfa^33)MOkHKADWlLdg)W}(NOAzNCyzl-9jH8er#%O2*-z<2| zr&9#wffLTJ`0Mi(bS~aSaRlY%&F*0e0C-o&QY8ch(Y(JR?&2_s<$m2(Og+6|4Ds_G z>--P}5x|7IB=P~9S~CUFnzJYNzWIx^{LVL^1K6*vJ%IhH^>}FW7B(^6zK_&YYx*{_+e=sF#j2flAcnmBRc{ak}QCwsGrUS_h03o79~a|IGB z0Cj`xZMrXbIUQ5QyyNO~Ddsdr(oQ6+cax)=TTF2t0W8KO!^tJQhc5|7FxUB>#8+^j z)IM~GH{ydQp|4z=c=%Y?4sV#%eOB+lgSsoZ_2)l zMCVw$fuiPCzt@wN{paU)_iXsijnNQ+#Mg}0TX_AwCAAtzwSOHk;xncC2gp$t->~L_ zm5Az5)_()R4%vnJ%$ZwP1*QNTEsneUoPm`Qy_RE~bGpL`9rx%;6<{j*`fL#by`}t$ z3x)ZE@ZR)GKTtL0oQcy7*%qD?&w0k%>8xCA{5x*n$(yBwjz7nILC4m;LH|_BgE#%JE%Up&-<3`LkX~ zg~1te>vWAP)eCODZDl%z>Jy)+#{IE+hI zcmt1G&`;kkL5$CYeVkv$Gp;7QBfC_*)CIRE_!X~d*wa(KAM@?FGcU+Ifo?3-l5gi# zK?ekN3%nWO;htvZCu+{i7kL_(vbx^6#f_=k?SiiWoE^(skEd0HaAD8g?rJcZ*2ZD7 zYtP*$&YYCnHDD_oIK(|$#_@E~^YWtNnR_hW_pS*cWyY=j@XkrO0q_FGqg^&%{Yfpt z`to$Vi5l64cqYh=MjNvt-*sViid;L>lG99Xo#MNNZ*Qr+6oe9hBmB6ha2Xwz5c>J? zK3PgXjP1S;@y}xr%4bm_lY;!}SQ zTT|m&jmdo<4-825sa^5rDO^8a_wxm089o)BSKXOVRJ2Jr!C6bhlKV{`2|zn_LyB3_ zyU%^SBu*A8&gd4U*4+=TN>`P_PLrn*>PAjd_`KKp=!_6bIE$ZTVCB2EKNWWw zwNzi})+^-}m&e<9Q!@duH}n`$)8719Y5Iyu(D6R&N}fkVzE?{>cIOV8PF`nPf0ZuV zH_3Drw`BlVw`<;cT;fBejAnA>We*$rIrQsn2)tQj&R1dT!hDQKRjpXBPF9LuJ6hv~ zXFV^?%TW%iQoI ziIk;XRFE@r*bTt7W<~p6VOXA3ir6OKnH1#ZCxd&l);?M5E0xs;pwWEb5ak{^+X*32T`7=BRrUwlZ1IM9o1cEqfNy)RbA}*)L;eIVvV9x~a@Jh+{ zx{ex3NLq~P6@BqsyjVtq;}<6hO~18jMPvj<6%hk`U%eHY)Gp1XtHF@18Y$+n4K2u) zjGvAkVP`@wxTPL@bxxc6q4pBx10>J5d9Ou()_gZ_IdY9>A~nwZ z`&Uu2aY=Xu_w1w&rYH9<&56SuR?$M{3h`TNFO&hv2iT+{3i~*xC6)7Nx4pkEr}9_l z3&YefBea}TXv-sHZ%Tb|xi8dKlTsSDziLo>NxFt<>Q)Z~RD%j60D+aSF-SMe)CKWt zuBeW=VwMhh%?_VbdAmcUGA%5Itmw7`Dd^z?T%TH+TtLK1C!+?Hl$aQ=stlr*#8*SD zg9tcy`HN_?=fZZKU0Vtnk!pM>J`~w_f+! z%@7K@gn^JPF}WT2ezEQ5)mOQ8YMjom_rbR($X@ChRFzgNHt&;Ae{WRCDY|EO#1e*g zDFKLCC#}hsKSvOzFZp>I2|&=Xc?2|U+r0qLjpxCKg;ez2aF%YTxD)LmJ=JYCC&f|! zxC&k=)Y8tfGc>b*CGZdai%tFCRoS0nxH(_}c2#u@ToK~ioMn?`4d<&;3}bBWQ%3kY zOC(jnJ%pXmNKvu`c!;@GF-&I|bg61OUO8*wi0?bffCY$}#bV~ZPv}R=K;#W1;9wS> z7XM43pkZWq!t6#0Z*dkpvpv(*Z2T9I zaRLgc?bZGtA>qW=Jn-KXMbTi`sut6*WZU!QQ}JwGLX{p#0(b{-k&oW@8DFTw@t&GwbTSSH%2d$87-jE{qE}t68|nZ zSL*qW7pPg3Qa!3iHs!9Oi;>tZqg`u``+VHce03*yxkgi?qGcd?34oOF`Zgz28)$Kb z_xC7RD%C&l46D-ddK}RW6i^T3O#?qF8{YwO({Isi4h7|MrJcV4dW#Pr_Gt)u{}=ygI{&A#P33;^MT|5JiV(gPYVq>^1A|GPu`u&Spe_@ zz9AbM&n{AbjNDqXS=C$WhdDjP_E!_=-J?MwK_Wh{<^kbi5lV6|&Fj|6wVD=KPX<2o zzD}SdGx$_^G5-kbIacm2$9~r1?awcP);=`pX1E`^4nQ;qkW@qK8M^)qTF2W#w_ao9 zu2eA04nIcv-x+>n{9YDA{#@sJGsauvJXDhQDgv1_e6`OziClSp1G4;Q71<0*H-fM@ zD(=y@@v2GzKQ-{^B=`-nG!%Yr-(4V zz2e@x9ss?86&M{SLJ`J*)zz24L$rcul}Rz^bRC8g6m1s|=APqeGx@51Fz2EF8ewG6 z{#90{h@a|MAxw1v$Ql^ymA;cKMJ$QMRc-s ztcuEy#N4HC?llss7`dVK)pT@>yJ2XB_V)n2B3{&zt^Cq%Yes~?5AEoT_02iu*W30L z)E6c$Q}v&=1i))JLprP5$Ms(N?uKx3=I5Wc?EowYgv)V&46bIx{bBM=PxoQEoS3uc zebjBi9Q6I7E8=KU==UFbdX68hfZwhFi%wQbJ`OX4d4v2<%Rjwv0z-+;feEg@i^EEm=FmRwztt z#<~z*J1xY`W>TPy){m!!Tcm74VnYiB)ugJVqMN0xree-QlQn1Oa$zx5XDe@(xU`~6 z94GNUWQ}%tY|ny4QbfeU;=F2EYY@iqJxIgd!=2rHW&LsQhu@=ihZ%YDPb{CDBpasE z>fqh>9*D&L*d^7)yrKn>{4+ac8G=8>s<0vCHTZb7&jct?MNBl4eViOE)u+4ZQ<4Fy z^*g_Gk_0mY&)}_@){iybJ-9eJSI=3DuzQoDNZ2P4Gq5;}RC9s{FyDjXBcvC|=|GOn z@)5!+s>;&^dfEyu9f8(6Y{#uXmE?)XRp`?By2wWWwNPbB%{v&}Y(4m|y53nrwh79_ z0&wA-s>+&Yo;BMC7-N(JLnyoX%G%2F)j9>y8~NqhkIT{i541A9>igT>RAIr-j=(OA znlHkl1o#KDpDAACe|d*}GOwY)EE_()ozQMdLn5G=H~Y}=O}@$YHjjC4$1zvut<;95 z=drl_$LUY4$1Me>C%Om-^`E!mjN=B+CF8*SMbi7*dph|jqlAmp@~p>n*Zzk1nt}1n z&^<(LRL1rz6E$}q6E2i3BEG0OUaiF0`?XQoK&+x^U((I3MHw{w$nxUEk=L=nvEsH>BX(*~$u{H+s8 z0y$>3=NGzlxGaV*6&vs_Y4hkr;!D?5A!iGI;4b=u4p!dSzZrfgs86p7fuDe630tUa z_zSa6y5f3?0N`C~ct~Xml*+~*1Px)2nN|kat64Tu4W#9l%px?+&vpKvxslZ^ z^kv@Tjvk3cOwHuXr&`-oFA|_Dg}Glb^^IGHrv*j|{g#@lf`*FuMolxs6@9RHvicmg znH`%iOa_0+JypvLw<|-~B!^-c38!nMeGl&Ew&WNsa%`xbXC5efP`6N;BcsMiN>w3g za#E~YMlIg)qj9fW3s~_--D&Wl`Y)daw_IT=*ocfDhhD}LtmcXt@!l!$-h-oH*k&OM z>V{fw(7(kxU{}NcSZpGSm0q}cy`RD)sm{ZO&O?X1Qmqz#c(Nj1!<@OeRL2>kqmdiO zp~X=vY*MKtPfk`VJ=7-;zgeUO%rgK>y2#m-u#;yt9VciQvoOy_c<115^Q#!x(>HnO zgUsm`z=D5n^c>KtDHF#*mOT~@v_MVqbZH;MHf?-b^}rhO&RI0%3XRdpeYH4FeT7{Q zE;&oUCJJ!1N=l3vVt(Gto@t)(3j8H{{}n0RU#$I|sFlgH?|!E?LXO0U0f0Y}r!TXBps`IC9JiG-#W7t{oSi zSXS`k658WQ+BEJ%sbV_B0NPAxyj$XtNletK}Tzq4p)k4PGGbK+bDk-kL zm-=j6Zi9_&+|<*_yZkx&D8;9dY4Ns(gFX-l8NP`qMtK7PX@|hgW1BBe_}xAy8Q$CE z;NZ+;RYa@iD0YmY^18C!z-ZBvkGtV|<272&V+_PPEb5Z4q5>>fExP)F@6t~jHJ7gik@zH7oYX^9^dm+Y_xO)}Ix$d7~;DVs; z0-@ZFInrx4r0bQbrcO|k2Ktb3o!qTHL&h3^$XX_ANF&x?uKMsp6#f4SIRIA?kRquE z;zq{ckO6LAq`t=YY&XH-s6owtr9-N5){_nl6lElg>m!sBErBH`(|-(Hc`v&!V2nkY zUGS@!6=TtzTg~cc8eJ7=KIq(4cT7Y`GX)M+({HD(D)n(<*=iXAXF;q5Z&t&hg-liB10DQO_u5zfxrLw{nuZ5^(&Kd+=jZN2;3<$b|o}>y#0SAWc>@r=H;2P4=CXtP`a} z9gM|^P+d17j~mE^sq4kas!*j++b1jM3&KCSWDZ%!=>eWO;UDfm&gR>jsPe|-Ma6}M zp^;AFC6(MMnwnW=KhTUtd~k4PCrEVXU#^w|quoM$XJDr0-*seVb6DPpbyaZ#>A=u5 z&i3V2z2k?kEdxf}E4cCi7+U(szXmyO2CU?!=GAFgPY%j5+d_+bYCLlCuqhIz`f&$} z*qyZcQee^pGt=e>cEHt$yH4)h{)0Wem$fB6eC%bLS)APu1V$DQodK0k?bX%sR#QSi zum3g6^OemCYMEEQO$)G^ z1)bmPWqf0dgOhv5IDaSqv-9yRHzh4Of5mlBe3>sX+)K8eLw|vU7$#4m&#r40lN|h2wzg&jat&`2?6@n6DCY`t`tGu z|0;Z99nRWLE}$#-b!895^)N%#cn6#yO>&JXSG2Mdr}fB&RP*#k_vb$w^~HzPD|V0< zhk&&0FYV6vq}LL3dWwsaP5kKAfL7F9J`0ToS;e#CQ303CvyXEW|DYR2S~9fcU^^(}0zKv#j)D`(%x|mr*;ECBnB4 zPb5CXP&`(_>wpM`J&E)_67w}YY;UKqS&*jA$Gvl@8tw#Zz6>(2*$+LTyQeM-0)ya6R-;V$;f-SAS#iq^?^LE z#JYCyC&P|v3cW4&{`n%{qWR^p2Fl|8&7I#6u_RH}e!V!r4r2{O=yWbjt0P~MTwMx0 z)jeDG^V&P7Y9!QRYn`LsUwfHLT@SPauBX{J|7;D{kO*t@Uhospay{}d&Hr1?^l_IV z8z@7k5T66ng8}g?Fy#lLL7`1V&xeowFb$P13&q7@xcpH#r~QC7YtDfN+D|gCyjU^7 zjg#cQXQ-vK`ttB?K=Gg`8Mi%SPK!4398aNimf(ERq^fy&D<)&(knrK|@=No(33^@P znpPm*{m}0Ewn)G9!!IkqNi~tQfjmB+wM;!1sHDiqruW70OVd5irJ>UXKc z-%8Oo%vHsBc4C`$<*xi66_(S#R9G_=fk>D5*z2=hXMQAk99OYwQ{{!Ip-QsZvb~Vq zEdSDQ(E&o3BGL`G(81#NN&Kq9PYaUnurTP7}h&f zA(F-Wmr6|D1D;jRzv1HQ&}LbIVOs`jOy8;m82&d_t}(;i_!!Sc-rO5{EXOA?B4Qr` zc_!9pG?Y;Qht1l6EDIKLJP{Kybudez?V6DT>MiWG6A09#f=OuVdS&mkg2N!aDIp;p z?p{du+xPA)A*t=R-9urLzRNl)LYKH$?10865lANj6Uv zR#l(I9g5v)e>KoMoQE)I&tOcrl4&+*mTm73X#jl{7x%>$Z@q9OG~-_07Pt|nq{S>mv;7p)WsYzzsO@G zepO4Q;^U_T-I;$F2#D40reR8wQ8Z~2R48Jcg8*`|4bFC`FHV^461nO=?e#*k&$mkx zB2mPty9|c93g;888Bvsyd@F0uCuvW~(;#AJuv0(OSMqdq#)t8n8xr36ix10rmGNw8 z;GV`&8S--{XYA>YHXUYjdFB5495%ft<(988TxH#@#+Iu#w9`k)_s|hUSc9fTxM^%giip+O^!(*+{*hcxf80G#Hw!0Dz1 zco+}pv@m|l=XqWdT=>K=%5x4CN?niA>+1c8`98G|gV)2T!ynL6_L=mplsX##HgoCm z`2|Tia^s45FRPi+DHWC0@0_52dI1)=A*$JB4z+QlCYy9m2dCeEkeX@PPUw<4yDU*?Wy&lLXjavKk3PeFV_%T{K961qMed(0mP%=e&VKbLCOe;V(3Zqtb3 zr927}m2cJcg#$Zhc!O)YvN&(jb*~`F%vhhjI|{eWPhD(Ik6qwx0&*biS?hfl+A?xZ z+&UUme2a!LhUkHe%~ZdRccp%BM&}yFr1d12^(5OyK?SXkH?w)XYjd1UD^R9B{EP71 zSIESE;s8l`=*hfhBc?^Ev=s3gqjUatuBVByfSy53gj7J@DTYZuYO$L+Xsry2X6a}?pd#+dwkMW4iUAx;v_i;3x4Z8^B;b7WD6%g zm*YrTL|1akpa#esuaYp0u8h0@$JBR#V(TQZS5DCN9Oya!K23*x2=q$E8aHq=m{ef# zb8TDH^CE`D=_5J=S?=?6a?f*tLL$hA*$Su3a@)h{TUeUuARU?>4{Y0Y7;INY)41VD zN@_s`r>Yca25sfv39cpMxavb}MQtQHxBSwNh~KqnvuZ}9MGVg^7ZuVF-Zk%2&tDb! zmKzG!0Qe6YB#Ms@rSTYyz$&qT-}+`0C~Qi$)D1Z3lc>AqYFOw-FR-TxuY+lxhl?8& zZ=QdsK}lTzUC^AN2DkT#%T66Sy=r=JEbJAgQz1OsTW|POI*+ zAC>l}qv-@XWS(L`|I#Y#QJ8wO?ygky$iYh?ffpVj6l zplBe$VA387|515Hbs$dq%2{imxMG|$nDwe;kmDqX=VtPT@&>eD9PkUQ>}|@?=i=Xf zI4K@HZ{|5G84w$`VGPNPHY@g7bYZF}EY&V@__a7-?F?0YeA7Vcm*P+JyDQJx=vzb3 zP*jQ>mGazDbTE2RKfCv&qKHp1V6?KwIs%?-5u)|9H^VL`y9Dzm>P5l2H#slV#!=jA4y`5%Ey$tkIE}m;zO5!z^9dqNN29}7kVPln4R-%*O-x!R&fMW)^cTqg@H3SwiFAaunO5u*YN zvn&)8K0nyIsssmVlEisEh#8knRM=gqeira`)@U%D>wAsUU*HK4Vv%r@?k?zn!sve* za9np80^RFKPw@mPN@*Uq-R1`iv-5Q$yzw;zxJwWBW{!M-CAC2K_J(#Seq%*Y2=(Ll zR}Do}snaRUK+{4}1X61?DW)}Sndr?$Vc<@5b^s|`NcSV0>ds)Nc&?crmaoFd0m^cV z`wanatYn^jjCjs0P*{%TZ4)#nVs9Qq$LrC>K|A0|I%VpFgj$F*`IXqHt?76WrBQc2^ndaHO1GNQ!Uk_HQCd0OriNyVx6{3 z9eysLG$1-8ny~--D>d%JJy)Z<8wMKnrbALjl%W+aJH4EifX3nGCemS9b;2Sw74BM4 zaWqv22nb)gPLSWw+3M-@LDW~FIyDsWfAbx>P((_{F5gf=Gf9@FvFYtBNkLv5IaIRi zk@?=+s;Lkt)Vrc{bJ$Q_8n)iNp65?SgH~TGX`H91fsJ~&wX5&Bs|`{tiXO;SNv0|j z#%UP|$XMy&D&1lF?S9+Zb-rp3Ktn@jc(Du<|;}Qb!+ic_@!7!xQq_bjB97m#V93YR*~O z(-QR4t*9$?Eica(tGf#GnV4pmQ?OlHCVmO60On9g5}}JDc2D9XfE1?tjZQ$$>NG++ zQCttW7ZtF#BtKcf_wD)WE{2sC3<`!Q#-Xbgq5T78zpJ`@Q<>Gt{(GWWtKyKl?gw}$ zmDP%y-OV+p8D@F5b0(u1Gb49*6P0KhTH_u_0W<5uJ*B>IO4G!xm7?N)aMTJ67Jj`u z3MODi&m~C;Us+3s&J6n?;jwShWU909i#OU9(>(ZoPwDptdW4s9|O(RmdyIZjb}bd(fq!~NsJJ7hgV z2Qq|~y2$UOQV={*BXD=VnYiz`!h&|9ciE7C!3|D&)ufC9sZGn@`KV1`wI+_mOlQ8* z3R6|pt$N-js3>%}qmw!6Spg-w8~Yd#_YgBK2}Can@#47CJTr72Q)Heke+5 z=`oSIb5em1GYGQx6WA{Sh@j5r0)E%&^P(2CAIShJNF3&u00bjdGkU!YN)(G?CAqay z1au5L-}0~~^Dahe{~WY(8v8VtX{SPT-($pi*Ynxf7#c2|KMb&&AAAp1fRCDYh7awS zNN91ax{Ui%8t>39afl-ZZ*7J305;tgTA<>NQQ*4`lXrdah^*LEz0g z;yjOVoCg`)1O3|+_dhMqNv-kw8K6p6L^H0;#wdICY56iprtpqwaT$Qtn9n3S|n> zNNjdsuZz#Uy}O0OFL|hVgiT3I+l_bF?mpA97O?z2s;Mpu4aWRVF+>cgeZs3}ZwDc+ zMIJoc>4uuT>_(NRYXQ6An=LW3pPjiRfGBZ-KO0DOV7iMs*H6o7b^!|!=B~33imLaY z`?~-SqdGlMvYkC`*=yRx$tr6`c&tQ({@Do}m|1=&8J7o?*($+6oB zeY1WnQr%eTSOK5e_)35C$pc2^mt&r4h}Jd6>YuW4&It6glRx3^5zXKQ!Qyt&AEW0ae)w7m(az2|H_e8X@4g2vGAZkNfC! zT|oE#{BPX*cj{x?7#6>8J4-AhCQVTMqCdcWiYbPhh$NY+O zJ^#9}RoZ#E)M>b*r057!GEEgxe;uTsB+lOmv)B&kamRyiY}vhl^k2U%P?^=)a2<~P znpS$vnbrOrHwsTSBxOVP_wE0CC{u>e!ahtc0xGCuMz}5b$pI& zP?jNz6`WtaPNzqD36oH+*&_TnBkbe7Z;K>p)a@IP&cO_c!Or;W{h9IN(un>+_>$CH^cBD=%%f`4>qMT#SUqWbj7GYrvDPM zH)2wyX)a#t>?@L$?)%~3TKJ8Kf9@r~MX4Mk?%wkUTE&t)s9U+LL=|I=It5(`wAbPy zmL3pvyj$U6!F~mMhAt~2v528PpZNe@;mwHLzrG|qlRMDt-Q5!5XM~ik@L+b?_Iq=( z=izIQf^M5Q%mFGFQ6L*N0^AqguSKN>c%GYtjI-f713}-26fja`QM!*hd?D5ia0YNy zpwQSOF!Q>)r+z{_h7sTU(-J0Y95M~*$ovz zIniF;TC|O`YLi5D>D)OGEvPOmQgBRMPE{xhj>Hiw5gl&Px6iKL3VRx>NIc23%A zUf`h(D{N$bD`d61VzI(kINm_AEa~omea_Z4)+iNdhoTTorOTZhvU#@32Zc$^PyyQC zdc}b%=7gO`MQ62)`8Q`rfmf4h@2TU8VJS7J+3}?Z`yDGAYN91r_AM2c=4*g?!T8ph z9$=JBa(iQ^G1~+5co=!^FgOXLPUivTNO$q8TK$|dh{ z5l$4Ec7=jyknf6J^AunpleFuILU&6%bFF7XeAB@tCuoVAt2Tva+pYgYQp%$Bl1LK{ zHWtFwaWypu@a3+`XAkpbqu!Zmpq$6?g`}PYNR7h#?u2-Rt}t9a%Y0wSdK+)KXJA=? z=^B^W`@Qm*0-%rON1Stv!#3SBVQ?@@YhuW-A!Bs?MO3HZKd#e6lw}sE_SU{le@MWx zg)o-(U6TdK2d%&FH_XtvNY8A4SMx6+r^{!QZ)9WaOk=y^j9!@MKNnZ`(N}hSgat6! zeb$h{6OSs5tOTz&psjj4q?$&&D%h+*`;LaK(FT#PI(zQr=(p76bE-PqLoy{UC=Iu= zx#w=f$lIakf>yK@XUQF}H~08S!YlMmalp1~+tSWe>pUfp5g?OtzMq>zvbyncX4lS5e=7*AYETA*S#FW zI)3L_nq^15O<6KbS`uDA$RLTg@7>XtMO2-m7;&VDs~pcueemm!=TZjGOI~h+%KTHx z6Baw8bQ5(jrPhNYVXbFXF`s8!;!&1;2X-QlbZ=tDKRK}T8=JS)?vmwG6Q@@rZDSc@P55KdhOF+eY zf->#}OnlAGi>X3oZh#+ky_#9px~0b(n)$TQagRk8z^e5!9dJyh?wPRPZmL^hF=%x+ z7eO;H!zQ|dNbKGE;BuI#p_%RM)MocQ3=VS!VJu9352q3=qiJyZo%}oW&u_8ohh*ny z@x$duAT^c=+gEE&e;k2P?tI77TU*_r0(-6rj<+`oUcfqG0Me%*JDXMJyO}>% z#A(Tkw^_}Wg>Y7;y};&rcIGTQG}+h;c55;p%^2W~mv4M%IkyVb%qqRf+hA|1)Sc3d z@Km#MA%PI`sitecSUcRpU3`)CebKL~q;wwB6B<&)RTz^D`fW6F^vM?0%%XM*x6J9r>=Jd)R#Z*}B9zUAA84=nQ6}Z>Ibflto@#au| zbZ*re9nm&kmR(E=w&(?*9-gF3>_bg7vjT@pCp+f`C6z)Isyc1Dx()-SVsW*F$DiA( zMQb{ja}w*QXbZVInB;1RA^K$SG?6sabd7dSwz8VlpDM>X&>S{o9n53fg~IHbDhkSk z$Ma3&(0;nykjD{}P;UENUw;1=+zh5>4~SFAC( zpX$F>#QsOr`oF`l|Dn6y{{Q?zwodr%PYKP993W8R{1AQ|m9x3KyMcm%yL;0s@dmsu z=|SJ$R4Nmn-CQAD-ynYccyV@C7c&GbbKgQo-@{Hypwp9;lqjAw`vF;`*ic?Q_k z*Rvf{+8K?Hja7Xq%|7&wJ4Z0RK|qlRIJ*sI&0-Te zbA++06m*QJpIlxda%Sh8-U%fB}wX@a#)OR_V6d*mIV+ge?}2{LC9xi%haQ_;r|a)cW^ zHbw>x25xGpJjI>OE%Q3f4Nmbq?BYR6&5qGo8-IgbO;+KuebbZM@5;$^(P&9wKA-1C zN#QS$|9-RJOPo8(ag>nR#U7hW9VM|y?7ek$)t?TfeSf&OB4r8Q$cz<;qtF$nY9Gbr zkvBY_JJ)Q}dS34y@9l3Mt?60=CbDK{i!G$G*$xKTGCZ2-%>k=OeRUOgSB*15wUzM& zHB}xx)!M4uPgyNS5^-Frcx4oxdJ50Q29bk<18xnuf1KP0-l}ZT*}l4&XzqAKKj>y>>_%c4D+Tngt3%9cN6v{R-b>bLtlvul+dD zjl7~zlvTyp_Ht$Ex~d1ETti!Eq?bbWo4saWIV07GgRUo)t+--f*`RrR+GO)~F@kDK$8UBk*?7_o`1_g|Hmr2)QMT09ywqo3 z4h^1TS{?bcGQpY`=UgZYD4LtTsF+V+#5w#d^9uOarQDDZX!I?^b1fi;T=w#Ks?VsU zRT+Pdj*j7zd`ZzSu=s~M!4eAj_=LEG_yly!1T>8Jgm@}ywC_Gtz?Q@gH$4LqWWgt| zC7t5cAZ=r6y)sn<^a{V;Z^|wjRNOk&Ol8vX->}q+8WEecLsUuTiQZh^z-a5t1l+qf9J{F7qqM%uOw_i8%GHyovUMHD-nL(6FW-6j0 z6iH7q_b#x3^w-;eK2I_DruQEAz$*7ZM7`Df>!+^O%=#joiRfL2L%3!ES2}aQ%b!U0 zy66uXIeNxS93nM7`5x%Rs1jzx#$R`xv|~~ZF1PZsLQA~!Z_kOFnmnHbvX8?Y+pGxyy|L4VK1 zez|n-NM`uVH%k`3j({apWpnUl#GsnBtl-7* zKwT-KLsj9!|BMU_|Nn>FW0PTEU|<7vx|tacpR8S2zC^eOD8k_B>gTe~DWM4f4aq%8 literal 0 HcmV?d00001 From bd27b88bcbd7831584e30fdf04a877fcaef076ee Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:50:56 +0800 Subject: [PATCH 269/311] Add pricing calculator screenshot --- docs/images/calculator-screenshot.png | Bin 0 -> 60967 bytes 1 file changed, 0 insertions(+), 0 deletions(-) create mode 100644 docs/images/calculator-screenshot.png diff --git a/docs/images/calculator-screenshot.png b/docs/images/calculator-screenshot.png new file mode 100644 index 0000000000000000000000000000000000000000..5eb8058d968acad958f8aea7e09532a396914d67 GIT binary patch literal 60967 zcmeFZWm_B3!Zr#m?ga|P3luNzZpEQE6!+p#+zIYlthl>dad&rj4I13-O!xcj=Q{u3 zoDXk4gdrr8$z;u1w-T!GT>=#e9|;Ny3ROz-n-UZhEU@`b7~vi8BatE+3jLgqM|NHnG;&`SYXuCHwA_B!KI51GP><2Rk$2^xbwCsO( z+0ohL0KPq0GCC15|MSQL zV@y&Kcqvvi3-Nyr;2s9M^S`h6e24VEFBO3Pe;uRGyFdSPkRkFy|9ddKR9R4z|MS{y zEQSAllM6&@pZ~l2|9f(P5tjM?(a1r-2+$7M?DU~ckbZiA-!(P?@9tWpm@QXnZV054 z_q_W~xOjJ$(k3&1;@(Z*th7=WWZ^54@|QL# ziI?Z`x^#TLxB@GesM0Uwi3UEOtTsMR{!*;=Zo`LdY7(Shb38Eq{rUe~7BL|I$;rvt zSrtF&<*HZn3oMb|X8;ua2(z+Yeh$!~;wj#aI9Pj&#aL=khV*IUS*2JI=07oIMq z#>bS&>7~@YExIXUmi=hM&&MI~_MYEDG@BSp=6@O*AaMb|#hK6WI6`VwGLaox6(=+ z+^*%DWAfabh_#Hh=B=sC@mkp*$;g=X*6B$}Nf{YY;#cr-RFyg{YS`%;Net!MO*O0I zV`J9KMX`6*cQQoF`EVwOUBM)f$?DwNAs%iT^P%4ql^u^ISCi@67k%n__2w({i(Z!7 zz4C$&-=24{i%(wQsrA~QZ9%W*yNq=vXBqse=%Ti*b|;N?;@Rv%^_tgL6T}&=A|RMG z_Z4fG2m<&IrOjGj6WL4E1)oM)gVBjDf4MU=L=r5uTPB^OIc_^bCZP#FeE48FU&)z4 zdU0`~?tI1ojOrMf&UPKGJio@zW`=eMY)woncbS^C6Te+4*+~V~exItxbc0xaA8UpEjILn*BuEox4E3 z>{3l6K0C|vhyyvJV|F!@Xl!gOviIX*V3MwVVwMO< zi1aw=`PCyYH7zVORO7GxUQe`ug;Iym44cYOU&kxY{_ToRhSrIUdXYlcge684A00hC zs#kcBM*^)XB$G+_--7qXf!*0lH-91}B{w4IM9a67^6hB^!6_7W4+}9M^i57d_nNFQ*WYBlz0PtJ=^?8M%|v`PurI7 z6J%s`l+D?o-kF;*9QZ^JRM}9Jy47d*O}WKep@=l*c-%}@d6cF^{PwR0Q+Zy6?`{)g z9&jBG7N_!`h$;f?q@Zhu)`%lJUn-i!F{10vCbvc^I~b;EfMJ}Mb{2Mfp@l`oQ>pgG z7~~u(gqc969za9ovhJiVwd6Y_2!~P|4Lt9NWvl%u!ur}$rA%eWF|5zfj-k5B9K_^Q zGGuooLBfr0-*wm>+5&0U`nZMbc6hlr5^flqp8oPsXWn1o1k(%7L=tELG3GqqPNMDnaxxG^o6JwpeG@zh`{)S3qcJj6V=#+StuRob3{O<}J5%qe! zb;6~~e0O=7T$q9BzSHRK57*{?7nCn~Gx=ACm#y*cbnV#0XXN#@#i+%ZDJ(-!?;Q~} zFdgP^-tt0zakP`FcKi|W+6hAaNUI|BAW@PMbSZE%%IffQqCE}FiiI`7_w*IzzTf8A zj?SK!n!4z%exW48!NK9I>9`^v^cs5Vesc_Zdb#VrH&AF3kEB&C`DCfQ#k@Fc_Z6kt zBdg-=ekVF4B!qLryTv1#_-N`*@IFhvQkTNsJYN20HMFt=V$fF>5CG$J7Z)wqeg$cn znxJ;EoGS^88c(28L-^0h@q712`wyiPZG91r5-$VYWTnc|^j4vKe!BcHy%r@I^_LR0 zntY4oVUO_;>nHhLJ#E%CHe1wF)?6YyLP8Wbs916a)G20W>*YGTraxFIHaXNno!U+} zbL^J$Wk0G7?UkmpdxAfl`op8S3LvWof2h~IIilL$QLof(v)CJc_OxyJD4oQBx}{R9 zMsJnI<8;*d`qJ@LGd?kqQW_iP>B+WZ=oIz}oD1u@@p8A}G{Ij&(m42W&|;y2h!B&E zsJ6D|_G}}XkdWP-EuQtO$XWZ<_wHm?GlLO0&N3y~hliHcuhlb;VRV{*EmYgwZ?ZZ& zcP=hg8mymA7xdEGnvhWwOO=LQr}$f3de&i?nas{N-EWS>0)lN<+P!W{REdLRpAKs$ z2{_IDvOb?B(Q5^9A$alCTFCRd?}mU)O8e=43L!e4Rb`7t>g3THkDic#dfPrl9e4_p zb-b;6j6-+-S*=zsoG;e`kVMvK*^l;T!}+mGmUKGxiilyd26|Z77MC-QP=_K)BO6-2 zlhr0=Y*<^H5@CNh@7GsfVc`<;VNh;=WVJ6}cT<6KYh^+RpH0_dpr@R)w8dPrQ%7lK zRCu^(U)saVT{pQ-i<7cT$LkTD5kV*+hCT>Pc31&fxf_fnx5o{va+%1WQ>#>yx1Fg- zCuPr^^vFUv5BcMbqRe~inXknC>J6*NT&-F9D(4T$|IF75dW(SeouA<=oG+gHdn?cVwu1x?Zp- zQ93Tf%=J_8j52q=@aT%WVKv(4rJn|Fb6NBn{-sUxnEy&jU zCrcJ6ByI8tXjCKa?Ve|gHmunrzw=dL zT{V3>@4<@Zr(B)98m99nge5Rul*<)QZYA9?s99+?3U=>XZ}TX9_sQ?M+-3c6tuVU8 zG$`^D>Gf8DZO6m@H5nQ+y%NuaR-?iVx5fQ0<%X%}OmC&gMSpnI~Cw@ z>Nf~3c9ps&y55KRjD_-?tKph-nV_ve?_tej5U{4%+g~~@=rOuv2k;|r@;Durh{k8L zc06w*8#QF_EL#Pm5wUzJf}uEnez-bv!KYo}(NHk7q~<5rUteu1m>Jk`nCPe1nyo($ ziY9!R#gftgKanPu4+5Ize<03<<$n>GT~479!|g@@b*D z00gQ(MUK3MG& zT-8G3rE0GhT1KO_G68g0Ozn78(AL9BGAjYQ#qK3F_DOle8l2@ikxgRrjZjPY^f!JJi2sLzw39x4%Lir}G_Rsh3if zXmraR&z0Lf(FZ1Q-QZ{nARn9afCPQ;N>z#A!o*Qgs8F&}FjF|pilJqv6OWHd<$vp4Pd9YkuxHLUx_o&* z0vDK{Ho@m{nYMRw;Y(ENWiP$w9GnV$A7xtI<^`7&TK#Trf%g;q`TZNEdK?z}{~wl7+ad}&d$YE;{Ps&wkZs5eIA zA`SE{*lAC95LMVJQfCfWTD7kS(x-X9aYf2$>u<=oPIi$+;yquC4_sy+@moT#O z2L}ourTh~~20LKYeJLvrCv(2f6|u$|JbZ}FR?H_OC(or*t@LDCe=Ehi*m!e&j%Gi+ zN-{YSh#|no#QE?M50Bs@0j^@Ur&--bs7<*g?u{l!gIo%mqE7o1)j0j@9`2motQ==3 zCIvf%SUMf=tQ~?Qm_cWDDuX!1a;D>XdUjm=>Uy%vpVEWnlg9v2KZ^XPPjZyF*)q~S zo!h<>1@h{1jN{#VIHRd9o1M_A@!cVJ22+C5$NZ5*4yVse4jC^p$Qj_nv4!iiGd>Z_ zaxf9E^B>s)RD8BiOuw?Ic_hhpTWZ}qr172}AA3W`smS*Gw4#UW>gKk3g85zh$G$hd zdGFj^hBHdL#1qCT$Vy3Rv^bYPB2tG2-|V8i`a0eQ<*cg!C?fR!9*gSxpBi-2^hxHQ zw>Xt*?pZpmj9qJOFOZfJWOu0uQbd!HL{;=`(4@)dr@J$Cmo<%!a7XiO zkkI1d?cH4_FQ{GW1<9*>P!wJ8d77@WvDuGFuC`LSFzO%sex;!H3ALo{e}siM{`M{C z+qYnef&TtP+MB24QN#&rkkjQjRw0YSTfm_@0bF1Um#W}P=n!MZ6^He!oiy%#jxvfPveG-+Qk?9qV;cy`x2$4V z^pDF$-Y@NK9-Ow*yYvzxm@3C!cN>fi9z3Y!0WWzPJN=OaY@fV$BFD!ZZ^8=Z&spX9 z14p{tmn!8m!I@ra{IjiY4snmPG?4Z?-I>TUrw$mlN0y3{^)}(PX8IHkcAOdeP;&$N zUn^sr4c2RqkI?Tm8_et;ugz~JPmKn`$Z9;7Y6>Fp>n+xw!xHI9^Jhh=by&)W z4X}MY&e~fH+S~`128i8yWo81@$_t`OJjZlwlUd}mPWleIr-e_4KT@pvzkZF| zz15yo_4D_~rPs(=)>qu^Uv)37CIE9GFRWRBJ?>7Jn~Izlwsr7}B(f9v7w6#7@h)ph zKi({GkALJ;lP3T7-MRhy?w;#f4~jZtVmd<(LKYi~{4X9zKptHD(4^P1X0pG-6Egv*+37yTl+2&`>*fV&Xq^-fEKhSfP!UsI! zD2ok$`dPz#wd(kXL{=9nwF;L-Rr5WrC+0*8K2Ne;R%zWqMD&`MW?{73 zJs)}(^pezl>+w((TxgNu=$u>tP+7wD=G-44N5F~SVzyd}B^{0`Yn<*_aWaQ&VKBAb zH$H+d?_B?b`A4n$&AkUCmTtJYR$D4d5J1$$&7&#YSTy7fv0U z&Mc~` z!jnk${On)dE(~W1g9=*T+8>Ka=irdXoA9jd2F-s7;|)}#8ER`DkS#gpuodQh;dPl? z1BkeGr!2*Kb?p$_YRZ4ua&wjWsL=Gjf=mj z$DVxFAWiXg>mJK{MC<<6+zTq!N;ATbpRhO2&PYDs5n?4`VSRLF7+RS)&!X2|vExc> zjyT-_z2u{H1VE_?(%E0+eW5yAt?v0WzEEJTRfj+tqfeqNHo|M$p))VXJI$7ezCFVR zz6Hi)I-6TfHJxlWmYCG;@NiP`Q=rg8w$ zNaD0IcAq7VE8yhewdG7DiH5|}2h*_TxcU9^5Swr7_qFDn10jrUs)i|H|BR4f)slEx zRazb%{3(X%g%U<59uIlM^?2zE;KB(-L@W^QBQ8>FYi>T?OB(%knCbnHZ`c=BLO$I) zn;xG5sj3o*AV4m0AnDS$lj-tsGv8v1cRx*GKPX{pS75iFEh;2}&sb%)S#5}_UsnkT ztD_|bR9B@M0JP!8M@1HC7ks8AlSD>9LPo#4LdU?NXPRW>W3b@nc=%?3UMwWPP39=D>gYNjO+$yu1u z@iLmfH-RWYrB;L#tV(k%)@g}`H$X_8$VpbMr3_6d(j~d3Bo_g@7vu)C-93+T| zK4K{QQI(gVCV>4_u|pa*4V>!-@ZdQ6bEvN|N1kW>-><0Jz)j%Hvz5xsVqNS6V9bl< zf3uO!cdftd4aK6AIGd~dg_#Ladr<457xhaKV7Wg2O@YpZTB)!gf(236c6p^v;nk|X zOM;V#CPuCHS|U)>o3SuHi2vqSDk?(D#qY2kO``5Hbqu3e&JZj24-p&vJGs4LaDlKn zL&PT}*o2=5VVYKobRGNb4YZu|TJIr{6K!OxpHFB7s+E{(xtc#THH9^$zkgo7gXJd8 z7WRF0+cP9@>zMysxK)t5CPO1>g79;3PtT!v zi^6tm`-kh=At(e(i@IFs0poQ$-1reX-I#M3R`|uuyA5w;i3*+_uV67e3Hp>+-QB%j z{++RRM%^@l25^qI4zH`+XRu64LLWORm9r85@DI=_D=B*GM|^A+NxgWv`zW2myQ5TC zdl$bcSBzb$#)GVnu6&Ws9I4cd}pT^IWd`yCZ3(Ha@*qh+9 zq-34RDw}KT9t%%a@M2B9X;05Pa`JZ>ZbtIu76x7(iRLRcDneNunH?ZsF`-vDSaasB zS6y9d28Igz4~4iQOi5)>a1NknyWJv1`xdLE8b+G%#!gAsjBF@zT^B18WvL zlIZBlJuZf+SR!P~S3~M*e=Hl2zy+Kp~=#*;Hr!$%QOt zjPx~<`aQ1pv1ESUG(7zcINf8ahg?f~LI?7Ix!5{vl~c2kJ<%%`)70y&Sh zIV2?!JzOE3g7u9JlF*&!z}n?{BAhLkCMFL>##24NSpS1BdOoBA1MMb zE^UROWwkK~KVRBkAx);tnM!C1z8ZdbsVtkU`aYXTG&Hi1&WoR!oS1RSj-qSKweta;8_Fk`u^=R2~|Go~z?pH9^gzRNkC@Gu4W)8@W;{KohBJWH- z-W>C5`A)*A27lm@Z*(HgL@|nz^8LK^OK*Tk0x@O(5&sLXuH6wliEKKHn6}}3nJ&Nk zuB2Tkv&PZnu*>-IKM6C^JmlE}(kv6tW=gmq#Bh9_utzvPilw`Sw{O(>S?l=s_vbUrZG6^^I$49G5w)uddwXypyd+(X`fhNf67A0kb*9GF}cCgm4aq9{BJ|Mr$W<5V@%+v~A z_n2G5ez~GwNf!vX>Y6P+vH|oU=R50ZohU4Pj~j^f`ggO3+}o>HWphbjBpMNKm<*KhpJwFx^NO7sA;~#1dMC6oeTxgx1+U%f(4Uea zB>SVx$@_Uhc&=+sCHGSD@*Pj(yr|iqaB$Ry%_(c)!_vq5n@)69vyubL*jg zL61{sst!~2g24-NvZNmf{T{xbdrZh15{Nsi)5#8tdEjcK>`N8KIFZ27h&dvf>1gAU+O% z9UPFSS!>#wkW%sr;0RL;nOJ$=WH-LrC63`vJjGFR^JWNwyn7yRGUT5p6&rJ)A5TY< zYkWhj50K)bUYv$7vnvJC54e!<#Y<>DmH(vwuNFY!jj+S2GFkP3oU;8Uv4t+>q%Ef9 zpO#&%1z%p?WOerq`FV@tW=@TSX^N&&Hesr${zgH)U5g}Fy5;>JIz!**!~ z9+5k&*}%F-X7`tr(_ZtmOn#%iD1$Qjaha}?kk1t(BOSDt5ZhB=GL&V6`}*u4>c#IC zathY-tbX%>E(ylExHyWdT};w_1ZN&mE&VFbA4$A9m=cBLT<6dg2Jd{?x;dUllsMvz zGNp+nh6hCFJl9_cY<6#q#X}OpZeJxFnW1m}Om)bSzis zBr@_c6OmC!2#5@WlH=neBSVrCLqmb8U8QHBRkypfBL-O|L3-_jX>hc|KHH4oB(1NN zGcl8A?DNF{36oHAV&Vg2#g@@^`}O6C6n*yAajC{c9Z*ds^@mxP5bdJ-sf@uBsV=2E zh9y;n$+8vU8QCNiy@f3K-!D`sOSGB&{7K31uG!HCYkU&!Av*BB>H{m_T_D=ZFh;SCUqkqd1LXfcLOfsZecd= zXzJbS)dT@VAqY@o1*$+GZ(-qb&3f5f`L=A=!&c|96bQ#2x!YzJ{25>;VSC+HHa4d6 zIP6C>dvdyCyy{jPZnS&Zw!c0Cv?@%8j`8>Es`gy_GO!Yrhr#suTTbr|jU(ne>nC?A zSC}@~*n3lnn5vCi90&dIgRSIwGIx_I^d&Z_FCV3_lR3>}-<|a99kNrk~_&=*t zcPiV%g1Y~F?r-6T@qN4L+aFGsistY^&s{SK`vE{6<(A(Mz;Pm`czP_Y#NX%%Fkx(K zEInJdpOGu5>)xB}{kF5{sOU1%toyf2ySlrHBo{uPhg>PO8?i5T%8R*%01LjefPc%KZ0MV&d#CbK>rsbh1dgI*7*z|tj$Avq4r zob?U^XmQB0vFSWc&D|VL4ganWW-pFn>lWvNQe4Jt6iDk18rU!Rn*-YIify}M{hgSu zd-r?vv|pNEP-qU7UcS47(*aW>nK` zHguaMV|d%#uelu0v#vV6oA%a`=FJvSTR)?3Y$sq|fbAZcB8pGHP~bVeMJ zlb?_tm*!1M_Y^ogd1z%d{XOxeK=&L|G#FLcWy=1NZ~apNl@so^2Aa>POIoR0SD%>b zld7)~?|L0}qPc6iRPjDH#O6mwyF)Nq3>G&xclU=?d_OVKH|qf!PyMIA)(h9i;MUjq zA;9);xz&tQV$oO!aL!u2GSv@o8U@dXh&6{zrWwqU)sDC+hlx6Q2}UI)#GgkWZy{bQ^bMbjr*?dfyW1NM zz?q~GneN-6p9WNuT4hPh{&xApMG~?8T$wLtF!?f!ZY(}wvqs5{IxJ?<=>^0Ej_NTG zhM`vFU?&Xx`ic??Jg+Z%rr=ydifN!6njdv?TFHVawY>)K0NQlAQuhFWv@~1k122r8Ahd+5 zepQCoX=-|xRc~GqXYMd)2~OMf)wiS4^h|q^Op*$SNPnuD@_(cosk*^Mdp``SBegIK z#XhA{L!l)l(!F7^-u|F~KzK0;_#_6bCIeG7X-#49ydw7!o`wc@vN^Zu_kuKmbnkub z@hPlB3Hd!MZX#E{YO|oMH_S+U&I+vt8*OdES+p-$jwi5Vh9<*_)^jDQKoBOI4p?G! z^B8OHA0F6kreRJ`78idsStQBuUxAQ*S4+(l)LYDnAU~Zf)k=ACmW+1ZwyRtmv~?U$D$de~DjAwow$0;bSr&3UfL z@j0wquMeiud7T?mxfH&H#g2D%b+tWR3m2~dMgWht_E9yNVr#0%+Tm2Lgo-0u4f)}- z+qFXDv$!U+g~`SJVUnK8)Dyr#leT4aL@MuZj^=v2R7+G4L(%y=Os@9xaw4i3Ie^b* zxmFOV{ewWA&vmo2B!Q0q8oFGgL-!@`YL-Tsp|M%uKU4Yb>GJk-?7O5a9SLaJRZJX0 zg;J4k>0!d`!}m#H(;|Lf4M@GWwtN*8yAGcb znR!*USZNPP9pF4pYJmF!{#q@G+{ks$#B}=D}Y4|iD;QVJtjUEdXXL2dKtw^!ZWnE0BYrkG@ zR64%xT_4U+f9kjS_3OOltc}VO-4VK#6A+`qAMIOqk|zE!VGt~uFA#}Z@MLlWTq4gI zP^S-NB*8+7a-m2d5|EiH{GPG;*LtP7-ZF_%tG_F)mm+S%gOQTsuWT}lJ?Hu%d2_pv zq~zeyTxsN5JpeOTTGaou7I6^1`=?9%+unWq_@DOxHi6$q|8FEy==6vHn-{s3)izk7Yk zpd&w84kiI<7ux73+SdUr0$0}^F$wTMaCdWrS=36j@j8$N)I~)_^W}wKDR^8CSsZRp zEa6dpxv_C@;4~>WJDOjBMi%XMAb5i79&hV>fu6oHDQFZ=pWbj&4fx%!JM%BvaRQ!e zVGu;DR^1kKI03&~chR1cmG!FK z0Q0)z?KYKH&}GBJ_Kui;dU~2Tg-JLb=5)Cs^K~yNr64~qPpD}R3=R+PnV6U`!#rJT z^B9IF;d-J`uUN&B=ifVdU1gm}69nnnG?+~p_7*9k60l$1T>(?~#OPT^k`omR3+7f! z>$PQNtNnH_;)U%@yVsNXOrg1x1Sv_k$*_;#WsM>X8Iut2;??7<=}O<7*W(Jinedl) zzV4SKHlmV+qNxth8!R%;poMNtv>jurL6V>Eg5E}z@>Dy=Q&O8jWpr;!u z8e%`t+lvceK9PC|JnqIJ-fg^H$qR;p2#WA~!N-gpveMF+0#e_;^|tK{RO&TUlHc}`R-OZ^+5E}?>@yi!XW3J1cp3C)GobeU5rlTYhiu8w>z7l7q?>U zLrs)qz7=?4$V^KxsL-uP5}Zilbvl|=mgn1IfKha$RVni607x&j^t&H#`r(aPoOLG` z3x?HyWO2ab=~(XS$!NAAoIS3+HcYCt_Q6gub|Fc|^D-89A z3W|hTRije=S6JB4&``g4!VH;ZiA?XXc!}}X+v7?zh~7+xpmV18?Gg~R$tx&8@|xv* z6*1}^7$7h$*KAUHmRWTqulg0|Kr0h3ES;ju#W1 zBfF71&`F01N{t>79-Gto4|^N>v9w6taOP?J@SF`=$rCSee-g3OvO>t`KSAe3;%6K~9I5V}D z>o!1IG-}i2L4ab!VY?Dkx_J#u8l+rYKiqBRn%OxLD>dtXzPKOOIwUcKgI+$y`c7!n zn)1rm_`z#qL*p~NRJ7wW0I?FNzXqzEhmDit{2n;;x~mml9)|;>DiRxxSZH#t2iZ;p z9Gkrcv!2{3Y;2spPVqt;r$qv_%A;FkQKkQTc7 z)tl=Jff)o;`NQ{XKRSUNBU{j9s$qosjmKF>Gy?fWY_xd5NIZS`ar@)raR@6}#0GgZ z|9p|1k#SrE)%b4#6Y(CD8?YW|$nSA)L2ty5c2ylKxt59R^MIYoBPT6#%?iQ>j13|8 z{qBtXj>vhaP7lu!A?2lEDj9;#Pt;wCXUVH#l;?D*I1H{<(rydUhkSKiVUD4LuF-Iz zUt-GFo5C4>mrdr!YUngp+K}4A89RDHd$$e}&?AlP4^tpro;^rlnup3nM|TVu)w3Hw z;wwLS;sHz|mrg5xGF?yim1+T<(Ck?pg#hQ%rw~!V90THgtPFN* zHw)x0YGt~o3~1ETl2{t`Neo()nPRR!Y*vd3iCQP^f+k^GBojrp?Jqc?G|Kx|jO{{R z@xd5c#))Ov)PnU#v&=fk^0k42ZxQy?{1KnHr%mEh*e!!{&x2$>^9uaoncM3`MMrng zK1iaF*(HbNoEud;OuItU@PTI>{#};Q5Q%35vt6;zqwcBPyZqNOcpTZv^L5D-h??VO zh}kd0LbjQ;90b37fdT!?V6ksdwZAsg6iEOJ_C6Z%lWWjU&fst`iPx;7A>c{9B&-EoJF ze#LK7KV!}i0)o&?cCcO8mql&JE(HN^z_a@^5tn+}vlf6}>C$foW*7zEtnANLfMBQl z&E!Iq_z;lIEK@0-u1Mzv5)k-!9|P~|uHK=zjtiV``ZHHo6UD7b*>?NUUYxS>#D7&420_*kGht77dqrKg!i@Ocb>uM7a7wDF~fOwp3soDWm zF96AZ?i|SU#ALL+|K@;4#>YeX^K`l1%;6tG>5>;58cbqPMWX5k1AG#lkjn>ZYPwtR zPZP(%nMz(y%3P_-@`?JGR$Y-+1LV=^7^r*M@gA@!^2#py;C+y+h&0j)CJ}@3Cwi+zhhH66;IWkk zTasl_SLB=sQZ-l}fLMES^LYwdswK0S0D*)EZ^uvhQv5xbn~} z^N@4+W;Ud0-+f`tE&BrmA>~gW5Kxdbt3A0518WspBO|<@L#Y(TKOoq{lsjuE@T&mR z5&-KM^VSOFQe)VukBYt`pi04$2A>O7g|a0S6%{q#9AzKmNS-WKzdt-5wyew^ZA8I> zoimef97!?be^qiH>a;s7*ZSKz9?BBhxpx?|>K}~aThxEZl$NgBX0m3{yb~@8}X@N(@Lp} z?b1$A#Op(nURKRyAGu2`nu2r<9Sjd%av&6`3axs@^B)1+PAFRBmv_{XA%)lHqYe|< z&wsZ4bHb^hqPhtn*e+fsmAWAjL#e?^VXIvZ_3n}QMSlm$*+QoH)$^r4ZpFJa9=jpq zZg(Mj1+4lW<_&wL*AyO{-d()0s<{QV|%C z(x4CBn$GP2SE=8NV-2^2x4RX2HC$Rr)}jAVqN}F|i6$Zn(WO%5y~~8DOW%qp55h_u zPMfXdIv~da#%|6XGy}rcYLi0 z4>&P_?VH(?ymqAM?l%X_7We1GKx-=TgInEbgirzK^#KU~+if?n{>(}b_&KR3leAvg zVIPAAM09V^SL(>>NZoHvLeI50EQ5ASH0w)f1njrdrsSf&iqiG&)$lZ|y3BQ`LwhjZE1S}=ZD7@u`-nX0s_ zQao?^aMxNKo9xc@v3;mA+>y3ij#u%d7K*(*9OFMz+V=(^?vCu{8#Ft+OBmmtmArB_(QsP$zW`x-ic7pwb~I3V zH=IZ4QIF=uO0oIL^V1%`^2JYcQhx7MKAs9D5w8~`ZUZFM#tQ6nO!6~*X3_23+LK|( znxZSvWx$OE9VMFvyy`$Du`ihParRjQg*p2jph73kLYBs975-t6EZ{le?*70ix=Src zKn(_&e?$K8fX#Pdc!aUtECPvn=~01L;PJc%oZ*ENt>#wvOL$5wXn1e+v+zzdiC}JF z6ZY0Tdo!IXrn_Lc`7RiMh17D8N%bn7v$a}@sBbY@WP_%&BtW~PPLcdelwE-#aic6nX-#m+_nyE;9ftxizbOS z_T1{jl@ng$_2z@XvKImVo-6Dfpb52(+jQ)p5ni76`<{H`*uwr+59p1Oo{p6*9{Ko< z9s(_N>$tM;yC+WX;a1r!RFa~IK3HSh&f*97=onWMI_@*z3cQ~8oZV2uypsL3PYdsY zU11%kN7_Pi^vt~p{3!)+WYWB;RPqIBAMDJFG0vjHBHZ^CO>hiekJle;boVkGu%dad zpEjk!Cu973t}zfE)HezB4aM;J<{oS*KjV{1PrV@_L#1Sx-24qGwk1Htz7EMXa9QPW z!Jcq$%BF=!Z{@bjnFQ}_@#~a0nrnZz<}aF^3P#9L|H$96xLE6A3nB%w?3SAFdpkfM zZ{{z(dwQnlvt4VOv3mYwst=O|-sxKM1}M5gN3QVoXObsS?IVtQ301doqIiy_i6GlHRoJ+y4iejQ%54&y(d4nChYNuXM%3=j7J+=AlJ7|7rnR&q!@JfX;KF=>=YMDU{C?@R%=CzfsF#wYQ;iKfX`{ zdSquY*bY`i(~w4e=EM)Lvm8;lA71xMD<_Xmr)t3)%{Q0@9yBqNNL?F5D-ANLG}f@x zw&Dtaq@BPae!f(g*{!>rFY_;0xY?BQ$ud!L2P&l&BY%2tb|~(AsoqjO1WCE>!vs2| zeBz3v5?hT<(4l7`Lv24yJYU&>y$4{89~4R?@nMpvqO0L^=i{{AcO!>e`SbB_k_MXS zp-2a6XAI!6?Zi+@{#sJFjZirZ`642^cO<9}wVb8UG>9O{OW0hZ*I_oW4l~PpU_2wO z!+3;`PEffxyruGzP3mYA=(tqnR=etIJDtbqNY{+Wl>-fQvMjg}2o*&xI^irZmMd$N zb2YB^MQmS2cmnNfycQU=f$~;`kWSoI7HnuD1IPI_fw_k(QtGmV8d~!foK>e^zyQMZ za2*bISlsF#HaC{$xy088Dz-mYIjh9P`;tN`BzSH<1I3TuRIe%$bkj0sJ~DFI?;<=3 zPol!q)&jssu1{ih(!;MT0MGe2ipZ4=uhIuf=Lf!yzN8e4E-eED-wX9GP_bLldFz{z z7oa&-@mlkXG|c+`QP}x&c9pNFl4dSQ!C8LO1p1ibnVbay`s~9l<+>cxs#InsoihEG zn%a+?0%n^+$G`j^t{SX1Y#fIN(G;@LWxQpRJB!ig11IrKYENaLQ7)=B!I@KZ0Xzp4 zkUv5f3c-*aDGG_-GL|b{{_uMq%9P~8+=d}V+f*II!AhiYMdFe8u6m!gHn+}aXXHz9 z-GP>{PV>=~4=4H@hxPL28DP*xgdp;F$O;b`9LX>DOJvo}HptCq+w}SLvMmxmT36)G&$cTv(LkKCN_HQe#JbXFM=dr)ZsVK**A0O$ob zcg9M1{9|(T2J5A5&p}b7f6cwIRL-;6rmOoWNDunE*Vtf^L~HVot+)oagg|N640b!% zGoUqo2^s6rgHMREn0fyMl>X)VMek{nf@g%;fN=)eTQlnx9d|zkESxqZ5YUKLlPH@r?~1 z8U2mCrEsKyAe65IEn$x8rR({npr>vTO9cHQG@5r)WF(MAh);3e*y~nss!bDYx^@Ut zO2&PcXV}qCP4N{`8Zj$$B0lQXtbL+OLBB;ghi|E5{@266C7@+y7g|w9EGH5O05^zjOd+(tG&HXmZZaI=h}XiPEMq zAD}fT0@r^f5yUv#pTntOJUZ@v0e6W8=L8_UOdVNq#D`G;|877)sBC$Em780egsm!a zn*&uZ$$ymMsG)4K1L`nb*7u#n{7!Ih>(1@ULStqUgKvY_qijqDzlZohZ)aYIwmb}t z;3UDEC`p38kLf~Ft!c*tgx7^a`<>ielML?MUSS|IBehSPThaV1i5Vl$ZRj^zq9{y_ zBiRYj&|pq9r4vAURLH@2MP_V+spUJV$k^D)$ea-KVwRFMlE0io=Q%2e%@#CZ_=7yI zUB$m+5^h*YR@-oJ7S;TS0rkgPu}{S^puo1Jbc)!%R;e$nfRGR160l4U>Ojs-O|3%P z?TN@m2=v{h&K59GnDE&u+b@TKT3Qes7SC6~$8%A~gKhS+)|lV%&^30Om`<*RGBc1} zbO9l9^izUXqshjHlW<0hC2_{o!MUb+yZ?u}w~VQ?>%%=;ic{R(ihGN@LveR46nA%u zyL)kWx8f~MDehjZI23m{i|2XY3fN03%BVnO>t!M3?_}p&s`!Ti;YpR!|Inl1AEJ>Ii2;Edk2X*PAlGH zeApF~@xZEvumXKn!oNHBoq3~Fa`TIjmCiC$O2%a@+iHyx@!yInZW9ZZw)}aTNT&pE zjzM~7Of%NCu7^8I3>>Z>^RKKs3(j|86m9k8x(P{&_5NN!wHP7-H59})R2^vVHJk} zEm#SQvc;cd8UxWpVlZh9*h7A=UiS#pV6yEt-VxS|GQI9xktLfzCv5qsyI__j!I~_t3A0`TNyVNB_87xma$hNZuC+-;8njy>6JF zHaC4!Kfu7j@|=U_P1K^tuyu2(5z0I2jLUXy-13qf*Eeh(z?)u1??lx8 zN}ExN5beh_Ati68eXJ`8g9Ax=jn@0udqCYk24Zb)lMMUcmE%5kuUw?VjGfV9;Mp%- zMibTS(Ru+IkJQ}1kJ}FzFOG$_^4UOZq;$Jn4OersPTX@)miJg@3~@*PE9%+UcdqZf zKNs07$KCuF6$N6#Quyxv-sw6b>g_x1PJqLl3q)YRnJy^B1>@h$2I2QOVz$Niw3UG0 zZ1saj11X%HKK#Ylu9n1!08PMa2O>Sq2BY}+`00UKt0I-UJ@G-KcBu8|R5>+JW|$Di z?)(XJw-JTD=7BJ#!-a~S!D~-ON!f3G5eGiF22s9j z>NVn`>kN*Z%7B!S%Q=1hE}uux)#YraeTAoC=Ei0{T_QaYede0ZZ|znwF!9NE&qGi* z_UfM4=gTia^jQ)3&8Vlp`7s-rk3AKd96G*yZ(;x1UcllE)8y;lmYQY>S-xUgY3q+l z$hTP9HKD`#HcbQ>0M0C$JaKB12K);yNk(krvtOXlg0tjsb_@3X_EJhT0x|LFQe_xQ zDa~NTCdFXgtH&NJ3mJRScrwGW`TmZ!9+6pCCdrR-avwhm84(5}WlRE@F-PKwuxAcf zfn*l;y_|v8rsb$o6S~*SAKZT?dfW|i7sOzcr#r7(|LXoGMErpsBR!qv@)hDIE;5WT za~V~XJI*bTbO)^Quy4jJ_u00LBx}4acrR+C*km?gmhA2ai%%&fIR$_$B#kqeuNCA# z6*j2q(dYT7+1+u+;?5zU`C!dmN$56>%aa}B#Bq#&sY4ojvqxo{&~((3x=@m%L1!CQ zJa&J(xJ5GbSbB1<%OI1qI&g_Ra`qXuu zG0$rA(=lV#yifR28G)1rHqBnFT+KmLJhhUi0{DZly*P4v^W@3UC_=H;s>-~=G zJt&2De#_Z*9kFlWO%C8BpCA@BJF?8@VN+e`;-TVmc{JQbCRxpP3^8>ebYy$JvD=i^8vTFbx?^)e6zhG0+i6$gaq#MU}<$&K`MV8ihuN3GvA z?`{(Ql{zXK)9>jq-cO_{so%wX|M2bQv{GhdH<9reIGT3RI?amph1U8KhB+AgFHrz* z^&1djJsyk+oT}^tB#c{^&)pD}aP|ze`|TBoA$?%$<9QrWFetOax}Hv9WWMd%3=Cs3 z;cFxqFbW2ov>sGToI}}S>|9C*t-3~2(P+_vehoX(Za@fu>@FI0Fd(#@hqj4_ivAW)K^~ zwtF99pC2}35X3LZ(mr^ylRYdE`{sO9&lg8*SD4T|;LZF6ISI3BfaYroa+YFwrufgu zG*dnH4nS_Zl< z^LnsG3M2F0{C;!i&sZY5`b+k2Y*%hgv) zgC8%#2B1eNb5?r>3uM5dopt}4&Tqti_=e^!2PW@JKji)gYa=MmU9!>&J*Cb>mp!-p z%4UTr)b(Ou#~&)rPM(BPeJj4w+bjA`amC{@w%ZY#-j4rz#dX#F_;%+sRJ++epMaC$ z!oI1S!a&3WeR z2=q9bq&BNDKC2aM8v!k+2hO&zvRmr6iWZc~W7^KwE{aYQi6Ln7b>3|RP>8M<3X+$h zJ6Spl2*?eYcWBC!ZZ4v!6il8nnZ=i~l=YXSaVfBvi! z<6_Rp6b~e+o0u36GG*k4_%0$byk|#|xVt7ncz%Lk z**w{^)FpR-8DNtwB;4_4oK!{IDd+pce7QQ4t5LPiCl8*8yz})$8aEvNr(Se0Lj$(! z$5hpGDSq>o;;-7^@ZOksXFP-6#kg%mJjvV&$OV7ozmi9DiWGi!J(zei#2D1Dqh8jw z^8voS?LYf~per0m%*BL${OE`9L_@yTd-pf?VY9kx+$fp@Z`E=aTI?Pk4|A^Xk*so@ zk8g0ckb1f6C*RheL}Y#(yA|hMDM5K+_^gMY%`eq#ECVIjuh`Y?4-OVm@Z4vo9?6tH zu(Len_y-In`r5FV44gXh(-q`)682TDJg|0`nE0Q0vr?JIe+a(5OwUMUy$Htt{hEQf zeRG07Hk(hQ*@dAYoL3QE2Y33-^-4$+{VJ;&!S9b{a*2<3!%XOxl$j5(kc%Qfd_;J( zcmRXgeSk8y4R|9C&9LE4Yn|5VEc?(W@c{0DYyV}uZwZy5*fiV6M;~w&>`|F(AJ1bA7v7>K2nH8+nh%YZgn8Zf%cQv>wXAs54I6F?+Q#3LznqhKG1L zuHubGGI_1#VYl0YIh9Z z8SJ8+Tcg13w6dh3M)MZU@5iBpxD#f<9~i;?8`o>zA$5T+M2X0s&VDJ&5B+&BEMlAk zow>Q$f)fA|KoR!MBhLftW+9^dDUVowo!O0VUo8#eTh}t#94WP+^KCnEk1%1s99**5 z@~WWkGGvp-X>&2e#8t<{2-SKI8#VjCXZhT+ng!GZI%jYm6p$9K0Y%F3eeb1GK4 zQ+OvfuG2eRkjTc8d=tXJWMzTjs+>8=P3%6t_$ke_&J`9CVrBsUh^e0jy2nK8}$3%!|UruTSR6AOMfn_7dK&X^QzTGz|&R7BmOwDH)t zL5LV9OBW|Y!?rPnyCQFNa+V48yd+(43Qq_Hh~~sDdm|d{6DZw=^597*KEt(KXL-h^ zqb&V}rn}C*Q}D9Y^XlF^s!0Bbt?1YovBB!m3+uF-!vl>PkcJBXSLpMtJ)Fe_0Mpnr z%%3X;j;s2u;a0fYTNvv;RBEifI$<6fw3Tysdz64b^<@pu@AvcXmhIB4)ex_ zv$XecMTlwn0s{sACVT`CPIA|DydyooxsEZwD>MFX^!M?7M?cLJ@39+83JXv86$k+0 zAg*&cRzq9OFIZb%f(3NApw`GL7+wxL^4;0EHEcUg71)*nv$$FUUwMM>(uGFFa@y>+ z$!nR)exnsym3XUb`!MP{{zwA61DRCY-SeR#`6138L7n4bYA!DRcK_lWsbA@e}{NcINT1 zyIG!-9P;n!m(CHdzm@@Nm1s(Pz2Af`{H8eYU1It^eEXUnU$VPf@k=JZxx0n;w}Y9B zTNcbgE)G7b%M*MA$MOdsUM(ugj|IiU4#(bXaVk@Df1RirfchvoD|xoo+Y{rdZu@7c z<(6AONX6+}t5H%?(wDsR1fm1ZY$kX6I&x~xl~-T4iWh0WD&9JwApyCMZ@(Qx`1nv-CTQk$oDj*FoG#sum~AH{E6H3{Q{ z!#fs1G|asfL-=~Nqs4(3!Ve=DVx60hN|6)w-7`xqZhj)(UM`rmt+6S9N#Vo^`Y0=Z^D(#*C7kN&GB%GT})F#W2l?hyUCr0dlBd-IjiwV)$-)>BC03GDP{ruq#1h>L{2?F(_rixaIP?7o3=GF$`|)Mkt#^;W z;&@B6+$jR{f_VZjJGG|%GW1Nr=Xup``r!6}`H!GH5M_f9@PXaB$4>J0C0XRU2IS8_e;6D-B|R$-An{Rdh9ui z*$Uhy)DRGd`>=a+Gg<_a-)@vw6M1Z*TK#5v0Isj^A`i0dAI#AncyBMJ&Jxq=O@XEm zp_u;ms97$PM_M4?wiV6?G}WRp%hfmcd9Dv{cN63@0$P2$AvgEqXAWl9kGHM{74v7U zmqVvq4_En2kC#(ZarX48eI5boGl{0fah5xDa|}L9wA@!- zcc2E8>;dAcsJ3&t{|qI~3rfS~!#=8eFpM~Vn>&#l>s zY}2aFR%@puC8+``paik!FN|&=f9dY-hTbdw5{^mNy;N&7P~&~h(7b&8{SBb=SOt;%mS~#oOH|zR2qJ;;G3SFB5wgKppo887 z;E1HOv=)~Uyo}550m$5Ig#S!S(%uhXFn{uG4v2U$rV2cR26D)n+vbe-Jr)MWcUW_` z!J(nO$=b~lRPwJJ<^Zw@1f}ZAu*w2AuY3VohRoHr5av8wI-K+zoLImOiA*g?%bYMpcd?D)qZ9>>V z)Sh5N*N4Yz5o;eFqkdKIySG4U@pQWR`1q*rb}H8M40K>13+DylEegzGG!B=NTa-)) zpzh$a5YKl#1Hi*tm#=_MnOb?1&C%ixYJ%Sz7-YPQ)wYvv``^xUL%TxeG59=J!3Wh{ zMfzRpK<3}}zG1P^()DCw6CC%1ON}B}7sm_ed$%_U^6>wzL`g+JJBFPoQ>AgXBUSijf=y5WlpYOERwy-E$1|>eb7jG)5U{F7?z-o@rn?XtHEq--fsB;<&ng$!7 zTn8p2`AlvK@6Nx=joKz zHehY}Je-8y+Nkry=lk@a&#rx)qVIJIDq%mZ=d&w!hvIj6RFANT zSnXdF-Kh`=BA^Q-lcSX z$h^5bi157?>u~b&YX1-B+6{3BXf?GtZ2Q-EPFkgp#l^+DzreA$KRvjgyzv`cZnVVi z)qA$AmX()J$xk7@%Aw7@6Sm&C0wTyh{Xv0;kDXyE6B70< z#3m!H>vson^uW!w3l+<8bCDnq0M`e<{oVi{zG+~S^?_Qb-h2qE`}`QFFq!^(oxGXX z=5%Bsin+9D#SvjZPDb{hN3wj&eT ze`^7<=we&LdsF(Jn~0L{jUX6+Ir$2B&VBATW5gn`W=a(Em3KyCe`X)}5d$Z_+JZnb za{chk|5`uQiF24he+tl35);%24hyaCF*Gs-8F9+1qX0b9$7eA;Bab~AFZ46aiT zVvNA4Xq(H?)ydM_PFeX21Q_@@;Hq!WV+4r2c5@G^x`^U0ffudkqOtDX5s%ZiI`!&f zJ`b2vcdJo&4jRCZ1TR(bT#pL_I`v$oF*M{U9}%r8%}XOoHlRJ5G`iNabCy^V$!i!g@ZFc9Qv3 zGLfI_gN}{{jg0rsy57jP4G}+yR-K5)MkfFoakfe`Klo%_m`W}Ku(a4JesUn;Ln|K9 zJV_}i#DJZigqis-=-Ju*+(jiH&E$c>3@=x&)Ve(Bdh+<|2eYi({GEpI>SzTa3X#V; zyVD!U|EGYbmpc!z`j1Q6-fJxNjthCsUnu|-5k?--x{fX^DU7=Ijd)F05x zBN6mg)-%f^(UhT&VMPO2S6x-W$m}1$qq#AnX$3sg+>xP`87d+&)Tn92?8Kt zB4S`H!iY0KLCdQmRE@m(_bQ0f@m!I<)K~O=xL6Kay8hA|Cnx&#YLz#e1Bf)1qHPw9 z79mJPDsy(7=T&V6o8IfUJIMP8!84`WuIpn0#9R&q&7UZ)&dOK#T%cRR{*Vz1&oh1W5I7915zavE=z5oj|YobNk+rQ-iaCm zw;sIkN5hk7u{G$2K7RDq!OJuH^uT%~`e3i9lPYm+u&8IlED`#%FXaU$B3cLx!;W9V z5Dd%>wa=9Tw#3NDOT}sph}^%Jd$#-_Af?6DBmy%aHo^}gH5CpH4iN#_RQsIxRaj(! zxPJs$H*YJ6%+fI8A6P{imw!GABO1`EP`;<7B_|h&^@V)LhNgpo`X`Qq*_w$Xm0YyY zZQBog-4;o345y9P)KthYPTV5<{pZcjp#Sr|FslC<6l^$w0MTEN(9jM+Grn)%IkWKZ zcJuyUU&QeJKl4WT|CGB_ARtLo@MI%OSEbCZR+Agh+s^8g_%cv2G`TBeWRZy+ltzu; z5wUaK-NJ~gnWO1;sG5~pn4j{*$5_emG^%URmt}{BBfb$P3zVoE9oVD_MJKIIl4je* z(yMamPYd`6Ggp+kXe6cg8ykf{wJ|M3@XYbOa*Qo{fuxz1+rr`mo0gcGMTYa0+2lH6 z=|B_K)=uj=5egorVs7X1s2o>jnww~>bnu@YV%^FhuU>ck_JF*f{j*l+x|N@8OdNj3 z$qq@dr)7or`Dj1|ii)dP8LTy9SC#XxAIpZ1iL-cX|L`Ce zs~mZF&Quxkx(T8?t`zw7GZxtPte+%Wdn65|_bJ)^-r~ zl3`yksvM7$NB6N4%_oaDr|1Imhw&Mo{BmV&pGrbDWnbGh5|zA1P6 zM%8X*ZKQtGZkINlCDDbRG_PfS`oYC;SZqOVzO(leO)TTM;*94iMRR^xcczk~2gB*@V=C73bmQX}sO9mxY6)_tPc0N5M2D(Q;yKe)auPQ+->?&!g5i*_F%o z4a-%_hFdKi!!@;MNmYz*%$>)OXEnXJi#eF=u29sP)Xa1+t+om%3;It`_(d9tT=ABQ zIT+?y^6gl8=%a#XEjSNl$pW#FhRS~7zddv6&0$q5n&yS%?$H=4c$r-*t6xG7CiJV_@T;3k%!lnYwcLEEC zP3AXjc(+-9asBg$agWT+Xhk0VE2mBb+w^TD)NIhSTD8_`6oVi*1$bm*BgiXc5RLA) z?iTv5TDoy=6zHccho&=}7J!-f-UlNg4J%3vM zGEUtRm*607rd8^pOGd{Qyezf+WcjDXh;qyhZtAkK==R5ZXls&{;iW}?lG{OX&5leN`jcwXzM04Af(_Xpi^6{dE1<=qb)Hsg4YPsO!I?wp@FqxU~z;}Fs5 zf~|Q`=4ME839sWMpo7k5hQo7*pDEOp;KrTo&ffU8dfA>8?Ze)<$Awg)wNK;4E6=NthkacPNCT z!Yx-fAlEjrYz^`KH=T1!m1k2|rXZHyo^I(&n;H&|>k-e7C{V`4iwh2M+RwlL7Bzdd zA4Tv?h=<(P-X7_f@JBhCAMwW-&fVy4S8_u#-1_dIEmCEea zu+jx+fz90oKHrnCwTzcSjS!mbZ7$;t#bSH2?w|L>Yd zySK~fg03&oopasIiG#^-9&M&=@=_uG)#26R@p41GA8MgY#2Q}k+82fUL>0;HKb7x3vds5ovXXm7I`eF`lK=?TCk{s#Pf)u=W%1ef)y5 z5k`k)^e*p`f;|IBt5v^!s##`SYG8t@;YhT!x6WTb!D*A6yqJOk`%|t8f_Gzq#mYcX zuiEPAh`0Ykm%Yu{+V0^wGJH(gqKaJc0C~B`Sli~lla7FUY`w25GM$rj-}EpN&! z%s$E{8$&;LXhE(<83C`jaJO*e48yP8w5J+;F)}t1xIPh|Z3_F_O1s{3p}vw$pqvF> z_sw4e$h!Lws^Mm$cSkBLO?wNy=wE|J35}2+=|hs+>6#V2zee5as5Ll@REO!ak?;M+ z6MCxtzA5dp4(Ht|cyKC^mo2xOElN985Mz#051xCEq3uvqd&P8MWkjSzRjreMjZZ`A z8xMJw91I4HoQt0qfnL@R-W4e59-HX zK1MtZDyk0*O9j>I4ldV9;H>EB@>VFE8u4r8!)>K25=>d<;dc)-98a zVRcAm4`O(S_uCQ*QB3Nn$fD|&h8Gk<9RLisn=K@@e2Q-zzsxZ8jRWfMGZgi$mLLi{ zy*t-F{RTK8$0HOA&lwyzs%%UXUS0jIc~2Q^kiU}O)z)sasW_`h%hTn#I~`06^>Xek zDZEM;jt)Nxnb~_TSYCaevINKICj#23Z!xuqZ|cU?l`vOP)ryK}ztA<|qG2dX*_cvf z%FJIy3_TBB*3T$Tjo92*s618)b11YCMURXAPNjT~B{ z4%Z%&vS2O=bgFPu66&btrj1oMm3h zIEqc6Iy--b46i!Y`nv~ozFqH06dgW{_RQ{Vn9?vm7~nIy;|_J-EzD0&%xDi~ zH@IeM+3J5Pn^i@|H>sLL^qj^0+h8xLB{?Z~_+Ce9=7f^1tU$cl^G(uTFU>&nl2NI7 zGuE1w7@d>dk2PF=n(yH#*UMJ= zAd{B{RV6RUXq4PxZwVLBQm~V-wwh$aMCxk5N^`VM9)IvDCP^V4MdmPKYyYVA4d@l&JHzfqInUn4kbthkaHa`gS zTXaFtpR|a%2$#aQ1m7g|F#Z%#v8=HxE22U2*=Q682_7UY!($|b+E?zEbSH5k<3)kO z0YUFL7+Bs_g>#RxA3WG!Zjz~TuNX(-H)yGZcPgcv)VLm>t1k=a+lWBBy1UcW_1wrM z#b)O{J{cNaZ!KyLqPFN(-hV4w>DM}Q+};xZDCBeOCuR%pKsqURJSxd+RdE-r0zRk8 zpM`v|QP1FBNseiZa9g>2B}Q zk`n7{(0QHGT$gL+Vp9NnXvMOqY8QZ5BpOw`*LzAFHp%vh$aLT-6&UA2r9OyozV&0QsgAiULJ+tq#15W(qF}Hi0TX`MTg@X~#3R>@%pH5Y9I@OBO%$xEsInR1xh9M#bF!oW+;yV=)GW z7&91MV!;5iEh<92=W4I4i)U(UrpTh8luAOqDmj;*d()ds+XV;DJWtA84y24Xb_mgo zFbO_FawcaJvSr7YgyD)W6%`cI|G8$P^<}6l><=*sekij-#?oVow1B&2sj2HDO?y6x z0Ctmv5F=s@T@le*(P+`tTR^w_ttBJ{8A0)Q!T$hx>-#-DMBC-e94Yjr7W8BYI?7!i z-4?VQ$Um&3^n$cC_8HRa+e^7Fg$m>Q)|%)!!k`SVf7B*j7Zzxe>-!;dfdAV%3xr60 zNCyd&#-#@`=&3$4+}ORq1sUaUaNoTQ>_6-Nf0GMp6W@Z0rGM zf)B}$r0=}tkSkKm2&{a1oW3?fLVo)`pHDiUUOz?nJ`D}q1xFEB*J^a$|DkgwvLX7w z&PB5~Kx_g}Y;seQ{%KRHk60kMf!lk3dX#Vpw3GV8#_KLI`8+RgE$$kWspQL!!TYvs zU-9tY-}Ka*q6gV3G-^2)1n<{b4qO3ZX;^$tLD|Q)E5aq0UCzBqj%D{7Ce?)Ud*t~x z(-&ODmd$lm2jE4hSL-mkrn6e=b$SKqV>7khcfy&Xqjx@SI8N>9QHoV^JL9oiD0K=4 z)rW-vK@V?@Z8kNTYMEM_eQya(!kVnSG~HwvwS!1-)O?pMQ&au>56DFGG55|_SoruH zCd;~X$&ZZ=x9?NdmzkY#{5|fY)L)L!v$Y_Qh_)zH3k7%};<1}A>Nr;@)P9OktCWq- zI=v1->`6!>eod5F#QYX_cF-cB)ovh+SP59Jc$!>@NNqHgti;wyDG1 z1M7oWJVJKvr^27$&cp~ca18|}h3H3BbOB%AvBuwBDzYT>OHC)#%>n-YTFvH5wT{@E zW+Bm;!|ElWQ6m7fy3h`XzmAHEEbycPOe}>kw($^z@N;X-4WO_N2fB1L%2-Rft(Pcb zy<^p>w#FD-vQr9~LXW{w&^o?7--QA)I|_@61aU-&+7L|dn<}+0G5%QtJz#@o#ICEa zJ8d>fVPww2iOy_w>RJ2s>unFzk@`Hfff1MQs*vT6NgQ;1&x=7r<(kQ7z#P@VQ_V>t zg_<}HWMBojwVSu@?(W*`)(@M_6!XDoot6)8Ry-~;K;_&_1O^hCq_Rx4+D}wkS~duV zb(cJkEQXNFzrF^cG$SsK5gCqeJI;xQW~f}fYH(;YIa+dlpUr}jii%4guNd5yLcrqz z*>8FX@aqHpeVJU|tqw_L3k#ae63;O2p_;GyOtZfd(5X};rrJRlVuaC)g|>B@OOatR zmBT3^@1>=rc#`q!@$vEU5}12#h_=(rKfADTx!L({QjTo79JO8s%4GjkjRtdBJ0d(f zIjqI|AbELt`TKW44f|SzdONhN_oib%(G?ZqM8m@1b}~Ic1fJREECCa1NO3Ii_pa%P zr$WDto4n^M&BO!?>^Zp8%v3I0NnUg$JUoow=fbS1a^saGm=41sK_NYar!32gR==Yq zGKT2|6~RhjZ2`1(HfGDsPOyYz3vG_}RmC6dNu3aX_oAN;oS&bsJ3P!!^nNf6tW?Y5 zcCY(XDD~IwtQYQ5ixUCPZR;;-FBKOTF1zWAvg6?y+1*r?~X1fA%YUcOAxe$(h0M;8=DEhrakP&b|Z_rw>GmmEUcJx#46|c3eU`6d30Bs~`A$S=PI$$ChvzRVh@%`>RDBCc1eYo1_a{2SwxQK*LV$~_4edOYzf##&+pw6yUhD=!j zIJLkvzj&-$;_q_CEg_xaF5G9>=Gbi}Rqj}vGXMHn)OqfVnv7`Y6+@pj`aA|JG#VPN zU*fR`zN*sdXr1Y)%DmyZ>0jT4(5ScQa}EfG^=@$Q>6hhyvLJxYl1 z=dmPMzgf?_;MW7ODL!Ka_G4fbq(;1Ai*a4g8##F?Zh^9)rXp{$9i*JUf*CUBL)VG! z+nS3!-~Hg+o-M*$Bg;R{`U5}$(PDZ1wb8slE?;h9hZK*o%CeON3#6?A=o|%<%G`kL~1=jg@LQd>G4%n?{ zhI1Q3l6Ttx2>c4ay1QyJ$B=wIA%{!5NvFlJiK~?vn(=Y7xmnNm^U^E>2NH1* z!E$L2jFY@9q=tW^$RkBeM>B$bB}1RZi%4YmXKW1~B7=fR$m3GSNsr-+L*Bjrt0Qa2 zP@;b!LBdDXb&q9?vqiUTIb8;Y+hj5$rIL-S@BR8d>*54qMs#t~a;`Gsh<2vzXszuU z03jPK=6={@a$k;rej{DIj%zYjdLrN5zr0XiaEE!9vuoR#@if{yXYi+Ttp*jo1&{w% zfuyOqXwsIksmN5`+Upf`fKj-8oa>Xk(p(z}4GXJOuI?OHWrXh;P+pZty}vr%Tdhx- z7)gwek6a9qAmD8{nby>3E3Bf|&)slc^9thC>T;1yIp!;1bd~_k?bEqkZWgJ`bs;P? zC%GXhcyycR*K<4}33h*2^Bym;vqOS|+08B3a*=V_A%b3Sqfhj&V0!*ye>tR>x|^|E z+BKR@oHJW5x0DJQw442E`^?m4tCX>=nHr`ilEs!R2bwCA%2fP#tT=&olF#XH16IAz ze`^7NKNp^8ge0_bHycc;0x35gIOC=&LHua;Dj%1g*$n8*l%cJsD7qf#y`USuXzNZu zvo9HvORPPQ=c8B+KWWh-e>OF>)#*a|Bf%+5S*cY^&&CGT@^|W?MYx&KW~~(}j0V3~ z_G_c0BxgcamQ5giF?sg~C#dD;yG@QwlwW4W@|hJ1xnwF>1sE!jZ=CWAZvX0+-|@>iQWwsimzXtu&v_J;_UwU^`wB^ zORqT+GuJEk>+d3xD}yCbdE9u6nCMKNCJBiF1ALF0`8Wh1>E^z}Freu}hi;REzob4K z(p=DFln+FEGTw)YT`u7?G@SkPWt7|)1%uuKE3yO@{gIy;0iP?BSNN!h9O=qJqjG(+ z#@=X)@C?;G5L^;hp1UUui#~JPiXXh}e&TJtNxpKmbqPO;(BYOwJB zb79Wfrp~BuGM_(gtU{Ik!;Jihd3_z#>rxZ@moKFCVZvX*A@sB=!UD{(-nS1T$^2ujaY>6R+7zcyGgbU=@Oqka7!d* zWT8H6kQ`^nFkW}t3V2OJ zVWHiv3JP2c7y+1tk)LA2qQivHQSeVs2S6Bhv$}-7qKhDoY2Hh%-uZwi5^{audjyad zfwnn}{@X|zQRA#k`nmoMSlzFn(8o)aXR8TvgZRKnGosZXs&}~DDa$AA0}c0)mLoDO z0v0hV<+;P@@+W55^Bp1~eFlF#4cOU+-#PS)u*Dr<;_flcaqzpJK^D|$dH+nQ{>?=R z{TxNG(C8>JFseK8nPhHAE=yCMMf}48}<#(V25q8tppfXYcwZN~c^)8uj?j$00AS{`CKR}$Kz!)FQ>3dPO zZ;G}z4%7(uh1VURWRmaBFP`v2#79OtVYCB@&B1mhJW2^@GMToUY)b!BUxCR1%&0hI_N$976(k61kcMBcs|huO-)ailfYL2 zv~yFd>~fI2VKFi{(LXJ4 z-?+?#Z<~n(zz{eAukA|1Z1{2;`C8Wcw60$9IuKEvw{E)x^e^8+nIF&%(-|+7y@`n&4V)v0R2eA{o%y;%zGAbF}jY)-=zJtKQD&-1e`mWJ3J|P~D-NxUH>2`YuHQ)B74n6pc zuVrk-B@`+QGS$qkb_Y#VEUe_TH1kQ@8k?ZNKojGFwlm_agam6=K3T|+K#~#yb>-hr zxU_0zPQQf%kY!(LE16H&1Ul2z-dgI=KYs#!8*@ufBrzJ^{+VYzG({3kitl18ea7m$ z2`Ilk9P<(%Ni{Sg)b;mUzC~Ykp13}d6kB^T|7zte)VVza0A5f=ycD8vs1mwUrhC_8 z9$m#VzUJ3X_am*@!OQP@%|%=yY`phG8VliZQ+~Hph&YYqZyOFq{y#F4QDBOo7`+NY z7z3Q%Nl@B*wp-8(9Cs-7ael5^p;#ce^&*XJ5>xH!_O26A)KOTxl`HxtLtz`g>;Fxe z?O51^jYaX@o@K^{QhD0}=Xkc(YMFj)@>g0ZjFZCZEijVmS9?p}K7Lp>H#0N)^r?GY z1m>S?+!qQ8aPmQg&_un}Qr>FPk?JxI^`NIDDf_5^}^|wC%T-($Y z1o-~{smA>Oti0-#v9Ylc6#PQ+3@d5z?i)4PFLK-e$cdy*f4%|V#32sgLqDY9;*#d( z`ah79l*awe3uTUJ_xq2SsE-c$W%7s#zn1w`P*e~Y6a;q`^X30m`~$_zl8}^?7MDdt zh)zC80sa3dI@<6L{80SJ8pSXw4IhsUUW(P{@4of=KY}71(5>z^$&;7{h9K5HVeQLtGt{(1jram9sZ zK*v#v-F0YBXH^GV-c@W;b;woSoc?%nEvapoaOcd5XNKjojQQtkp7(vfj>Uh{sn_rk zSNI+__G36>yIP0}k=`$_Ck70UTBjxXSarN4;IQ^VSh(79C>!DGo++hL*u7mSU&sv#hVny&zX_-G-5TT=>oO2aWwwsmd>m8d&7U??;}8e$FZGGEZ%SUiL#Szxx7#~Qp+8^Qu&ac#$9)gKRAqB4xWGj43M!>vdw4BWWM zv9D`UxySZ0)~zVNqo@3M2+vu~9rwZJV;wcT_3=}cQ-DqPkvC>*`lruyX-x>0Hb-fC zgK}XVTW*J6?81|n^S zYaP}Y`mTih(D-~{_>UQbUl433K>Uy0zF4d?hqrDPrd7)>W<8>LTU`XVE%2!RO?g)W zy_=KK$|n&ZifFWhcApaeIOf`Fzq0UR9k>}hnD}`N-z3c~t`Zhw z*Pgc_q5Od3dPvDH%n7;mp1&I_yIb}??JI+*EL{IY48!N7P}l)8c+cH_v80>%lUZ2R z`V(9#v#ajh8&o-;$c)*;Tl%p>1)ANxE7WdFgL{URn)E+;ZnVvL7#}IPFqtfRk6dXZ z-(gW^|IF%#1M-&#O+ zExu}-c9NwY_CNb*v>g&N%T%kHXIIX91`wIEF;WUtNZebVji`SW+fwr#*6wTKBPBW z`_PYGaYzV6!w_t4cJrFL{tx!v`m2pE>KmoB6lIgut7A-}^lGZ@6pSAF^gvGMO{8&pvy9&Q|{~Nt;6(`f=C# zYK4pXF|h-s`;#1*H!iiCb`8siM;j~JeD*&c)a^bTf_L^xoSBdd#kl}0(|B~eTKAF$ z<~rAVCH;&pBbu6|ts-Zr$fNE3IjaY8BDu86LPbZqOvMTBO=(rT>m48i{6mClRdyOW zwt)2p%ON-7);18cyHCJ-4(>izIwL@IT*ryI40rGT>8~R8O0wChMJ81U^}zLinuh-^ z;cI@u3a|8H4&|ZxhZ3L${pFPLAN3o7HWy^IHQjOz3@#!hIceF$j?Ws+3Uf7-Tezs< z648Ifmd?lEf*LWgM&Ub>NgDy{bUar zsoYg&%&qC}RS2c6Nog5%R*ZSiNh>1CtR14bl<;{368~8-^k$xb+)}%8+}duL5v%m2 z*h+-m?lLiGb&eg-n~^Y;$t6OuY!VAiO$JFz(t5-4^Fi-!I=n^Cl=TD@#dc~ot$V`F z3Y(ykEvdi>1jvs*KzNXK{nC`?wBwzP!NVgnDjqa#^p3D(jubx+xc1DLzS8D}#W z%a1UExi3lMYNky7Q}*otAl;nprwC441e{cC-;gTKcbcXSwA5|*RBf8H_%DZBFDK9~ zVNcG*rrDWHFwxQSrIK;_-!Hv?*~bAIr^O5qi>adJFQl@O531put4VgD=U8K>Q~CxK z9KdrD`_PgD-J(jXQU5g6P&Xj$CUg8O*OMVeeBqF8_ER00@0TQuO1xDv@`+vXQ~ukl ze0VnXlOpcjtXs+?vxl0Gjn1(SWf0g(HDkS(znX4Sk7c@gPd6}|de9(Z_a)+JxEK_Y z(YW`u{2;I6Jw?mQCrC)D$7O1bKH-Kw5A8}u)h}(wbYEWXXgMlxb7L7u@0m+_&^}Y{ zYH0*_4HLPrp(;?kezq9W-LMQfy8 zD|cnl&&~Hd%d`sP15#!op(_(8zeyi*CBiIq*|g!gqa*t@0qn-Mf-ArcX`{Fd2&u2U z_ZQD+T0(sJ-mjU&m1yWlvv%~Oouy>?2S?E-(D*@|*I_5~oA!3T^{3Q9&R*gtMrGic zDp>8zP{$#sxO=M$|6Yon3D1*-!dNj`#y=V+qualagG+b35Wsv%St^n1N znj)5Q*C9PEzvAw@tmcux1omoJ?IFcKxcY?-(}U!A$`=~XPotd;tyoMX59YdEpGvZA zZa@i+qx_^?$GS*0kwC&QK=0Q)_T0_lMIM#GOmP{Q4=LUKLa{9pTsbvoKBOd^M}9Nq z#`;>%mdc-h%Z){OoI`goTXg|M9se45DdeXftcO`*q_QBildAJF*_=iWLo4P{{w}aZ zRh{aP$YnbLuC|_D;`eA?*QWFAjm;m`5>eGsolYGoOk@UXB0mZj2a`I+uCC^@5;^px zV}wQ#XoXI<&r)cd!$opNS!De9IC+@p8ZolUC>C-iZV##MU|tzR88i;D6SaJZmmV)H zad9o2jwcZgxuId>x@OAVLz93&UBYqpR2##es5xVI&95F4^i1ZV37Hz+^I)u5NA3s~edvy+q z$f-<%OVXycysJaXOr8+lzg<=nB0-pIcnl9dzr$bHk1{X)pb(R;u%J{gmZ$nJ1YkW1 z(^y=tp$zjMv)|UTu7ZkWH##L*i>~%r1`E|*FX^r<(CX4 zI_2M8bOpUXV|D|&DP=fI^eZiF*&(+`8TNvDivFlW-)gS=NS_$_JWB^{bz1GaVQMSk zv&eYtqFv=;lExgOgL*&~vf_Um29!vM8qFFkm$!k|_7h1G!XFNH4c%Ndv~k z#xY~-{PuWM$T0<3jVLp>AR)rvr^FptE##ye-u((|IWmZH3dy|!1hnF-Va9G@vi4G? zvRR2A$j8ceTp+-3IKp8;=I3{117c+D`0?d!W0tsZ3$GnZ#O}zdZs4c?ATd8CayFN} zIWg$#Rkd+Q#K36<`Vs017@28!8XHj&HTItmn1ofQR=67E$gJ+M`pF?)t6r%Otry5)e>?gCjnh5i< zn+FN(K-teat=8|>>T*I0PyefAz>6ttsYxmLK&|&6hrz%FF!0e#@131Dp{kO61P%`U zIQ}?sLgi$w&(2KsCz0?v$^-(x)V^;vxLnmMy0w1%hE_Qj>Mk%Ya?brB4;19u%l(3< z`7k&>NlgtTeL;X0`@$3w8X+x{m}%n|Tg=4AXMI7&yIEWkR2ZnU$&By@Nso6yZY0b{ z!E;8#u6%QK_1AOh8xP2%`LQFtL_WbInxNENO*{>z@=MYmlS*+7*RuHc%^g3ChuDy2 zGfH*Ie5IJZ1I^LO)1n8+UJ2#u=<$GLc>zJw0@&Afld?&*CdNCU3PS6#N zo|ntw#?Xs2C_JvxL`FnNY*K5++Z(A>UgDV~n@_sh=upnyaKv1=6#fKS!u0wHt{!RR zVNlP3)Ei;t90uUY)!E;*Iv?GozS9>WDO}elX#X#3$MepC_L{-Voy>6p00NdYZX%=B zHcioDdyZ0d4nK_0dn$XoR>vWrT0%QLVynZngISUsAk~ilbn1ID=S}SKJF5-qn%-{K z>UxuaBS1UBvgvn_O~Um374eF$%g?E%;m^zZo~B@RdF2lb)WtYBZy%V>E&Sp28(}Mu z>3Ca>FslCi-q;$038dNm#*GUH%O{hlDF!B;S3cC6w7c;)S&aoZb2P%o@<`$sZXaaG-|NUINqLP5d>3eL{UNhiM1or7>GMP*jGEr1j7Wii=5LHD z_Fu+}(U>sOGI?wR3XPhe^=zHwS>_v%9aa6k{?)XNX|$_p zB)ikSYY``h4VRT1M`%_19Qz#aN~Ed6(St>g!R}&?kVF>`v@owH`dWND{kNgn7|qs| zt!(jjMT-r~5|PQl*E^7O9vyO^?LPtaa2=fNU>ck^UuY z-6!ef^{x5A2|7yXfL_1U<#<>++l>4Ne}yBU=ZhFEA-kC%8!)RWXcX_~=FSyRxTkH( zGkVsw7$=fEcw|`pcJuD&Az&_EyT;NZLxG-T-ST_qjOg|Jzf1j8)+4bN%_p-2iPQ#kX(0ruvI^%zV0n0~PTR@MB zT}@Z%TX94K7sUJdC2!ncd#{tDPt9(JrydZr+(U(G&PWQH74!Z`mvB1vg|H=16dk~I z!pcD!ydI4sejT|V^SY;JvCg;6I6z-Yr031ovzdnNE0=FL*KyKL{kvP_82S4TVGs?h zYzFrG_N|ez{$~|?f&o(Tj>=NYQf6F{2#q=Y(!UMWiZ%943{Slju!(-?dqUa$;)jn-+v`R04sf-J?M*XLT1xe~#^%ziW@n zzT^W&BWV>J9P6eYi9WFr8T~eZS^Z~G!q2B%^sO^kgYcsTT5qv&Z=crAFQ#VahS;{2 zmE>cjcI0WpK4VB|e5t2e()H^$M?ka`0RfMoqw8iCu|48uSvgok#Trdtb459KvHK0v zJ8ptD728<2mKq7X2x%63^o`0#KsbBMUzP|7^pgc`Dsd2qAah?&uWw?yCA&818)02O z4$71c5JKHrGS3Ss1=vgg#yPCwKf?oC5*HB^dP+%Z&erM>{|3pc5`VqXcEnnSZ<8@d zf0F)Chf*K-@xvMSYE zO5TBW`vm`B5}po5VL8TU zH}!h!tfQp!D9jl}gkHZH{pkL4O{V3#p(US?Gd*^x0cGu<-S*O}bCO`UqI}LDTEF<+ zZQL$Qs-*P02jR4OnG_2x`jg0V_1Tau2BF4v?M5dffa!moBmEWT)u=}SY+?6p0~eiE z>Tdxw-tY}wRsc8;gGg!eEI>d_53H(N~cp1-q5XI#+*v1Xo zsOy+dpL!o@3k>4vdCmG>7+5EP>@=6$-9W-)aTpLTM!G=T+ zD*YsYjXuMMbiPTUqt$*!lfG7^3c^!rbTZkCk)J;_->D;M-_dRSEO3V1MDXd10K1V8 zThmEtcNbMTb*xwS+E(kky@W+)+&EOdcJCJcIFO;EUnXELor~bCk6v_kYjq$#k${{cBYIj;Y97qQD&hpqiL>&?uUnx!tWG& zDjM7Tk2exjtX3(EYZb@b3pSrgMOpH*m+U}Ptk!Zu+-Pp)iGeKNt7<=cB9(& z!-3ec8`6VOev~QUyfS>J@86?&i(cN3#aY~{8O2eF@V~5W{6k#7?YKj`lzcw^nY>U< zRrdaCVI7Wv=RO6s!|>>^WY_&I#TB`IS$ddTgH<4RB5r!E<=x`WK3C7^QcK$`T|wW@ ztlH|xs^~IbqyBD-Rob2u-<;g_P^L5EhRbM=CK}O~+<_=YY2dM2JUtF_N-Dj*_1Hb` z$=GgaK{GEsO_7W#SWoi_6ih4ue|`SULZk@!`@=#kaDyLbfB0i*btW}h^Gziuh#qnD z@bE?1dR@#X0xk3|g9lwaopD+4PL@$Q?>*~v)^MBoe$JY(loKL)4`p5NTaR@ZV-ehc z$mq9x9XVi}w%MN_QAaLd<8}NXvm#uSL0^g!y?m~3bR4# z)|tx92AoX2N;wsuYH0kQDdgt9LZsaDwQeeK1`&nyVQeK={~tqJV>Q{_yR#RgA)ke) z?g2CaG<9tB9w9EY?)Bm}q1qIBwR!CV5{7Skdn8wW-3s`4w=^Sb0;l6SC#S2@psAf> zXE`J2ubtT(#>t#@i`IP=VHc2!W@L1zd$M3F%DAk=6=(+fw-Pp)|Cv3=q3DvypGEj5 z-B$pwuY}aa4!NUv9Q9(0K=%;f-2*4xrWX^r__Zt%(U?_aXc747E~0_EXd>{P91h%> zZl%U*Z9mNPjdr1yaaIF@;u}t6e*$WWk9#TDA6RQ;zyHU>&iYc_N@AR^@yV?Bn?=xr z9oJzW?z;!2$N3NP+*U^?`lfz$(Ajx{LNAff_ocdaWjKZ&B}&CuA8`#`jdy=sms=i1 zsV$URzDXvh<-~qhN)+u3QH)xlsKvaz@9n^n3`=IkzLo#%oyk_0PWK3o)%&^sPbL$O z+1R>G3}I8%xT}{o?DB8bIw}l74R#smwf`{jc&C4srhibosTKPdX z`n(7mMTf6;vCx|`dU9v*4M=nZ%P!K5XJMNNhP1tweMCmVdW+`X{R{9XtRL8P{tkKC zqbRRt#$231&`R^6?=v_n0Z$|t*8hPuGR_m{hHzyzopGt6jz9{A63%*P76npYa+>JMHKOMM;CiIPxC(Flb{|Oxf>??^PU_{MZ};r zrPmn&dkkXq#t@8&#k?*}zsE(bJTJHmu{d$uY)bsO@15DX-^S!{`Em(& zZpBiK?)}Y5$8*|a1~dE=VyNvie;ZMlWILP30zkXxT7_+^z=P&~n=HtQ$VSxv4^qy= ziJ}K9mtDU0!r6w^*i)8I)a6Hi36AG>K@xH88g`~T`!)&Rg`8OSGB7&&*TJ~gmh7eTXvuXQu92rna*q=w5aM9kfN2*w`8gqoK zeIr+F@-y-hC8viAA*p-e>scIj3(+`}@bdq_whvmMTf@<#t&U-lIUSZNH zn53m`!g|a2_E?GjvbxultvtKSX)o0-dFsyVm^j@{RVY1eQEf}d@C_l$^*EHb@vMf8 zdxfzUt?nkvVyZv=VrBrIR&*~k`h9|XPoPE#I`3a!?IFhNfrbfe0#JYO+?p%6OK#9M zacwswqjrQ~vL~tD=1MFLJcd5U$$Z?=6DC(6!xy0U#<P7RK9ypbLY?U)u1L7ualeBlcj3!S)nuR)T5PKU9t7#1L4E0(Q(%*MC zFHCz@IV?V?wp57qt17Jt0K3F`YP5N`)?xGbEC7Q%2}%B8iqJ=xismF$ALpSMENEnv znjqQAEZemf^{~7{v`RlY-_-)Q`u&$zM^dKR?-vRrlibarmz8UCJeB*HU33DT*x*=8 zPCHQZ?aktCbCWpk{Nwo(YVE0N7z(m_bKMb1k3_=4 zg9@bh?kA%nT5tNwsLi1o@8vnFE2)6!hhC&*)`$DzRFyqpUz~So&BKpJQMYzn+xZKK zWXnV(upz}KkQw%AU$}EMIlW*VzjB8D*hKQ~14XV!m$Ryzeo_NGVYdH6qwuXAC*IIN zSg8s0CG+r9rbEG1j-T~vjM2?ar2M0Kk3$BNB|ddUNo73TwDR1inBUxQxx#Mcdf&OW zB9pk8-M9X@d8uDcMp>Mr#-)1xxC0`t>c(Xnaf&3`O!dC%I*jDr?|fc~U|tXt-ho%~ z)p@HwFgtQ99n9y!n3Kt;44pU%PtcBgs0q!^U3JYdKPsz?+Bv39^>^d4Z?ZJ+{CPa+ zLVh_7PiTs_NC2+}7xzvUlEG{#_2?;2=NYe2WZb~l7SmG7!|aqFMS4%C8lwC-Dy=R_ z6Q>}1-xNZdrp_)NKPdK^H)UegkJ$fAY}!ovJO;RsXN8CxmOhUY=m4(+XXoPZG;eT*+NJu-IobUm@)`%$Kq;Z>0mJ zeG+wj9rR~wk3##|v$H+;yMj>ZCWW|NXcO9>B^_bL&-O}m7^8!{d6*qr+I9-!=ATMu zekw`p^>-hlih&p3P3_c3zAfCa@YM6XK@9IZ1wOF;lZu3*IWw%L48)(A&}7#(&j!k1ddX zbdU8l4#(C3Q}uowW1q2ZU+Qi71C}71bv-2@;POE!y)~dF+{>M%ogspL+}Jt(yRg_a zj7?j@CI<}G-QbV=sUh~{0qrvj-j`MJPel1PzCD3ptrBEn2T3aAT>*0PsskFG6=pbH zhtFLxZYh})qlH0kJhRZOQeT-gBNt~F^r_D!PKGDPZE+;Iv6$lH0HUc#^>FkhV@Vhx#e_5c{RT2~TckG0vx{noR@ZjH@qBx-@1@K~Yz3p-QvO1flQ}(QzT$WdvQyMGJq3Ql zT+S{fV=W|bs)2?=IsOfal9ooN=OYJ;tw6x=7zAP5Kg(#kq2ZE9Zk)Rw4X1V)_o)`u3lKDqElNZ}zF8XIR5W^&nl(B0YPGc`2A=jB>esqAaDHBvJk8?%xO?|mS2;OI(+=d;nmOf z^jPVk4Z(6E#kF1zVAOM%jbf3Mb~12HmR;y%>|?XlI-eUY3fK4yd6szc<#$*THPM%& zup|StDS<#M@345YOugF+)WP*k0P1E;p zTS$0fr9+|aPmnzB1L-^PQK^nM_scUUBf)YE2saAE2=`-~rYx%@+5fQx>@HbovY0g@ zk8s#~-ZbUuJ3ojT%=E(7W)-2dF^p|l&{dFsWM-lHih}iX*Tm2U$ClK|ye-$$mLq-l zf&T=L2qqd^tdQ!SLQ~uiv?Ga*S1qf9S-^5>q5+C~5gGg%l7>85EhY^u7D_(V!R+7E{iq5% z)1Z-l&)3hyqeUD~EK+EC1Z!3ko?CdpKlQ1Xn$w*j%6l#@Z|8VzldZlAY(nYAW$KCT%0`R~bXi>TjJ@tp1*6F|zVN$v7 zDKC%ghM&iCl9)G`LLD^|&n zY>X$8Yc{Ay@l*&1Lw>v;S_}I=q>~@R5WenyFqaxK(Dsp#^0oDA;q%1(XTjJXS4vZ8 z@o!Z=M7(m@nVng>vh=k zvm>ZpevlKqAnbNb(k@D$qvbI(kK!R~v0sccTDuuW=I%2WG1W2Vyq?OiszX(g;eX<) z@Ow74R!Sats2y}DS~8L4!I0SzY4EDH(aL6y#~>3BHP4p;>29^Wy=LC(dTPpHEOD2f zjWAw?yuO&N@6|qAa=B=1RF2oWdl<=iAR)^L|8#R>A500ZdUE2dDUvH}skzLysucYs zW2k%(w?vUht*dE9`1dxY#U9&?c~d$(WsZ4aE;Y2$AIupv0O@p#)v_AQ6?2O?vI2!( z6@5;fpfV))gus*Y8=SU)5dReMdMqp~YR>OJZNKElqaeI~Efh0^fTi~fx9)NJitksF zfpmAh&T`wsx1TO0raIbP{wxHxIgCa@CS=i9KD=$Svi6(OqZDXoS>~QDUOh6>V}g3i z7?B9OxXRWdv#e-ET3U?1`9jM{RLa}#@F@Chkd_4*TjdE9oNDf~l6h_N;MmQ`q}Hxs z_T=|2%Z~?kw`31dcY$Spu0s_v;8DF?d#_9nLRk#RD9{p9H(T42CIwdgc8G#8By{7$ zPJAkh?OqMlgR9+t=5E7?`Eo-|ohRGg281T#wN`b`-kPpU?RmPgLX_YCT6nYb?PDuU zH^VsGs}wG)h?*8fMln=i(|gYkX*AtWw|nV0YkTj$ZG{=4oX5Fo$*iZk-3&WjvU+Mf z=Ntlza>JT}eza8OijLiOfxL#^tyB9VxuzMor4P=wQ%@^`@S8U(fB@-9blbZ6XTPEg zA>88~BoHiU=$c`J;)M;1?uf193XqGK9)>nJLe%n7B9`RXwa{vXHu84e&i3YOY=lch zGjUbHIJz03T6Aoe`^@=Waw5%fskhrd%~Gg~0kEbrk)Xrfhqy82d=#ynpiqgaUuWA7 zqgkIQTMFh;|3GVr&s%;wA=o;chA_o4!ADN zb^5(s26A*SsZy0%2YPL%ja<0{70w6hMH|*GxzLIk(@VX&W-rdMY%U;nyX*<8hg=ei zd}(5RWd&5c-?6a1 zed!QZv0zibH8Bj&)@_7x!O|jf=|+iVnHh3HicD*D{PtF0UvC z4@B5JqiRwW3#>|O{D>_a`il-$TeR{M<~*NQN`S3)8g3j~E2;GlbEC)StNdM?g+U)% z`2s7iFV<;N{4??%-#k%q>^a<R<`RT0>0ub?)jlHA~d$3pkTe7SJf#XM|WZJE5s5Lulzb8XLL zgK?ZReQ_b%-KMX>QUmAjw>|p7OD3T;=37os?Rg${am+skcgco-+_$SlCuiD?u z!qTY8I{@Kyu|aoHglR1TIiWG8O!nddqLk40lhK=z4#Qh@0v!jog^gBod?&RO0o^Hy ze?5^2@L3WM!q3AwBI|dG0*iWw2S!XUqKqM>$>8{^hx4LBCnt67&7pb4Zx$@1f*Zv97_=*ko*y>HC4IcL?nxPr z##dNMsNGTlnh@>caz2(1a93qrQ-rsa@CjL~>pS$Oh|Iq?Vnho(6N=t<6`sgN7tcwJ zheF%5eD{{@u@@D3vTYu7MsGjOp|lj`TWg*X{>;f&jY(0SS439!Ci2h>v$e8%UeG?J z+0Ql*kYZn?QDj0Hn+fX)ni>Qp9-9{&j5QyO9ur-zi*|Yd(a0XvJ38-$^D07M;SvBC z7ToM#Oz%^62@=jMEm6-!zz!=L=3T8R^lF(S*6{4*l*Y8{v8$ZmwfHx;VpdA zE96-`G2aKjzpc9l(w+t6Kf6u(G37wk{odwHbHwBIrSp}`e!N05+NLXjaVez%i^!3t zW!lkE!mAAWYC#cdt=25ES9&{nJSI90bQJT3>L~pfo~OgrHG-1+p(2e2SI0%Gxx56! z!|zhLj63F>2@VUq&@(9>#L7HCBjPVB;X#jXZPz&q&yX!&phGoxM)w5YRQox9_|s+4 zO22G`N*P&|a-(+MJix;0hQn*kbxlkb)5Oh8P1_z(dsN=29E>&yHS=vQ3JV>^K_=rc z)p|y^o5MaAjY$R5Ek3(W87xreytvjxqtohY&7iv(6ps4(23J(3uQ+Q!x}4=+gUqW} zXLGy!tz=2&(QaI0n5Z!dHE=WL4g`}-B`Msy05hz>*B+S++3SHxfa+4pZq3VS3N zKqDd`Sq${V_`{YE%3=;>(c^%O3+C$BRt*!K(qag#K^^KZP$uacmiHcEe;HOU==Z+K zsZnNBu0?5J#jkhH!#JlI#px#T zj@IJUWYfAOKYKLqX5xex7S@NvN0c7f%g|lQE5LT}THWOcOq@1>DS&@IxBTUj?@UDF zhm&&@L%73r;{eG}{9l+!bc|vcCcNrfw<;LB#^Sp`|VG zhQi{;1L%e`e3As0{g0hkwTnQA7grjT>MO_ZPa9e8hg+D}k@1l)_=Hb8l3lZd(E+_;~ zdrZ*M_1)1@Uh4Sd5=SAm_V7!uzS#t1_;_KwjeQM7SJJtHGp5gUQSqa>2za~a#C{-= zPnJ&m=c!NL=NUH6Z`+43>bCUAtH-csVa8+`8w*RB`9V~U0%;QNsLYJP=&`vT4KcF6$CjBWGrt6up75(R)pGL>8ypv%?mDQL0Ca&? zk5@V_D_c9*lr${c?yeO>f9#=m-6*6bvoM?&kc`zgZh;i6_Eij+K#hA7wk+Do67D9$GsE4JQswuFV~U?MpYGi7SEf6XBXN zk(LG!D}c1!oE`XNjS0HKq5!L%+VUc{C1+BSy5aZM!V4=_l$a+C<}8{>bMcn1l$C{d zePDj(LJ>u0v!GNlG=UQKj{Yq(;jL=)A>_Vuz2;bGo>e4HCC4sJpzd6d4Q;FE+sn%T zqqdHwR!_-!uh4Ierq-YFeD8+Z@ayzc{@UC5{W1nVL?q)80D%x(G9ac%oX|8(n165X zR&$Zu9Wo7ZU~HTSbLMzSG zN%dtN;cIo$$D+vh1^}G>_R^m#&2zag;Dk1-2=S)m5PEmQU@4fDKYMojMUGw$U^G#Q zFRyGbCHsXjD?P_~P_yK3*XneWwl}(2hr_O{fNM+Q}+*CHtL0WbY zZ2)|HZ0ie~u$G}zu#)q{lwW@_c#SOS{h%FPs5E_;LxGy4;EE{ogl|5xCgP`Yz(@EJ zLdGkXruHtq-Cs`jZ|^auOEjCglE#OhPpALK79c5k-tCv~3*#N0J2uWE?LF0*Bra~D zh7PSDZWu22lTA`CMuO~1fx0Syw=P#bka~afTm~70{FujlV%+4qf4~i(6rp>)EI3^u<=_^QmC-Y3C}12 z+=|)H)%r#5Bq|*o9u`&CU_dop=unhw5@gs?_GJDTw$eaQSYX_1@v#|pT(jlj%r%D* z1etze<|`IQYduG4{4*giqW838W7CR;Rwi#3f^34rbfHhodMx`6Yz`k*CZWdVOV?k& zG4o1f&R%^OMV^ikV|DpgfZwZSVMmyE^?mjIecGi#@^Mv4o~f~wJH$Idh&H>r<F>U}gZ5P%ZhK$%51;9)-ApO-vtf@x(Q{_5@2^_E;;s(}S-B z!aT&FBf_8bX=qmBY7`{lOsfwKp|ic?M-il#I+?vwDZK~EM$oNdYrmYP0`Qowsfrzj zbO`^-_4Nh*V^JRhN;&(4lJQOf+Av;u3T8KjW3Qe|aW^Y_>%0c>!7ETf){(yrWO*<5 zJ?$Z_j!F20*m4twkcn}5u~*J3b$3?}F~T7Vqm~R~>Z7-StDc~6%EXRzX@@W+2s~}c zx?_>Q+K2ufm%O-$BOaLEVVfR9K?4Ixx`rk{K;&^?e@simk4J=UJKM3aG zPxQf`o&tf@PA)w5xh&dNOgDAR-Y{MKF+d_%w)&uY$c~}Vdh7y0msD7zI;!+XrDmheBO z6fXrh>xw-#JUAg3S{SVfI(qvFuHT`a#ml&u@rt8)vq(5_c1g}L4vwd=LEq^-R2*&V zF||g;L%*?N8*o}e0l9n|r{tWI0w>Qsb^chsp5ZT*&0uqtOYXRfhc^VYA1Y4hXckS2 z8CrWdwbN#WdJ{YLT|3MSo?|GbD@##BmjN$q5;;ELME^QG#WRvBa%$cJ2Y|y=W7R{- z^v?ak%%x#sv`VmJt!w}rEduVLjQEZ>uj3_x9h z*o)q(uPRDm>TQH(iMF}EaL9Q_+Or~DGVAPWjQ&;x{Sn*Br4etY*n?HhSd!f8{D4qWY5Y8&&xI$r1P$uOB=S`KtmaqA(2UZr0~v4teK|F>wCIe%EY|m71wf+htDyz_rhBqTOno|##_r@uu zE^>mcgLB(E1enQY zxkX`ypVWR3wfxC_gDq3UY^U=HxHg@mal4{lq6$-Fa8gI^MIv>&ND|aLV=;3pi%!+i zql^8B{nq8w`4Aqfv42t0`n6IOA3y=tO@pXrlchsST08_n|Ej)QF;$gBi zk@mG`=8kbU_Kx&QkvxJ)%=9i4Pa%QliS=A3u)yXy=;m`+V9jUDq}7cb8w#JUEWubH z(p6ZZ9$$@+KkW8lKK+$+R7wtJrgW7WrcRO7vJSO*lf>Qi&=Cop#$Xj!r|7j-SQ>1R zrG1I6TCLz=uU~~vn7?=HVfaG%J4XdL*Ir7-jS34Gr*!om-g~KnxcTG4X6q02H;=e$ zL3XAl9((0`np&}h3v&n8bUAkaU%Ds~wT@NM>5_jHshPwm#ne6Y!J*$(Rk=x!~jZib!)~rtnQ3pIHv#^D`V~)C$x%9AL_H3h->CN^3<0YA&H6abH$GRQCJc%@xqIYSuKAsX~7&XVm0~3u(UfQ zvB$={6)=P;?d^S4G$bMu)PSM*Po{t7jBkXvPse1P%uk>95kzPBc{RK_uJX*{t656B z^ngpXM~y5Y?H`+utYj3>p4z+fD@p0@)|61e#ydDIjzv4Zl%qUI#1EXr$%Q)zc$}Ks z9p^dS+gn(mfopYk^o%Q8Kp)S}QLmokpm>bkMt1Rem_t}ew2{o7zQ}FzgXml<@nF-p zbAM=el!q8h-tPRK@K_!fO<0k3@P+iWFEz;xC%&zobnWrMVEt|2THUi_42BUw4PdYO zEUZhzF4Yh1kEs8h|H6Ou{ofHw>*a&T|Biv!_^(8SX z*8h$l*#4i}NPT=n`R|MW{}xzS|DU79vu7Wkm1}TH+V*~XBejbhZcBNlXihoJfX8D0 zl05Fw!#jCy&duiqvdPbY`$l|rz~?00-;`fHvFVTRzZHh&Tg|fmH}%!Ak0@2EP2toOBlIp#7_6GGf{88y?l8?fzt|+f0Cl^i%W_FlCaD3N_kc! z_VSZVxCWr8bqx{KG#EySu@aN(>EDG<*;W@+dbAipNOV<0LNEW7T zKAvH-_2(i^ggdw*%>zs^+rA3Q`rD|0!OiOikqlZ?$rPHEpgJmxh}vk<$@vflfzF3M zONtsW8l#WVzSyeJ13GJZTAizowq)2QQ3v-sT<<)bm7mEhmCWkq{YQv%O8nQUPx|nF z%41v~n%#WxMOuM)h3iP20?*_WO2=1lmEz&#NG5*Hms=0Env@%y+R9%st;D1Bc#GU>?D5 zN#oi9JtseUIrrcCeN&f=#XPJ`Ij#J+)cFstVLF+-t)hmRTcFdPdt}hN;vhR)cddTg z`A7Z3KSpk_gOq&qCfUNZO-x=8pcxIJ+_Y-S887x~v9vmm?ii!3a#!g&j|F$*b4H^y z&Hq}h3%u}ZD5ne9yZtos!?g(P-zV84)|n&TQ-gOu9K6PeA;9-}vMrGmQmy{Y)3)sX zs1o)kRClu{8FcOp@VR;Flp^eg@h`B|PIoV~he`6EJ(f)P9I)A43HtjSH?-;l_Y3hC zzn%|S?2YggBQ%C-RGrQXd=JuT3P|s9OC}2PlUMV$t!yLi?WK>NVFvpJh$ncR`U3yC z24(b9VtR2AeO>Kujj!It!rx0xzW^V`>ysMfrtT5TWBL%%b$9;SP+-hBYiagt4)9H6 z|4Ebi1=qp+X2r$e+c)2pcG*@-I5;?9e@pgg!J@bf#qfS<8$h4hy}fTFd|6?w6VwdW zu@b2(=?K2xvZWh`aIMEENq7&^{^YBq^wQSp2{UV{b6u(b4jO0R{K?1gi!fp=%eAQD zM8>tqef~z8tdQ@kk%+zS;v~N7Zj@hv{1UbN#G|vg>Op!EZuVL;bbfoGdh(A?BaFMg zZ8kZ=w+jqo6BFiN^XPiwIR8`;;gcH*xqV%4bT2i49NR?u&g{8{6LLPzoFUFQ5+Ol61LgVZ75`|o5u zqE=alqJwtTu_@y+b{h84COPW5N-l_Kk+vu(IEWIyC6t0edZO zRy-Xd&Lh{TJQPw}C1bAWV2sPEEo@8jzO-yKT0`bG`u_DYg>>j@3(o$2!`_O>?eFUG zeIw7K^_>yoY?;$(ls>3R)FT)9=gbLM#wyeLPNsDtnjvhw&BsHqe6HuYdfQ(QWWi&9 zo&q|TN|Jy4YxxNy$sY7rk%h^cQ}C|3-U3|M$q#SU78aOpMah4L5u7;*x5|=V$-t!EJUWV_oSgU+@96CBlY7}BGl)uCTGZO z{svgPrdV;%JhTomCeyb##3#GT_8-N>{$7^s@{D?WRNOx(H*0TJ0K4BK!%g(1LoxJM zplCciSiF?JM%lsI#Gw`rFa?7?%ygz6OO~*O)zXA-cs5Bg{Nog^KMLe34aVGzpB@zS zaHrr!IlN@8mQ1+$@T8dvG39Jip2Fu4r2j}VUM8>m(^vMS(bP-%Nw%B*w5^}$kKfMQ zxxVkt1gk%#p=c9zp**=eFz&MXxj|u){k2LRUYkq8&0S6DRln#x*?J!Wqnlas2kV(s zXMAwS$h8Z=8iUyN*vIHxF5USXBAMIGrHqqDahx|AQj0amtVMZH!!eQho1S>>IUVp( zs|zx9^-!jgKE(I*!xWQny&;)SmdqVVegamS~pI&6i4ZXCZ9niJoihanZfi$JO2}jb8Ns(A8NwISXRq?V%aDF2DnntVRs_6}|0aQU9;07zbVR#SnD})djBR+xH>7;;5T29#s zoGzX#^>0EMe|&0$(*efiPRq5{SK7Q%K$Kobt(TG~nm$q8D7tx#==1uC3m!$&gSB>~ z^8DBTQ{8ueHMK?SVnIbjM2UbR97RNFQIXy}A{_*Q5UD{CkP-sYTY^UvQHr2|bZJ5o zLg*cWdH^XQ3WVOIghL>N7D^J}?wot?AMnN-_m1~|${u^o?6vlqbFTTVZ_cqc9Iy^2 z{N2N`1#>#@XVyMr5{*#KhDeo^H}@sJ8uSRA(k%pVf#lk-kHqBE35al>w7|RL>(rAv zF?o0MBuXr%f;H0rq5S!0XY|#GmWr>l#2CO@drVArm#e4Z)aLbFH7iN~NI-56*7-gr z#7IdFQF$|iw=oXgZe(0~BXUbzom%?_T;e8B8aB&M)i)K~6zN0t-N>EM%AGM%V1dw^ zDhn17dS;4JXUmI|O_YYh8gkh&1;Iq=9!wYQv(6Xnducr} z58`JJ!i4PPcda2In>9B02ND1vWvN#V=~-OKe$$rsFwwj}Vz$%Hk#^>J&>x#Nb#emm zu}`aT9?(AdggkEy{P98G-SsMTY=?ASCkAmPz`dNb(0du;?MjJy(R4-e7nzu8`r5L!K#6wUK=Q4Njud(-W9opVHC1={(I?ULC0+`eiD3$$zfnfhw!M+`Tfdx} z-~KF6(xeWEqjz-^l3kN}dMR0o@Q%1*WQlaRDh*Hl-+Au&z zTwzjYovxxc6{YWAy`ZfR>Ni($hUwTOW{>I6TJ>`M3}zVcYHGJ1)vC$5F6dyRR9$#CiG4mx2V~@TT6rtwH zFb*2|y7RJkx^LU(Ho^krj?3}uNMvQ9+zt$j(jZ13MhBzb(Rt=lOL7>W|4DPytqX6L zI|pt?{1`AU+2+%s`P(M|2Lw_kZ8cleQwQ$tWa?U_hc1J)2IxjESD$D$8zx&<(OXCX zu`F(m557Ap?0!3KRJi)}&ULEtl|(rnRq&bcx@6BsoXHaei*WK3-uU4D}x~X=Q{P} ziLq5HH5B~t!Ze)M@lFvh_H~T6WX{}2!yqVuJ!a}BX6*t`?y=*CQ1$RkQ9z+P7k1du zC9S+x+|aC(zBiP24lmuyWxSHic)9yBZdmulP8xTMrB3_j;7xzR8u;ta?RECcxZr&4 zpMAeZe~wrSooch%9X7YgQa>`u{H0)BebZ`up^UESHd$nJRiSg^n&XzbXK$R}zTMU6 z+%M7JFe6M;m0y{Y2Td?r9>+auJ#`zvTq-N^Ayl_@GQT3+re-BhyZA4ku!(Da*-n&n zM~M$^lvMs@k2eVX71@#t-k`7``LKx|b;lzvHo=AurbYfnz3r&omNKK@z96sDr98kF1yl0N2~a zFM(rtJn0|T2jCc7Ni2LXAFe*MdyktNbcA?58SM?*d~0ytIiu${n!5d164I!Z1sO*a zKyKh1)>2e|ja2p9EA6gM!sM-Xso$dUeq?kQD3@7YDlZl~H~%VGXSTe|=mJh%BgmyA z3EGhrX2;Cc*N06YW;)k=u)C{Q6OHU^ks3O>XA}HC9ZAxWgh-7yTY%CSMRx*Rw%x?h z_O=Bp!{Gp2DuHceW8Cc+J`TO77a!+vw(&52tH{KsSMknDPx=lX+BcQMmsob5=XK9@ zub-rMg7=L>eu?je7{A;-cm_wgZnWzMtuR|x3AFC8%c7{3(*5L5&ls$T*NK(w4ZF?y z%1wK2&26dSPs+0FIH0eCKxj>OTw8_@+Uf4;ji!A{+*#fCQwtm_UR1=7!ySrCo)dfl z;PsSRxHxjQJOIJgs#H?~y1unvG#k)Ph6FndT#|jQtKqjz1=S!>{A-z~xEI zU(iOU7AtiZ7eZIp0s=aMH=rCO{9~Lpg5R}P_Ncm2W4ke}12!K)^uB){2*l}s9V7IQ zLwZR@uza5Ue9h;hgSY4DKnDDIP$JabV&8$MP& zX(R;EtO@C^NDx;cg&LUI(O!?#KClMkLhei3d5$*uZ{dyuj^SEU61JhO+#6CSr5v~t z67^23rIeCigxkEm`^QWws=vhlX;d# zN-E}GO4Sg*sD1x4_U`u;r9`NOgNBa@awIaIN#whD-XErttcB`q_0`kB^woR2$}B5X z0b3Y&^+kx-Fj>RnSdMnF5jw}NXE`Lwqp=UD;0cZoLq%jK9dZu;%BJ9CDiCE}#532g z!%KIoq(&Os#HFuO)hn94EoI01_A(s;{#%bmHjU2>U0htuC|>UFmk;UnfsS_+og#sM z@V19nNJt+!1bh4gHf?+~Tq|eO_{L1%(9i5}E{p4evRxZeGckj`!B_ii-b=m2oE{Od zDc59Avoynd_<`x+S{C7uT0G~Ite0=9%6eaR16B8XSki{W7@Eb)bH=^~axuEy)o~K|#U5 z>ogkC)Rc-B<8VZ5KlE*ro5H8&=lP}a<}j0>p`j1QW>GJJTwr>a1 zXOaBsVj%teUApf%*N=Yy2-fTqWnXe8GURR2eLDMmlm*;*g&0O+n1c7`{CB z#cvK{ztd#_vnE5W8Eue=2+&OtT%zqN$fkcFTO2865mKEO~h=x*HH zuW)j5oXz@9mnF=eJOq8)`i~r7dLSX8AbmFae;~h#1gag!%sf zwVQ3FVF%=q?ik+x)9U{prZ~*;x?uXv#mEm2G-+jMeJd;LK8%76H{EbgYm?7lSl|p zxZHJFjrC$@N95Wr8EYM=h;qr%YLk@AOMRfS__2x2hVHNZh24v-B+1Fjjyu3={W?x# zeGk^oJy4skVk~Ket)vx&R;(u_C1EgBZ&zWJXYU1pj<~jNTSb;fU-DlSOy1jPkWlPd zak1d($wt6lwf6u1@OQQ~!~ut_7_&>y7@)iBU&1OsP*>Z%-8XWB^Y36`kkj&GOyGq$ zJ;)^X%k0`trJ42!9);S#J^v@I+y%$sTQ(C+FDl#nFDetUUr?*F89Bkr0?or%;n%mqb0Y5L znqNKp5G*4t19>%KpX@$*t3%9Z`cL5VKt!S`f=%j)yEur|1I?u^n1I3f6Vcy+DrIM) zy|rU$o`rhhV9m4N?0R~W9B1&$S;bt={B|||)2Ah zV=qaX8v{)V+E%;|)O;#9FH{zE>k30a^1s$+F)P;`8Zue!!TJxKHxt+=9MT@edigvYcHXUV;c~FJqrrQju3Mh1b4a&P3_<@5#+A< z@2?X+`A) z#62PL96kImT2M^i#XKnNxJ08d{bKBSF@3&aVU1zvyu4)y+QaA01rW^O@#Dw6xZ-w` z3a{>NlJeOYQu8J!U3y@06?4FUnzI)ul2U1ove^$-7NVRHj zO$K)*VmZ@ZT80H_@!7&Ybt5>ybrA%v8P`LH*i+Z#<$J-`PrSOv z2}zUZWA1;ubP0e{hj%OzxtvM7M-R_~K#9G*qYmoiDgS?EzEH2YxXeSy*L3{WEG$3Rv^qtYB=R2N)(x-^$}0862!HU-;Q!jw>4&g z%NdbpG1ZY1>d{wVXko-y9~|=@YYN%m?k`eSQkqv$`Mm76b;Ny zeK;8d!a@nFea0C35?d>8Hhp`d)~oF;#B-?Eu5PtygQUE$jM=jra;Wv+eNHZ(wD450 zT>AFyTSs>AOn`(;yGA|P7$88h#HCW5wsZ;g3rKOe0MqV$@+Zg6_h zwTlbLMz)LO^yJ`2$wo-uj!S^)OdAJ~$xtW+x+F)P2SFVqBk{)o(@~69SvO#?l2aY~ zE{aFYC3==NJ@1IWMCo=;P+uWK9c>8S6G#>{o>9iv-6OK+`pwMJZo9O-1y#BcsoGAFT@Q)N zqtAcq3S|LWRIfoCbnEb=?}W_#NE}%#HuC1zntNZdz7B;UZGiK(nM6dd#!^YGE6F;; z@*UO`8{(?5ooN&b#XD*)WCby|=?6WPr1L^TSbL@!YTqv)fYAem#kD^HocuOa#ka2@ zj84!NKB=khtY>W8-Ii*;YHC&`>v#H$5&I*g>|iRIYT|p(v^Z6}aXy7tzR8>m zcIayBuxGL0ubZ*(v6=t{(gDztGw$D?R+N)%aJ8LAAb6BcNv%G4>E&`^~RGaxr9YK&QWRFQCu$O`e zVY5Lk`IBm_Ouu%sPoIs(#ZjAyw%-|3`&-T1alm)=twKm_YHU}&X9;B&mj3Oes!&N;8OOCG)rqcboQ{?hH zV$m2328&(tfO`lq>|6+Ze0+t4g)fisTdU5^=)UU-3R)Q-86PN+#yV63c<8Q6n_E`I z)}mWx`IFt}$)7Pvbk&7yEJZ{!bhD98z+VeTvk-A4`%Etj-xlSd70vBQ^}yvvkX*ik zr_n;EqajWnPbL}yrc{1MEx7iW_m9&T;^I89H;pL&mfN>lE*|&p(F4V&wK>8SG|8#^a_Bs;((>3ScZN8!IgZriToc{U`);tpx literal 0 HcmV?d00001 From 5f839e45de319d50b758a0d750a9fec13bf3250d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 01:53:09 +0800 Subject: [PATCH 270/311] Update OG and Twitter image to catalog screenshot PNG Replace social-preview.svg with catalog-screenshot.png for OG and Twitter card images. PNG is better supported by social platforms than SVG for preview cards. --- site/index.html | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/site/index.html b/site/index.html index 1c227f6e..56a58b97 100644 --- a/site/index.html +++ b/site/index.html @@ -18,7 +18,11 @@ + @@ -28,10 +32,6 @@ name="twitter:description" content="Explore 4,587+ AI models across 95 providers. Search, filter, and compare." /> - + + + +

    +
    +

    🎨 Multimodal AI Models

    +

    + 1,548 models that see, hear, speak, and create — compared with pricing, context windows, + and capabilities +

    +
    +
    +
    1,548
    +
    Multimodal Models
    +
    +
    +
    1,487
    +
    Vision
    +
    +
    +
    118
    +
    Audio Input
    +
    +
    +
    34
    +
    Audio Output
    +
    +
    +
    28
    +
    Image Output
    +
    +
    +
    167
    +
    Video Input
    +
    +
    + 🔍 Interactive Catalog + ⭐ Star on GitHub +
    +
    +
    +
    + Contents +
      +
    1. Modality Breakdown
      2. +
    2. Flagship Multimodal Models
      3. +
    3. Vision Models (Image Input)
      4. +
    4. Audio Input Models
      5. +
    5. Image Generation Models
      6. +
    6. Audio Output Models
      7. +
    7. Video Understanding Models
      8. +
    8. Choosing the Right Multimodal Model
      9. +
    +
    + +

    📊 Modality Breakdown

    +
    +
    +
    👁️
    +
    1,487
    +
    Vision (Image Input)
    +
    +
    +
    🎬
    +
    167
    +
    Video Input
    +
    +
    +
    🎤
    +
    118
    +
    Audio Input
    +
    +
    +
    📄
    +
    141
    +
    PDF Input
    +
    +
    +
    🖼️
    +
    28
    +
    Image Output
    +
    +
    +
    🔊
    +
    34
    +
    Audio Output
    +
    +
    +
    🎥
    +
    4
    +
    Video Output
    +
    +
    + +

    🏆 Flagship Multimodal Models

    +

    The most capable multimodal models across all providers:

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextInputOutputTool CallPrice (in/out per 1M)
    gpt-4oOpenAI128Ktext, imagetext✓$2.50/$10
    gpt-4.1OpenAI1Mtext, imagetext✓$2/$8
    claude-sonnet-4Anthropic200Ktext, imagetext✓$3/$15
    gemini-2.5-proGoogle1Mtext, image, audio, videotext✓$1.25/$10
    gemini-2.5-flashGoogle1Mtext, image, audio, videotext✓$0.15/$0.60
    llama-4-maverickMeta1Mtext, imagetext✓Varies
    qwen3-235b-a22bAlibaba128Ktext, imagetext✓Varies
    + +

    👁️ Vision Models (Image Input)

    +

    + 1,487 models can accept images as input alongside text. These are the most common type of + multimodal model: +

    +
    +

    Best Vision Models by Use Case

    +
      +
    • + Document analysis: Gemini 2.5 Pro (1M context, PDF + image support), + GPT-4.1 +
      • +
    • Visual Q&A: Claude Sonnet 4, GPT-4o, Llama 4 Maverick
      • +
    • Code from screenshots: GPT-4o, Claude Sonnet 4, Gemini 2.5 Pro
      • +
    • + Medical imaging: Specialized models available through various providers +
      • +
    • + Free vision models: Gemma 3 (1B–27B), Qwen3 series, Llama 4 Maverick +
      • +
    +
    +

    → See all 1,487 vision models compared

    + +

    🎤 Audio Input Models

    +

    + 118 models can process audio input — for transcription, voice analysis, and audio + understanding: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderAudio CapabilitiesContext
    gemini-2.5-proGoogleAudio understanding + transcription1M
    gemini-2.5-flashGoogleAudio understanding + transcription1M
    gpt-4o-audio-previewOpenAIAudio input + output128K
    claude-sonnet-4AnthropicAudio transcription200K
    + +

    🖼️ Image Generation Models

    +

    + 28 models can generate images from text descriptions. This is a rapidly growing category: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderCapabilities
    gpt-image-1OpenAIText-to-image, image editing
    dall-e-3OpenAIText-to-image generation
    flux-1.1-proBlack Forest LabsHigh-quality text-to-image
    stable-diffusion-3.5Stability AIOpen-weight text-to-image
    +

    + → See all 28 image generation models +

    + +

    🔊 Audio Output Models

    +

    + 34 models can generate audio output — for text-to-speech, voice cloning, and audio + generation: +

    +
    +

    Key Audio Output Models

    +
      +
    • + GPT-4o Audio Preview: Natural conversation with voice input and output +
      • +
    • Gemini 2.5 Flash: Audio understanding with text response
      • +
    • + Specialized TTS models: Available through various providers for + production voice applications +
      • +
    +
    + +

    🎬 Video Understanding Models

    +

    + 167 models can process video input — for video analysis, summarization, and content + understanding: +

    +
    +

    Top Video Understanding Models

    +
      +
    • + Gemini 2.5 Pro/Flash: Native video input with 1M context, process up to + 1 hour of video +
      • +
    • GPT-4o: Video frame extraction and analysis
      • +
    • Qwen3 series: Video understanding across model sizes
      • +
    +
    + +

    🤔 Choosing the Right Multimodal Model

    +
    +

    Decision Framework

    +
      +
    • + Image understanding only? → gemma-3-27b-it (free) or + gpt-4o +
      • +
    • + Need audio + vision? → gemini-2.5-flash (cheapest + multimodal) or gemini-2.5-pro +
      • +
    • + Generate images? → gpt-image-1 or + flux-1.1-pro +
      • +
    • + Video analysis? → gemini-2.5-pro (best video + understanding) +
      • +
    • + Need tool calling + vision? → llama-4-maverick or + claude-sonnet-4 +
      • +
    • + Budget-conscious? → gemini-2.5-flash ($0.15/$0.60 per 1M + tokens) +
      • +
    • + Need free API? → gemma-3-27b-it (Google, free) or + qwen3-32b (Alibaba, free) +
      • +
    +
    + +

    🔗 Related Comparisons

    +
    + 👁️ Best AI Models for Vision (1,487 models) + 🖼️ Image Generation Models (28 models) + 🔓 Open Source AI Models (527 models) + 🆓 Free AI Models (81 models) + 🧠 Reasoning Models (1,306 models) + 🔧 Tool Calling Models (2,350 models) + 📏 Context Window Comparison + 💰 Cheapest AI Models + ⭐ Best AI Models in 2025 + 🧮 AI Model Pricing Calculator + 🔄 OpenAI Alternatives +
    +
    + + + diff --git a/site/open-source-ai-models.html b/site/open-source-ai-models.html new file mode 100644 index 00000000..38ca6caf --- /dev/null +++ b/site/open-source-ai-models.html @@ -0,0 +1,845 @@ + + + + + + Open Source AI Models — 527 Open Weight LLMs Compared | AI Models Catalog + + + + + + + + + + + + + + + + +
    +
    +

    🔓 Open Source AI Models

    +

    + 527 open-weight LLMs compared — pricing, context windows, tool calling, reasoning, and + vision capabilities +

    +
    +
    +
    527
    +
    Open Weight Models
    +
    +
    +
    81
    +
    Free to Use
    +
    +
    +
    375
    +
    Tool Calling
    +
    +
    +
    231
    +
    Reasoning
    +
    +
    +
    269
    +
    Vision
    +
    +
    + 🔍 Interactive Catalog + ⭐ Star on GitHub +
    +
    +
    +
    + Contents +
      +
    1. Flagship Open-Weight Models
      2. +
    2. Free Open-Source AI Models
      3. +
    3. Open-Weight Tool-Calling Models
      4. +
    4. Open-Weight Reasoning Models
      5. +
    5. Open-Weight Vision Models
      6. +
    6. Open-Weight Large Context Models
      7. +
    7. How to Choose an Open-Source Model
      8. +
    +
    + +

    🏆 Flagship Open-Weight Models

    +

    The most capable open-weight models available today, from leading AI labs:

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextTool CallReasoningPrice (in/out per 1M)
    llama-4-maverickMeta1M✓✗Varies
    llama-4-scoutMeta10M✓✗Varies
    deepseek-r1DeepSeek128K✓✓Varies
    deepseek-v3DeepSeek128K✓✗Varies
    qwen3-235b-a22bAlibaba128K✓✓Varies
    qwen3-32bAlibaba128K✓✓Varies
    llama-3.3-70b-instructMeta128K✓✗Varies
    gemma-3-27b-itGoogle128K✓✗Free
    phi-4Microsoft16K✓✗Varies
    command-aCohere256K✓✗Varies
    mistral-large-2411Mistral128K✓✗Varies
    + +

    🆓 Free Open-Source AI Models

    +

    + 81 open-weight models you can use for free through their provider APIs. These are ideal for + prototyping, testing, and learning: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextTool CallReasoning
    gemma-3-27b-itGoogle128K✓✗
    gemma-3-12b-itGoogle128K✓✗
    gemma-3-4b-itGoogle128K✓✗
    gemma-3-1b-itGoogle128K✗✗
    qwen3-235b-a22bAlibaba128K✓✓
    qwen3-30b-a3bAlibaba128K✓✓
    qwen3-32bAlibaba128K✓✓
    qwen3-14bAlibaba128K✓✓
    qwen3-8bAlibaba128K✓✓
    qwen3-4bAlibaba128K✓✓
    qwen3-1.7bAlibaba128K✓✓
    qwen3-0.6bAlibaba128K✓✓
    llama-4-maverickMeta1M✓✗
    llama-4-scoutMeta10M✓✗
    llama-3.3-70b-instructMeta128K✓✗
    +

    + → See all 81 free AI models (including non-open-weight) +

    + +

    🔧 Open-Weight Tool-Calling Models

    +

    + 375 open-weight models support tool/function calling — essential for AI agents and agentic + workflows: +

    +
    +

    Top Open-Weight Tool-Calling Models

    +
      +
    • + Llama 4 Maverick/Scout — Meta's latest with native tool calling, 1M–10M + context +
      • +
    • + Qwen3 series — All sizes support tool calling + reasoning (0.6B to + 235B) +
      • +
    • DeepSeek R1/V3 — Strong tool calling with 128K context
      • +
    • Gemma 3 (1B–27B) — Google's lightweight models with tool calling
      • +
    • + Command A — Cohere's 111B model optimized for enterprise tool use +
      • +
    • + Mistral Large — 123B parameter model with robust function calling +
      • +
    +
    +

    → See all 2,350 tool-calling models

    + +

    🧠 Open-Weight Reasoning Models

    +

    + 231 open-weight models with reasoning capabilities — these can "think step by step" for + complex tasks: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextTool CallKey Strength
    deepseek-r1DeepSeek128K✓Best open-weight reasoning, rivals o1
    qwen3-235b-a22bAlibaba128K✓MoE architecture, thinking mode
    qwen3-32bAlibaba128K✓Dense reasoning, strong benchmarks
    qwen3-30b-a3bAlibaba128K✓Lightweight MoE reasoning
    qwen3-14bAlibaba128K✓Mid-size reasoning model
    qwen3-8bAlibaba128K✓Small but capable reasoning
    +

    → See all 1,306 reasoning models

    + +

    👁️ Open-Weight Vision Models

    +

    + 269 open-weight models can process images alongside text — useful for document analysis, + visual Q&A, and multimodal applications: +

    +
    +

    Top Open-Weight Vision Models

    +
      +
    • + Llama 4 Maverick/Scout — Native multimodal with 1M–10M context, process + images + text +
      • +
    • Qwen3 series — Vision-capable across all sizes
      • +
    • Gemma 3 (1B–27B) — Google's vision-language models, free to use
      • +
    • DeepSeek R1/V3 — Reasoning + vision capabilities
      • +
    • Command A — Enterprise-grade vision + tool calling
      • +
    +
    +

    → See all 1,487 vision models

    + +

    📏 Open-Weight Large Context Models

    +

    + Open-weight models with the largest context windows — essential for processing long + documents, codebases, and multi-turn conversations: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContext WindowTool CallReasoning
    llama-4-scoutMeta10M✓✗
    llama-4-maverickMeta1M✓✗
    command-aCohere256K✓✗
    deepseek-r1DeepSeek128K✓✓
    deepseek-v3DeepSeek128K✓✗
    qwen3-235b-a22bAlibaba128K✓✓
    llama-3.3-70b-instructMeta128K✓✗
    gemma-3-27b-itGoogle128K✓✗
    mistral-large-2411Mistral128K✓✗
    +

    + → See all models with context window comparison +

    + +

    🤔 How to Choose an Open-Source AI Model

    +
    +

    Decision Framework

    +
      +
    • + Need free API access? → Start with gemma-3-27b-it (Google, + free) or qwen3-32b (Alibaba, free) +
      • +
    • + Building AI agents? → llama-4-maverick (1M context + tool + calling) or deepseek-r1 (reasoning + tools) +
      • +
    • + Processing long documents? → llama-4-scout (10M context) + or llama-4-maverick (1M context) +
      • +
    • + Complex reasoning tasks? → deepseek-r1 (best open-weight + reasoning) or qwen3-235b-a22b +
      • +
    • + Vision/image understanding? → llama-4-maverick or + gemma-3-27b-it +
      • +
    • + Edge/mobile deployment? → qwen3-0.6b or + gemma-3-1b-it (smallest open-weight) +
      • +
    • + Enterprise tool use? → command-a (256K context, optimized + for RAG + tools) +
      • +
    +
    +
    +

    Open Weights vs. Proprietary Models

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    AspectOpen WeightsProprietary
    Self-hosting✓ Run on your own hardware✗ Cloud API only
    Data privacy✓ Full control over data✗ Data sent to provider
    Customization✓ Fine-tune on your data✗ Limited (prompt-based)
    Cost at scale✓ Fixed infra cost✗ Per-token pricing
    Latest capabilities~3–6 months behind✓ Cutting-edge
    ConvenienceRequires infra setup✓ Instant API access
    +
    + +

    🔗 Related Comparisons

    +
    + 🆓 Free AI Models (81 models) + 🧠 Reasoning Models (1,306 models) + 🔧 Tool Calling Models (2,350 models) + 👁️ Vision Models (1,487 models) + 📏 Context Window Comparison + 💰 Cheapest AI Models + 🤖 Best AI Models for Agents + 💵 LLM Pricing Comparison + 🔄 OpenAI Alternatives + ⭐ Best AI Models in 2025 + 🧮 AI Model Pricing Calculator + 🎨 Multimodal AI Models +
    +
    + + + From c2a70841e2aba74360cc2a8c2ff5816d2ffe86fb Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 02:31:38 +0800 Subject: [PATCH 274/311] Add cross-links to new SEO pages and update sitemap All 15 existing SEO pages now link to the new Open Source and Multimodal pages. Sitemap updated to 56 URLs. --- site/ai-model-pricing-calculator.html | 6 +- site/ai-models-by-provider.html | 6 +- site/best-ai-models-for-agents.html | 6 +- site/best-ai-models-for-coding.html | 6 +- site/best-ai-models-for-image-generation.html | 6 +- site/best-ai-models-for-vision.html | 6 +- site/best-ai-models.html | 6 +- site/cheapest-ai-models.html | 6 +- site/context-window-comparison.html | 6 +- site/free-ai-models.html | 6 +- site/llm-pricing.html | 6 +- site/openai-alternatives.html | 6 +- site/reasoning-models-comparison.html | 6 +- site/sitemap.xml | 546 +++++++++--------- site/structured-output-models-comparison.html | 6 +- site/tool-calling-models-comparison.html | 6 +- 16 files changed, 333 insertions(+), 303 deletions(-) diff --git a/site/ai-model-pricing-calculator.html b/site/ai-model-pricing-calculator.html index 6f59d81e..9e441ee8 100644 --- a/site/ai-model-pricing-calculator.html +++ b/site/ai-model-pricing-calculator.html @@ -872,8 +872,10 @@

    🔗 More Resources

    — largest context LLMs
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources

    • — 95 providers compared
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources

    • — largest context LLMs
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources

    • — largest context LLMs
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — LLM cost calculator
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — DALL·E, Imagen, GPT-5 Image compared
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — detailed pricing tables
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — largest context LLMs
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — browse by provider
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — find the cheapest model
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — detailed analysis
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — detailed analysis
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • 🔗 More Resources — largest context LLMs
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • - - - https://i-need-token.github.io/ai-models/ - daily - 1.0 - - - https://i-need-token.github.io/ai-models/docs/agentic-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/api - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/audio-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/cached-pricing - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/chat-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/code-examples - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/code-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/context-windows - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/data-acquisition - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/data-schema - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/embedding-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/faq - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/free-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/glossary - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/image-generation - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/large-context-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/lessons-learned - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/migration-guide - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/modality-matrix - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/model-comparison - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/model-selection - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/multimodal-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/open-weights - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/pricing-comparison - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/provider-comparison - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/providers - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/quick-start - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/reasoning-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/small-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/structured-output - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/tool-calling - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/video-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/vision-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/models.json - daily - 0.9 - - - https://i-need-token.github.io/ai-models/docs/openai-alternatives - weekly - 0.8 - - - https://i-need-token.github.io/ai-models/docs/agentic-models - weekly - 0.8 - - - https://i-need-token.github.io/ai-models/docs/code-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/docs/audio-models - weekly - 0.7 - - - https://i-need-token.github.io/ai-models/best-ai-models.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/free-ai-models.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/llm-pricing.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/openai-alternatives.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/ai-models-by-provider.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/context-window-comparison.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/reasoning-models-comparison.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/cheapest-ai-models.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html - weekly - 0.9 - - - https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html - weekly - 0.9 - + + + + https://i-need-token.github.io/ai-models/ + daily + 1.0 + + + https://i-need-token.github.io/ai-models/docs/agentic-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/api + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/audio-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/cached-pricing + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/chat-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/code-examples + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/code-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/context-windows + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/data-acquisition + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/data-schema + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/embedding-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/faq + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/free-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/glossary + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/image-generation + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/large-context-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/lessons-learned + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/migration-guide + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/modality-matrix + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/model-comparison + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/model-selection + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/multimodal-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/open-weights + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/pricing-comparison + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/provider-comparison + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/providers + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/quick-start + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/reasoning-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/small-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/structured-output + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/tool-calling + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/video-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/vision-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/models.json + daily + 0.9 + + + https://i-need-token.github.io/ai-models/docs/openai-alternatives + weekly + 0.8 + + + https://i-need-token.github.io/ai-models/docs/agentic-models + weekly + 0.8 + + + https://i-need-token.github.io/ai-models/docs/code-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/docs/audio-models + weekly + 0.7 + + + https://i-need-token.github.io/ai-models/best-ai-models.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/free-ai-models.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/llm-pricing.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/openai-alternatives.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/ai-models-by-provider.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/context-window-comparison.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/reasoning-models-comparison.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/cheapest-ai-models.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html + weekly + 0.9 + + + https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html + weekly + 0.9 + - - https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html - weekly - 0.9 - + + https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html + weekly + 0.9 + - - https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html - weekly - 0.9 - + + https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html + weekly + 0.9 + - - https://i-need-token.github.io/ai-models/structured-output-models-comparison.html - weekly - 0.9 - - + + https://i-need-token.github.io/ai-models/structured-output-models-comparison.html + weekly + 0.9 + +https://i-need-token.github.io/ai-models/open-source-ai-models.html0.9weeklyhttps://i-need-token.github.io/ai-models/multimodal-ai-models.html0.9weekly \ No newline at end of file diff --git a/site/structured-output-models-comparison.html b/site/structured-output-models-comparison.html index b21c5a66..082780b3 100644 --- a/site/structured-output-models-comparison.html +++ b/site/structured-output-models-comparison.html @@ -717,8 +717,10 @@

    🔗 More Resources

    — GPT-4o, Claude, Gemini vision compared
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
    • diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html index 22ae87aa..aaeb395d 100644 --- a/site/tool-calling-models-comparison.html +++ b/site/tool-calling-models-comparison.html @@ -1074,8 +1074,10 @@

    🔗 More Resources

    — largest context LLMs
  • - GitHub Repository — star, fork, - contribute + GitHub Repository + 🔓 Open Source AI Models (527 models) + 🎨 Multimodal AI Models (1,548 models) + — star, fork, contribute
  • Date: Fri, 22 May 2026 02:42:02 +0800 Subject: [PATCH 275/311] =?UTF-8?q?Update=20README=20and=20index=20files?= =?UTF-8?q?=20with=202=20new=20SEO=20pages=20(15=E2=86=9217)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - README Ecosystem table: 15→17 SEO pages, added Open Source + Multimodal links - README What's New: updated count - AGENTS.md: added Open Source + Multimodal entries - llms.txt: added 5 missing SEO pages (Image Gen, Vision, Structured Output, Open Source, Multimodal) - llms-full.txt: same additions --- AGENTS.md | 2 ++ README.md | 20 ++++++++++---------- llms-full.txt | 5 +++++ llms.txt | 5 +++++ 4 files changed, 22 insertions(+), 10 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index de57deb6..4751048d 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -62,6 +62,8 @@ Curated standalone pages targeting high-volume search queries. All cross-linked - [`site/best-ai-models-for-image-generation.html`](site/best-ai-models-for-image-generation.html) — Best AI Models for Image Generation (DALL·E, Imagen, GPT-5 Image, Midjourney, cheapest, free, open-weight) - [`site/best-ai-models-for-vision.html`](site/best-ai-models-for-vision.html) — Best AI Models for Vision (GPT-4o, Claude, Gemini, 1,487 vision models, cheapest, free, vision+tool_call, vision+large context) - [`site/structured-output-models-comparison.html`](site/structured-output-models-comparison.html) — Structured Output Models Comparison (829 structured output models, JSON mode, SO+tool_call, SO+reasoning, cheapest, free) +- [`site/open-source-ai-models.html`](site/open-source-ai-models.html) — Open Source AI Models (527 open-weight models, free, tool calling, reasoning, vision, large context) +- [`site/multimodal-ai-models.html`](site/multimodal-ai-models.html) — Multimodal AI Models (1,548 vision/audio/image models, modality breakdown, flagship, free) ## Key Design Decisions diff --git a/README.md b/README.md index 20efc192..f81e15eb 100644 --- a/README.md +++ b/README.md @@ -570,21 +570,21 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 15 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 17 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html) | ## What's New ### v0.2.0 (May 2025) -- **15 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation +- **17 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities diff --git a/llms-full.txt b/llms-full.txt index bef40227..51bcf196 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -401,3 +401,8 @@ npx tsx scripts/export-csv.ts - Cheapest AI Models: site/cheapest-ai-models.html - Tool Calling Models Comparison: site/tool-calling-models-comparison.html - AI Model Pricing Calculator: site/ai-model-pricing-calculator.html +- Best AI Models for Image Generation: site/best-ai-models-for-image-generation.html +- Best AI Models for Vision: site/best-ai-models-for-vision.html +- Structured Output Models Comparison: site/structured-output-models-comparison.html +- Open Source AI Models: site/open-source-ai-models.html +- Multimodal AI Models: site/multimodal-ai-models.html diff --git a/llms.txt b/llms.txt index bef40227..51bcf196 100644 --- a/llms.txt +++ b/llms.txt @@ -401,3 +401,8 @@ npx tsx scripts/export-csv.ts - Cheapest AI Models: site/cheapest-ai-models.html - Tool Calling Models Comparison: site/tool-calling-models-comparison.html - AI Model Pricing Calculator: site/ai-model-pricing-calculator.html +- Best AI Models for Image Generation: site/best-ai-models-for-image-generation.html +- Best AI Models for Vision: site/best-ai-models-for-vision.html +- Structured Output Models Comparison: site/structured-output-models-comparison.html +- Open Source AI Models: site/open-source-ai-models.html +- Multimodal AI Models: site/multimodal-ai-models.html From e086912081ef4209078679cc790aeb758780d4f8 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 03:01:42 +0800 Subject: [PATCH 276/311] =?UTF-8?q?Fix=20interactive=20catalog=20data=20lo?= =?UTF-8?q?ading=20=E2=80=94=20separate=20fetch=20errors=20from=20init=20e?= =?UTF-8?q?rrors?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The previous fetch chain had a subtle bug: if the local fetch succeeded but init() threw an error, the catch() would fall through to CDN and Releases fallbacks (which also fail), showing 'Failed to load data' even though the data was fetched successfully. Now uses async/await with proper try/catch separation: fetch errors try the next URL source, but init() errors are caught and displayed with the actual error message for debugging. --- site/index.html | 58 +++++++++++++++++++++++-------------------------- 1 file changed, 27 insertions(+), 31 deletions(-) diff --git a/site/index.html b/site/index.html index 56a58b97..267cbed7 100644 --- a/site/index.html +++ b/site/index.html @@ -1208,38 +1208,34 @@

    const RELEASES_URL = "https://github.com/i-need-token/ai-models/releases/latest/download/models.json"; - fetch(LOCAL_URL) - .then((r) => { - if (!r.ok) throw new Error("Local not available"); - return r.json(); - }) - .then((data) => { - models = Array.isArray(data) ? data : data.models; - init(); - }) - .catch(() => { - fetch(CDN_URL) - .then((r) => { - if (!r.ok) throw new Error("CDN not available"); - return r.json(); - }) - .then((data) => { - models = Array.isArray(data) ? data : data.models; + async function loadData() { + const urls = [LOCAL_URL, CDN_URL, RELEASES_URL]; + for (const url of urls) { + try { + const r = await fetch(url); + if (!r.ok) continue; + const data = await r.json(); + models = Array.isArray(data) ? data : data.models; + try { init(); - }) - .catch(() => { - fetch(RELEASES_URL) - .then((r) => r.json()) - .then((data) => { - models = Array.isArray(data) ? data : data.models; - init(); - }) - .catch(() => { - document.getElementById("tbody").innerHTML = - 'Failed to load data. Try refreshing the page.'; - }); - }); - }); + return; // success + } catch (e) { + console.error("init() error:", e); + document.getElementById("tbody").innerHTML = + 'Error rendering data: ' + + esc(e.message) + + ""; + return; + } + } catch (e) { + // fetch or parse failed, try next URL + continue; + } + } + document.getElementById("tbody").innerHTML = + 'Failed to load data. Try refreshing the page.'; + } + loadData(); function init() { // Stats From 5ebb6f19371ef2ac3508739b86dbcf7f8c06a49d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 03:15:32 +0800 Subject: [PATCH 277/311] Fix interactive catalog: add missing familyFilter and picker DOM elements MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The catalog was showing 'Failed to load data' because init() threw 'Cannot read properties of null (reading appendChild)' — the familyFilter select and picker wizard elements (picker, pickerTask, pickerBudget, pickerResult) were referenced in JavaScript but missing from the HTML. Also fixed an unclosed container div tag. --- site/index.html | 581 +++++++++++++++++++++++++++--------------------- 1 file changed, 333 insertions(+), 248 deletions(-) diff --git a/site/index.html b/site/index.html index 267cbed7..10ba5a21 100644 --- a/site/index.html +++ b/site/index.html @@ -776,6 +776,43 @@ display: none; } } + .picker-inputs { + display: flex; + gap: 12px; + flex-wrap: wrap; + margin-bottom: 12px; + } + .picker-inputs label { + display: flex; + flex-direction: column; + gap: 4px; + font-size: 0.85em; + color: var(--text-secondary); + } + .picker-inputs select { + padding: 6px 8px; + border: 1px solid var(--border); + border-radius: 4px; + background: var(--bg); + color: var(--text); + } + .picker-result { + font-size: 0.9em; + line-height: 1.8; + } + .pick-item { + padding: 6px 0; + border-bottom: 1px solid var(--border); + } + .pick-item:last-child { + border-bottom: none; + } + .pick-rank { + display: inline-block; + width: 24px; + font-weight: 700; + color: var(--accent); + } @@ -838,274 +875,322 @@

    🤖 AI Models Catalog
    - -
    - - - -
    -
    - - - - - - - - - - - - -
    -
    - - - - -
    - - - - - - - - - - - - -
    Model Provider Context Input $/M Output $/M Capabilities
    -
    -
    - -
    - - + + + + + + + + + + +
    +
    +
    +
    + + + +
    - -
    - ⭐ Star on GitHub · - JSON - · - GitHub Action · - npm · - Hugging Face -
    - Quick Start - · - Model Selection - · - Free Models - · - Tool Calling - · - Reasoning - · - Agentic Models - · - Code Models - · - Pricing - · - Cached Pricing - · - Context Windows - · - Vision - · - API - · - FAQ - · - Chat Models - · - Multimodal - · - Embeddings -
    - 🏆 Best AI Models - · - 🆓 Free Models - · - 💰 LLM Pricing - · - 🔄 OpenAI Alternatives - · - 🔧 Tool Calling Models - · - 💰 Pricing Calculator -
    - · - 👁️ Vision - -

  • diff --git a/site/ai-models-by-provider.html b/site/ai-models-by-provider.html index be872b58..8628d6f8 100644 --- a/site/ai-models-by-provider.html +++ b/site/ai-models-by-provider.html @@ -4520,6 +4520,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/best-ai-models-for-agents.html b/site/best-ai-models-for-agents.html index ba86e0f9..f56b74a9 100644 --- a/site/best-ai-models-for-agents.html +++ b/site/best-ai-models-for-agents.html @@ -939,6 +939,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index 1d58d3df..018e6d7e 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -814,6 +814,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/best-ai-models-for-image-generation.html b/site/best-ai-models-for-image-generation.html index 9613b2c5..d7e194a3 100644 --- a/site/best-ai-models-for-image-generation.html +++ b/site/best-ai-models-for-image-generation.html @@ -750,6 +750,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/best-ai-models-for-vision.html b/site/best-ai-models-for-vision.html index efa2dfce..07f6a8c9 100644 --- a/site/best-ai-models-for-vision.html +++ b/site/best-ai-models-for-vision.html @@ -744,6 +744,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index e6494c67..cec18261 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1059,6 +1059,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index 8fec9cca..15464e08 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1512,6 +1512,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html index 04c6f726..f91c8533 100644 --- a/site/context-window-comparison.html +++ b/site/context-window-comparison.html @@ -1596,6 +1596,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index 6d87ca55..41a5fa68 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1487,6 +1487,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/index.html b/site/index.html index 10ba5a21..a36207eb 100644 --- a/site/index.html +++ b/site/index.html @@ -1067,6 +1067,7 @@

    🤖 AI Models CatalogChat Models · + 📊 State of AI Models 2025 Multimodal diff --git a/site/llm-pricing.html b/site/llm-pricing.html index f823ca9e..c90df5f3 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -862,6 +862,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/multimodal-ai-models.html b/site/multimodal-ai-models.html index dbc739fe..6815ac0a 100644 --- a/site/multimodal-ai-models.html +++ b/site/multimodal-ai-models.html @@ -642,6 +642,7 @@

    🔗 Related Comparisons

    ⭐ Best AI Models in 2025 🧮 AI Model Pricing Calculator 🔄 OpenAI Alternatives + State of AI Models 2025
  • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index ce33def6..0732a587 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -929,6 +929,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • diff --git a/site/state-of-ai-models.html b/site/state-of-ai-models.html new file mode 100644 index 00000000..f8ab7cab --- /dev/null +++ b/site/state-of-ai-models.html @@ -0,0 +1,725 @@ + + + + + + State of AI Models 2025 — Data-Driven Report | AI Models Catalog + + + + + + + + + + + + + + +
    +

    📊 State of AI Models 2025

    +

    + A data-driven analysis of 4,587 AI models across 95 providers — pricing trends, capability + adoption, context window growth, and the rise of open-source AI. +

    +
    +
    +
    +
    +
    4,587
    +
    Total Models
    +
    +
    +
    95
    +
    Providers
    +
    +
    +
    81
    +
    Free Models
    +
    +
    +
    527
    +
    Open-Weight
    +
    +
    +
    2,350
    +
    Tool Calling
    +
    +
    +
    1,306
    +
    Reasoning
    +
    +
    +
    1,487
    +
    Vision
    +
    +
    +
    2,195
    +
    128K+ Context
    +
    +
    + +

    1. Provider Landscape

    +

    + The AI model ecosystem spans 95 providers, from tech giants to specialized startups. The top + 15 providers account for the majority of models: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ProviderModelsNotable Models
    OpenRouter415Aggregator — routes to 100+ models
    Google261Gemini 2.5 Pro/Flash, Gemma 3
    Requesty234Aggregator — unified API
    Cohere197Command R+, Embed v3
    xAI193Grok 3, Grok 3 Mini
    DeepSeek184DeepSeek R1, V3
    Meta163Llama 4 Maverick/Scout
    Mistral155Mistral Large, Codestral
    Alibaba (Qwen)139Qwen3-235B, QwQ
    Anthropic121Claude Sonnet 4, Opus 4
    OpenAI115GPT-4.1, o3, o4-mini
    Microsoft99Phi-4, Florence 2
    Amazon96Nova Pro, Titan
    NVIDIA87Nemotron, Llama Nemotron
    01.ai83Yi-Lightning, Yi-VL
    +
    + Key Insight: Aggregators (OpenRouter, Requesty) offer the widest selection + but may duplicate models available from first-party providers. For the best pricing, go + direct to the source. +
    + +

    2. Pricing Distribution

    +

    + AI model pricing varies dramatically — from completely free to over $15 per million input + tokens. Here is the breakdown of the 4,587 models: +

    +
    +
    +
    Free
    +
    +
    81 models
    +
    +
    +
    < $0.50/M
    +
    +
    ~1,800 models
    +
    +
    +
    $0.50–5/M
    +
    +
    ~1,400 models
    +
    +
    +
    > $5/M
    +
    +
    ~480 models
    +
    +
    +
    + Key Insight: The median input price for tool-calling models is $0.50/M + tokens, while reasoning models median is $0.80/M. Vision-capable models average $1.50/M — + still affordable for most production use cases. +
    + +

    3. Capability Adoption

    +

    + Modern AI models increasingly support advanced capabilities beyond basic text generation: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    CapabilityModels% of TotalAvg Input $/M
    Tool Calling2,35051.2%$1.50
    Reasoning1,30628.5%$2.10
    Structured Output82918.1%$1.80
    Vision (Image Input)1,48732.4%$1.50
    Open Weights52711.5%Free or low-cost
    Image Generation280.6%$3.00+
    Audio Input1182.6%$2.50+
    Audio Output340.7%$3.00+
    Video Input1673.6%$2.00+
    +
    + Key Insight: Over half of all models now support tool calling — it has + become table stakes for production AI. Reasoning capabilities are growing fast, with 1,306 + models (28.5%) supporting extended thinking. +
    + +

    4. Context Window Revolution

    +

    + Context windows have grown exponentially. The average context window across all models is + now approximately 200K tokens: +

    +
    +
    +
    < 32K
    +
    +
    ~800 models
    +
    +
    +
    32K–128K
    +
    +
    ~1,000 models
    +
    +
    +
    128K–1M
    +
    +
    ~2,195 models
    +
    +
    +
    1M+
    +
    +
    ~30 models
    +
    +
    +

    Largest Context Windows

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelContextProvider
    Google Gemini 2.5 Pro1,048,576Google
    Google Gemini 2.5 Flash1,048,576Google
    Meta Llama 4 Scout10,000,000Meta
    Meta Llama 4 Maverick1,048,576Meta
    Google Gemma 3 27B131,072Google
    +
    + Key Insight: 128K+ context is now the norm — 2,195 models (47.8%) support + it. Meta's Llama 4 Scout leads with a 10M token window, making entire codebases and books + processable in a single prompt. +
    + +

    5. The Rise of Free & Open-Source AI

    +

    + 81 models are completely free to use, and 527 have open weights. Here are the most capable + free models: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelContextCapabilitiesProvider
    Google Gemini 2.5 Flash1MTC, Reasoning, Vision, SOGoogle
    DeepSeek R1128KReasoning, TCDeepSeek
    Meta Llama 4 Maverick1MTC, VisionMeta
    Alibaba Qwen3-235B128KTC, Reasoning, SOAlibaba
    Google Gemma 3 27B131KVision, TCGoogle
    +
    + Key Insight: Free models now rival paid ones in capability. Google Gemini + 2.5 Flash (free tier) offers 1M context, tool calling, reasoning, and vision — making it + viable for production use at zero cost. +
    + +

    6. Best Value Models by Use Case

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Use CaseBest FreeBest Paid (Cheapest)Best Overall
    General ChatGemini 2.5 FlashDeepSeek V3 ($0.07/$0.28)Claude Sonnet 4
    CodingDeepSeek R1DeepSeek V3 ($0.07/$0.28)Claude Sonnet 4
    AI AgentsGemini 2.5 FlashGrok 3 Mini ($0.30/$0.50)Claude Sonnet 4
    ReasoningDeepSeek R1Grok 3 Mini ($0.30/$0.50)o3
    VisionGemini 2.5 FlashGemma 3 4B (free)Gemini 2.5 Pro
    Large ContextLlama 4 Scout (10M)Gemini 2.5 Flash ($0.15/$0.60)Gemini 2.5 Pro
    + +

    7. Key Trends & Predictions

    +
    + Trend 1: Agentic AI is the new default. 51% of models support tool calling, + and 1,080 models are classified as "agentic" (tool_call + chat). Expect this to reach 80%+ + by 2026. +
    +
    + Trend 2: Context windows are commoditized. 128K context is now standard. + 1M+ context models are growing, with Google and Meta leading. Expect 10M+ to become common + by 2026. +
    +
    + Trend 3: Free tiers are production-ready. 81 free models with capabilities + like tool calling and reasoning mean that cost is no longer a barrier to entry for AI + development. +
    +
    + Trend 4: Multimodal is mainstream. 1,548 models support more than text + input. Vision (1,487 models) is nearly universal among flagship models. Audio and video are + the next frontiers. +
    +
    + Trend 5: Open weights are accelerating. 527 open-weight models exist, with + Meta's Llama 4 and Alibaba's Qwen3 leading. Expect open-source to match proprietary + capabilities within 6 months. +
    + +
    + Explore More: + Best AI Models + Free AI Models + LLM Pricing + OpenAI Alternatives + Context Windows + Reasoning Models + Tool Calling + Cheapest Models + Open Source + Multimodal + Coding Models + Agent Models + Price Calculator + Interactive Catalog +
    +
    +
    +

    + Data from + AI Models Catalog + — 4,587 models across 95 providers. Updated continuously. +

    +
    + + diff --git a/site/structured-output-models-comparison.html b/site/structured-output-models-comparison.html index 082780b3..43c4c26a 100644 --- a/site/structured-output-models-comparison.html +++ b/site/structured-output-models-comparison.html @@ -720,6 +720,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
    • diff --git a/site/tool-calling-models-comparison.html b/site/tool-calling-models-comparison.html index aaeb395d..8981d51c 100644 --- a/site/tool-calling-models-comparison.html +++ b/site/tool-calling-models-comparison.html @@ -1077,6 +1077,7 @@

    🔗 More Resources

    GitHub Repository 🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) + State of AI Models 2025 — star, fork, contribute
  • From 1854c2a2de63a6c16b63a943bcf0d532b000401f Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 08:46:52 +0800 Subject: [PATCH 280/311] Add bilingual benchmarks & leaderboards doc (docs/benchmarks.md + zh) New doc covering: MMLU, MATH-500, HumanEval, SWE-bench, BFCL, Chatbot Arena, and other key benchmarks. Includes leaderboard landscape, limitations, and practical recommendations for combining benchmarks with catalog data. Updates AGENTS.md, llms.txt, llms-full.txt. --- AGENTS.md | 1 + docs/benchmarks.md | 109 ++++++++++++++++++++++++++++++++++++++++++ docs/zh/benchmarks.md | 109 ++++++++++++++++++++++++++++++++++++++++++ llms-full.txt | 1 + llms.txt | 1 + 5 files changed, 221 insertions(+) create mode 100644 docs/benchmarks.md create mode 100644 docs/zh/benchmarks.md diff --git a/AGENTS.md b/AGENTS.md index 4751048d..a371f062 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -37,6 +37,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) - [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) - [`docs/model-selection.md`](docs/model-selection.md) — Model selection guide: free, best value, large context ([中文](docs/zh/model-selection.md)) +- [`docs/benchmarks.md`](docs/benchmarks.md) — AI Model Benchmarks & Leaderboards: key benchmarks, leaderboard landscape, interpretation guide ([中文](docs/zh/benchmarks.md)) - [`docs/migration-guide.md`](docs/migration-guide.md) — Switch providers: pricing, API compatibility, checklist ([中文](docs/zh/migration-guide.md)) - [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) - [`docs/code-examples.md`](docs/code-examples.md) — code examples in multiple languages ([中文](docs/zh/code-examples.md)) diff --git a/docs/benchmarks.md b/docs/benchmarks.md new file mode 100644 index 00000000..d024a867 --- /dev/null +++ b/docs/benchmarks.md @@ -0,0 +1,109 @@ +# AI Model Benchmarks & Leaderboards + +[中文](zh/benchmarks.md) + +How AI models are evaluated — key benchmarks, leaderboard landscape, and what the numbers mean for model selection. + +Data sourced from the [AI Models Catalog](https://github.com/i-need-token/ai-models). + +## Why Benchmarks Matter + +Benchmarks provide standardized ways to compare AI models across tasks. However, no single benchmark tells the whole story. This guide covers the major benchmarks, how to interpret them, and how to use them alongside our catalog data (pricing, context windows, capabilities) for informed model selection. + +## Major Benchmarks + +### General Language Understanding + +| Benchmark | What It Tests | Top Models | Notes | +| --------- | ---------------------------------- | -------------------------------------- | ------------------------------------------------------------- | +| MMLU | Multi-task knowledge (57 subjects) | GPT-4.1, Claude Opus 4, Gemini 2.5 Pro | Standard academic benchmark; may not reflect real-world usage | +| MMLU-Pro | Harder MMLU with reasoning | o3, Claude Sonnet 4, Gemini 2.5 Pro | More challenging version | +| GPQA | Graduate-level science Q&A | o3, Gemini 2.5 Pro | Expert-level reasoning | +| HellaSwag | Common-sense reasoning | Most frontier models near ceiling | Near-saturated | + +### Reasoning & Math + +| Benchmark | What It Tests | Top Models | Notes | +| ------------- | ----------------------- | --------------------------- | -------------------------- | +| MATH-500 | Competition mathematics | o3, DeepSeek R1, Qwen3-235B | Key for quantitative tasks | +| AIME 2024 | Math competition | o3, DeepSeek R1 | Very challenging | +| GSM8K | Grade-school math | Most models >90% | Near-saturated | +| ARC-Challenge | Scientific reasoning | Most frontier models | Grade-school science | + +### Coding + +| Benchmark | What It Tests | Top Models | Notes | +| ------------- | ---------------------------- | ------------------------------------- | ----------------------------- | +| HumanEval | Python code generation | Claude Sonnet 4, GPT-4.1, DeepSeek V3 | 164 Python problems | +| SWE-bench | Real GitHub issue resolution | Claude Sonnet 4, o3 | More realistic than HumanEval | +| LiveCodeBench | Continuously updated coding | Various | Avoids data contamination | +| MBPP | Basic Python programming | Most models >80% | Near-saturated | + +### Multimodal + +| Benchmark | What It Tests | Top Models | Notes | +| --------- | ------------------------- | ------------------------------- | ------------------------------- | +| MMMU | Multi-modal understanding | Gemini 2.5 Pro, Claude Sonnet 4 | Images + text | +| MathVista | Visual math reasoning | Gemini 2.5 Pro | Diagrams + math | +| AI2D | Science diagrams | Gemini 2.5 Pro | Scientific figure understanding | +| DocVQA | Document understanding | Gemini 2.5 Pro | Text in images | + +### Tool Use & Agents + +| Benchmark | What It Tests | Top Models | Notes | +| --------- | ------------------------- | ------------------------ | ------------------------------------- | +| BFCL v3 | Function calling accuracy | GPT-4.1, Claude Sonnet 4 | Berkeley Function Calling Leaderboard | +| τ-bench | Agent task completion | Various | Terminal-based agent tasks | +| WebArena | Web interaction | Various | Realistic web tasks | + +## Key Leaderboards + +| Leaderboard | Focus | URL | +| -------------------- | ------------------------- | ----------------------------------------------------------------------- | +| LMSYS Chatbot Arena | Human preference ranking | https://chat.lmsys.org/ | +| Open LLM Leaderboard | Open-source model ranking | https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard | +| AlpacaEval | Instruction-following | https://tatsu-lab.github.io/alpaca_eval/ | +| MT-Bench | Multi-turn conversation | Part of Chatbot Arena | +| BigBench | Beyond basic tasks | https://github.com/google/BIG-bench | +| MTEB | Embedding models | https://huggingface.co/spaces/mteb/leaderboard | + +## How to Use Benchmarks with Our Catalog + +Benchmarks alone are insufficient for model selection. Combine them with our catalog data: + +1. **Start with your use case** → See [Model Selection Guide](model-selection.md) +2. **Filter by capabilities** → Tool calling, reasoning, vision, etc. +3. **Check benchmark scores** → For your specific task domain +4. **Compare pricing** → Use our [Pricing Comparison](pricing-comparison.md) +5. **Consider context windows** → See [Context Windows](context-windows.md) +6. **Test with your data** → Benchmarks are proxies; real performance may differ + +## Benchmark Limitations + +- **Data contamination**: Models may have seen benchmark data during training +- **Task narrowness**: Benchmarks test specific skills, not general utility +- **Leaderboard gaming**: Optimizing for benchmarks can hurt real-world performance +- **Staleness**: Benchmarks age as models improve; saturated benchmarks become uninformative +- **Cultural bias**: Most benchmarks are English-centric and Western-focused +- **Cost blindness**: Benchmarks ignore pricing, latency, and availability + +## Practical Recommendations + +- For **coding**: Use SWE-bench over HumanEval (more realistic) +- For **agents**: Test with your actual tool suite; BFCL is a starting point +- For **reasoning**: MATH-500 and GPQA are more discriminating than MMLU +- For **chat**: Chatbot Arena correlates best with human preference +- For **cost-sensitive**: Use our [Free Models](free-models.md) guide first + +## Related Documentation + +- [Model Selection Guide](model-selection.md) — Decision framework for choosing models +- [Pricing Comparison](pricing-comparison.md) — Cost analysis across providers +- [Free Models](free-models.md) — 81 free models with capabilities +- [Tool Calling Models](tool-calling.md) — 2,350 models with function calling +- [Reasoning Models](reasoning-models.md) — 1,306 models with extended thinking +- [Vision Models](vision-models.md) — 1,487 models with image understanding +- [Code Models](code-models.md) — Models optimized for programming +- [Open Weights](open-weights.md) — 527 open-weight models +- [Context Windows](context-windows.md) — Context window comparison +- [Interactive Catalog](https://i-need-token.github.io/ai-models/) — Browse and compare all models diff --git a/docs/zh/benchmarks.md b/docs/zh/benchmarks.md new file mode 100644 index 00000000..4c11dd1c --- /dev/null +++ b/docs/zh/benchmarks.md @@ -0,0 +1,109 @@ +# AI 模型基准测试与排行榜 + +[English](../benchmarks.md) + +AI 模型如何被评估 — 关键基准测试、排行榜格局，以及这些数字对模型选择的意义。 + +数据来源：[AI Models Catalog](https://github.com/i-need-token/ai-models)。 + +## 为什么基准测试很重要 + +基准测试提供了跨任务比较 AI 模型的标准化方式。然而，没有任何单一基准测试能说明全部问题。本指南涵盖主要基准测试、如何解读它们，以及如何将它们与我们的目录数据（定价、上下文窗口、能力）结合使用，以做出明智的模型选择。 + +## 主要基准测试 + +### 通用语言理解 + +| 基准测试 | 测试内容 | 顶级模型 | 备注 | +| --------- | ---------------------- | -------------------------------------- | -------------------------------- | +| MMLU | 多任务知识（57个学科） | GPT-4.1, Claude Opus 4, Gemini 2.5 Pro | 标准学术基准；可能不反映实际使用 | +| MMLU-Pro | 更难的 MMLU，需要推理 | o3, Claude Sonnet 4, Gemini 2.5 Pro | 更具挑战性的版本 | +| GPQA | 研究生水平科学问答 | o3, Gemini 2.5 Pro | 专家级推理 | +| HellaSwag | 常识推理 | 大多数前沿模型接近满分 | 接近饱和 | + +### 推理与数学 + +| 基准测试 | 测试内容 | 顶级模型 | 备注 | +| ------------- | -------- | --------------------------- | ------------------ | +| MATH-500 | 竞赛数学 | o3, DeepSeek R1, Qwen3-235B | 量化任务的关键指标 | +| AIME 2024 | 数学竞赛 | o3, DeepSeek R1 | 非常有挑战性 | +| GSM8K | 小学数学 | 大多数模型 >90% | 接近饱和 | +| ARC-Challenge | 科学推理 | 大多数前沿模型 | 小学科学 | + +### 编程 + +| 基准测试 | 测试内容 | 顶级模型 | 备注 | +| ------------- | -------------------- | ------------------------------------- | ------------------- | +| HumanEval | Python 代码生成 | Claude Sonnet 4, GPT-4.1, DeepSeek V3 | 164 个 Python 问题 | +| SWE-bench | 真实 GitHub 问题修复 | Claude Sonnet 4, o3 | 比 HumanEval 更真实 | +| LiveCodeBench | 持续更新的编程测试 | 各种 | 避免数据污染 | +| MBPP | 基础 Python 编程 | 大多数模型 >80% | 接近饱和 | + +### 多模态 + +| 基准测试 | 测试内容 | 顶级模型 | 备注 | +| --------- | ------------ | ------------------------------- | ------------ | +| MMMU | 多模态理解 | Gemini 2.5 Pro, Claude Sonnet 4 | 图像 + 文本 | +| MathVista | 视觉数学推理 | Gemini 2.5 Pro | 图表 + 数学 | +| AI2D | 科学图表 | Gemini 2.5 Pro | 科学图表理解 | +| DocVQA | 文档理解 | Gemini 2.5 Pro | 图像中的文本 | + +### 工具使用与智能体 + +| 基准测试 | 测试内容 | 顶级模型 | 备注 | +| -------- | -------------- | ------------------------ | ----------------------- | +| BFCL v3 | 函数调用准确率 | GPT-4.1, Claude Sonnet 4 | Berkeley 函数调用排行榜 | +| τ-bench | 智能体任务完成 | 各种 | 基于终端的智能体任务 | +| WebArena | 网页交互 | 各种 | 真实网页任务 | + +## 关键排行榜 + +| 排行榜 | 侧重 | URL | +| -------------------- | ------------ | ----------------------------------------------------------------------- | +| LMSYS Chatbot Arena | 人类偏好排名 | https://chat.lmsys.org/ | +| Open LLM Leaderboard | 开源模型排名 | https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard | +| AlpacaEval | 指令遵循 | https://tatsu-lab.github.io/alpaca_eval/ | +| MT-Bench | 多轮对话 | Chatbot Arena 的一部分 | +| BigBench | 超越基础任务 | https://github.com/google/BIG-bench | +| MTEB | 嵌入模型 | https://huggingface.co/spaces/mteb/leaderboard | + +## 如何将基准测试与我们的目录结合使用 + +仅靠基准测试不足以进行模型选择。将它们与我们的目录数据结合使用： + +1. **从你的用例开始** → 参见[模型选择指南](model-selection.md) +2. **按能力筛选** → 工具调用、推理、视觉等 +3. **查看基准测试分数** → 针对你的特定任务领域 +4. **比较定价** → 使用我们的[定价比较](pricing-comparison.md) +5. **考虑上下文窗口** → 参见[上下文窗口](context-windows.md) +6. **用你的数据测试** → 基准测试是代理；实际性能可能不同 + +## 基准测试的局限性 + +- **数据污染**：模型可能在训练期间见过基准测试数据 +- **任务狭窄**：基准测试测试特定技能，而非通用实用性 +- **排行榜博弈**：为基准测试优化可能损害实际性能 +- **时效性**：随着模型改进，基准测试老化；饱和的基准测试变得无信息量 +- **文化偏见**：大多数基准测试以英语和西方为中心 +- **成本盲区**：基准测试忽略定价、延迟和可用性 + +## 实用建议 + +- **编程**：使用 SWE-bench 而非 HumanEval（更真实） +- **智能体**：用你实际的工具套件测试；BFCL 是起点 +- **推理**：MATH-500 和 GPQA 比 MMLU 更有区分度 +- **聊天**：Chatbot Arena 与人类偏好最相关 +- **成本敏感**：先使用我们的[免费模型](free-models.md)指南 + +## 相关文档 + +- [模型选择指南](model-selection.md) — 选择模型的决策框架 +- [定价比较](pricing-comparison.md) — 跨提供商成本分析 +- [免费模型](free-models.md) — 81 个免费模型及其能力 +- [工具调用模型](tool-calling.md) — 2,350 个支持函数调用的模型 +- [推理模型](reasoning-models.md) — 1,306 个支持扩展思考的模型 +- [视觉模型](vision-models.md) — 1,487 个支持图像理解的模型 +- [编程模型](code-models.md) — 针对编程优化的模型 +- [开放权重](open-weights.md) — 527 个开放权重模型 +- [上下文窗口](context-windows.md) — 上下文窗口比较 +- [交互式目录](https://i-need-token.github.io/ai-models/) — 浏览和比较所有模型 diff --git a/llms-full.txt b/llms-full.txt index fef01548..f70e8377 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -315,6 +315,7 @@ by_context = sorted( - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds - [Model Selection Guide](docs/model-selection.md) — decision framework: free, best value, large context models +- [Benchmarks & Leaderboards](docs/benchmarks.md) — key benchmarks, leaderboard landscape, interpretation guide - [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq - [FAQ](docs/faq.md) — frequently asked questions about the catalog, data, and contributing diff --git a/llms.txt b/llms.txt index a14cd37f..7ed92566 100644 --- a/llms.txt +++ b/llms.txt @@ -315,6 +315,7 @@ by_context = sorted( - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds - [Model Selection Guide](docs/model-selection.md) — decision framework: free, best value, large context models +- [Benchmarks & Leaderboards](docs/benchmarks.md) — key benchmarks, leaderboard landscape, interpretation guide - [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq - [FAQ](docs/faq.md) — frequently asked questions about the catalog, data, and contributing From a01896a39c3b2096f9f0ee537b1ccaab222b81f9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 09:47:08 +0800 Subject: [PATCH 281/311] Add AI Model Benchmarks Comparison SEO page New page covering MMLU, MATH-500, HumanEval, SWE-bench, GPQA, BFCL, Chatbot Arena benchmarks with model comparisons and best-value picks. Includes cross-links from 9 other pages and sitemap entry. --- site/ai-model-benchmarks.html | 649 +++++++++++++++++++++++ site/best-ai-models-for-coding.html | 1 + site/best-ai-models.html | 1 + site/cheapest-ai-models.html | 1 + site/free-ai-models.html | 1 + site/llm-pricing.html | 1 + site/reasoning-models-comparison.html | 1 + site/state-of-ai-models.html | 1 + site/tool-calling-models-comparison.html | 1 + 9 files changed, 657 insertions(+) create mode 100644 site/ai-model-benchmarks.html diff --git a/site/ai-model-benchmarks.html b/site/ai-model-benchmarks.html new file mode 100644 index 00000000..87cb65ca --- /dev/null +++ b/site/ai-model-benchmarks.html @@ -0,0 +1,649 @@ + + + + + + + AI Model Benchmarks Comparison 2025 — MMLU, MATH, HumanEval, SWE-bench | AI Models Catalog + + + + + + + + + + + + + + + +
    +

    📊 AI Model Benchmarks Comparison 2025

    +

    + How do top AI models compare on MMLU, MATH-500, HumanEval, SWE-bench, and Chatbot Arena? A + comprehensive benchmark analysis of 4,587 models across 95 providers. +

    +
    +
    +

    1. General Knowledge — MMLU & MMLU-Pro

    +

    + MMLU (Massive Multitask Language Understanding) tests knowledge across 57 academic subjects. + MMLU-Pro is a harder variant requiring deeper reasoning. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelMMLUMMLU-ProProviderInput $/M
    GPT-4.1~90%~78%OpenAI$2.00
    Claude Opus 4~90%~78%Anthropic$15.00
    Gemini 2.5 Pro~90%~78%Google$1.25
    Claude Sonnet 4~88%~76%Anthropic$3.00
    Grok 3~87%~75%xAI$3.00
    DeepSeek R1~85%~72%DeepSeekFree
    Qwen3-235B~85%~72%AlibabaFree
    Llama 4 Maverick~82%~68%MetaFree
    +
    + Key Insight: MMLU is near-saturated for frontier models. Use MMLU-Pro or + GPQA for more discriminating comparisons. +
    + +

    2. Mathematics — MATH-500 & AIME

    +

    + MATH-500 tests competition-level mathematics. AIME 2024 is an even harder math competition + benchmark. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelMATH-500AIME 2024ProviderInput $/M
    o3~96%~83%OpenAI$2.00
    o4-mini~93%~75%OpenAI$1.10
    DeepSeek R1~92%~72%DeepSeekFree
    Gemini 2.5 Pro~91%~70%Google$1.25
    Qwen3-235B~90%~68%AlibabaFree
    Claude Sonnet 4~88%~65%Anthropic$3.00
    +
    + Key Insight: Reasoning models (o3, DeepSeek R1) dominate math benchmarks. + For cost-sensitive math tasks, DeepSeek R1 is free and performs near o3. +
    + +

    3. Coding — HumanEval & SWE-bench

    +

    + HumanEval tests Python code generation. SWE-bench tests real GitHub issue resolution — more + realistic for production use. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelHumanEvalSWE-bench VerifiedProviderInput $/M
    Claude Sonnet 4~93%~72%Anthropic$3.00
    o3~92%~70%OpenAI$2.00
    GPT-4.1~91%~65%OpenAI$2.00
    Gemini 2.5 Pro~90%~63%Google$1.25
    DeepSeek V3~88%~55%DeepSeek$0.07
    Codestral~86%N/AMistral$0.30
    +
    + Key Insight: SWE-bench is more realistic than HumanEval. Claude Sonnet 4 + leads on SWE-bench. For budget coding, DeepSeek V3 at $0.07/M offers remarkable value. +
    + +

    4. Science & Reasoning — GPQA

    +

    + GPQA (Graduate-Level Google-Proof Q&A) tests expert-level scientific reasoning. Even + PhDs with internet access struggle. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelGPQA DiamondProviderInput $/M
    o3~80%OpenAI$2.00
    Gemini 2.5 Pro~78%Google$1.25
    Claude Opus 4~75%Anthropic$15.00
    o4-mini~73%OpenAI$1.10
    DeepSeek R1~71%DeepSeekFree
    + +

    5. Tool Calling — BFCL v3

    +

    + BFCL (Berkeley Function Calling Leaderboard) tests function calling accuracy — critical for + AI agents. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelBFCL v3ProviderInput $/M
    GPT-4.1~88%OpenAI$2.00
    Claude Sonnet 4~86%Anthropic$3.00
    Gemini 2.5 Pro~85%Google$1.25
    Grok 3~83%xAI$3.00
    Gemini 2.5 Flash~82%GoogleFree
    +
    + Key Insight: 2,350 models in our catalog support tool calling. GPT-4.1 + leads on BFCL, but Gemini 2.5 Flash offers strong performance for free. +
    + +

    6. Human Preference — Chatbot Arena

    +

    + LMSYS Chatbot Arena uses blind human comparisons. This is the most practical benchmark for + chat quality. +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelArena ScoreProviderInput $/M
    GPT-4.1~1380OpenAI$2.00
    Claude Sonnet 4~1370Anthropic$3.00
    Gemini 2.5 Pro~1360Google$1.25
    Grok 3~1350xAI$3.00
    DeepSeek R1~1330DeepSeekFree
    +
    + Key Insight: Chatbot Arena correlates best with real-world chat quality. + The top 5 models are very close — pricing and features should drive your decision. +
    + +

    7. Best Value by Benchmark

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    BenchmarkBest FreeBest PaidBest Overall
    MMLUDeepSeek R1 / Qwen3Gemini 2.5 Pro ($1.25)GPT-4.1
    MATHDeepSeek R1o4-mini ($1.10)o3
    CodingDeepSeek V3 ($0.07)Gemini 2.5 Pro ($1.25)Claude Sonnet 4
    GPQADeepSeek R1Gemini 2.5 Pro ($1.25)o3
    Tool CallingGemini 2.5 FlashGemini 2.5 Pro ($1.25)GPT-4.1
    ChatDeepSeek R1Gemini 2.5 Pro ($1.25)GPT-4.1
    + +

    8. Benchmark Limitations

    +
    + Data contamination: Models may have seen benchmark data during training. + Prefer LiveCodeBench over HumanEval for coding. +
    +
    + Task narrowness: Benchmarks test specific skills. Real-world performance + may differ significantly. +
    +
    + Cost blindness: Benchmarks ignore pricing, latency, and availability. + Always combine with our + pricing data. +
    +
    + Staleness: Saturated benchmarks (GSM8K, HellaSwag) are uninformative. Focus + on harder benchmarks like GPQA and SWE-bench. +
    + +
    + Explore More: + Best AI Models + Free AI Models + LLM Pricing + Reasoning Models + Tool Calling + Coding Models + Cheapest Models + Price Calculator + State of AI Models 2025 + Interactive Catalog +
    +
    +
    +

    + Data from AI Models Catalog — 4,587 + models across 95 providers. Updated continuously. +

    +
    + + diff --git a/site/best-ai-models-for-coding.html b/site/best-ai-models-for-coding.html index 018e6d7e..29610fbe 100644 --- a/site/best-ai-models-for-coding.html +++ b/site/best-ai-models-for-coding.html @@ -815,6 +815,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Benchmarks — star, fork, contribute
  • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index cec18261..946eb900 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1060,6 +1060,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Benchmarks — star, fork, contribute
  • diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index 15464e08..5ab40a6a 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1513,6 +1513,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Benchmarks — star, fork, contribute
  • diff --git a/site/free-ai-models.html b/site/free-ai-models.html index 41a5fa68..22c4c31a 100644 --- a/site/free-ai-models.html +++ b/site/free-ai-models.html @@ -1488,6 +1488,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Benchmarks — star, fork, contribute
  • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index c90df5f3..ff91a883 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -863,6 +863,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Benchmarks — star, fork, contribute
  • diff --git a/site/reasoning-models-comparison.html b/site/reasoning-models-comparison.html index 0732a587..647aad08 100644 --- a/site/reasoning-models-comparison.html +++ b/site/reasoning-models-comparison.html @@ -930,6 +930,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Benchmarks — star, fork, contribute
  • diff --git a/site/state-of-ai-models.html b/site/state-of-ai-models.html index f8ab7cab..e5f2af14 100644 --- a/site/state-of-ai-models.html +++ b/site/state-of-ai-models.html @@ -712,6 +712,7 @@

    7. Key Trends & Predictions

    Agent Models Price Calculator Interactive Catalog + Benchmarks
  • From 4b3a8ffd67a099e7df3f76f88f8aabf298c62650 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:23:53 +0800 Subject: [PATCH 282/311] Update README: 19 SEO pages, 70 docs, add Benchmarks link MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - SEO page count: 18→19 (added AI Model Benchmarks Comparison) - Doc count: 68→70 (added benchmarks.md + zh/benchmarks.md) - Added Benchmarks link to Ecosystem table - Updated What's New section --- README.md | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/README.md b/README.md index c64503d3..2502178f 100644 --- a/README.md +++ b/README.md @@ -570,27 +570,27 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 18 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html), [State of AI](https://i-need-token.github.io/ai-models/state-of-ai-models.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 19 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html), [State of AI](https://i-need-token.github.io/ai-models/state-of-ai-models.html), [Benchmarks](https://i-need-token.github.io/ai-models/ai-model-benchmarks.html) | ## What's New ### v0.2.0 (May 2025) -- **18 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal, State of AI Models 2025 +- **19 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal, State of AI Models 2025 - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities - **GitHub Action v2** — version, format, and output-dir inputs - **npm package** — `npm install ai-models` -- **68 docs** — 34 EN + 34 ZH, all bilingual, all cross-linked +- **70 docs** — 35 EN + 35 ZH, all bilingual, all cross-linked ## Roadmap From b9a8c181d1c4f1b3365ae1d84b272dcf5a7e58de Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:24:59 +0800 Subject: [PATCH 283/311] Add Benchmarks & Leaderboards to README Documentation table --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 2502178f..95fe1752 100644 --- a/README.md +++ b/README.md @@ -463,6 +463,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [Model Selection Guide](docs/model-selection.md) | Decision framework: free, best value, large context models | +| [Benchmarks & Leaderboards](docs/benchmarks.md) | MMLU, MATH, HumanEval, SWE-bench, Chatbot Arena guide | | [Migration Guide](docs/migration-guide.md) | Switch providers — pricing, API compatibility, checklist | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | | [Code Examples](docs/code-examples.md) | Practical examples in TypeScript, Python, Go, Rust, jq | From 0aa3fbf65f217951ee1a1577bfa81fa5074c8bc9 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:27:16 +0800 Subject: [PATCH 284/311] Add AI Model Benchmarks SEO page to llms.txt and llms-full.txt --- llms-full.txt | 2 ++ llms.txt | 1 + 2 files changed, 3 insertions(+) diff --git a/llms-full.txt b/llms-full.txt index f70e8377..d2a3943f 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -408,4 +408,6 @@ npx tsx scripts/export-csv.ts - Open Source AI Models: site/open-source-ai-models.html - Multimodal AI Models: site/multimodal-ai-models.html - State of AI Models 2025: site/state-of-ai-models.html +- AI Model Benchmarks Comparison: site/ai-model-benchmarks.html - State of AI Models 2025: site/state-of-ai-models.html +- AI Model Benchmarks Comparison: site/ai-model-benchmarks.html diff --git a/llms.txt b/llms.txt index 7ed92566..5cb4f27a 100644 --- a/llms.txt +++ b/llms.txt @@ -408,3 +408,4 @@ npx tsx scripts/export-csv.ts - Open Source AI Models: site/open-source-ai-models.html - Multimodal AI Models: site/multimodal-ai-models.html - State of AI Models 2025: site/state-of-ai-models.html +- AI Model Benchmarks Comparison: site/ai-model-benchmarks.html From 7f84188bd44fd1e46783756ced161e00f9daedf1 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:41:53 +0800 Subject: [PATCH 285/311] Add ChatGPT vs Claude vs Gemini comparison SEO page New page targeting high-volume search query with head-to-head comparison of GPT-4.1, Claude Sonnet 4, and Gemini 2.5 Pro. Covers pricing, capabilities, benchmarks, API features, and budget alternatives. Includes cross-links from 6 related pages and sitemap entry. --- site/ai-model-benchmarks.html | 1 + site/best-ai-models-for-coding.html | 1 + site/best-ai-models.html | 1 + site/chatgpt-vs-claude-vs-gemini.html | 627 ++++++++++++++++++++++++++ site/cheapest-ai-models.html | 1 + site/llm-pricing.html | 1 + site/openai-alternatives.html | 1 + 7 files changed, 633 insertions(+) create mode 100644 site/chatgpt-vs-claude-vs-gemini.html diff --git a/site/ai-model-benchmarks.html b/site/ai-model-benchmarks.html index 87cb65ca..e31555f4 100644 --- a/site/ai-model-benchmarks.html +++ b/site/ai-model-benchmarks.html @@ -637,6 +637,7 @@

    8. Benchmark Limitations

    Price Calculator State of AI Models 2025 Interactive Catalog + ChatGPT vs Claude vs Gemini
  • diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 946eb900..885f3db8 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1061,6 +1061,7 @@

    🔗 More Resources

    🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 Benchmarks + ChatGPT vs Claude vs Gemini — star, fork, contribute
  • diff --git a/site/chatgpt-vs-claude-vs-gemini.html b/site/chatgpt-vs-claude-vs-gemini.html new file mode 100644 index 00000000..9d9afa54 --- /dev/null +++ b/site/chatgpt-vs-claude-vs-gemini.html @@ -0,0 +1,627 @@ + + + + + + ChatGPT vs Claude vs Gemini — 2025 Comparison | AI Models Catalog + + + + + + + + + + + + + + +
    +

    ⚡ ChatGPT vs Claude vs Gemini

    +

    + The definitive 2025 comparison: pricing, context windows, capabilities, benchmarks, and API + features. GPT-4.1 vs Claude Sonnet 4 vs Gemini 2.5 Pro. +

    +
    +
    +

    1. Flagship Models at a Glance

    +
    +
    +

    OpenAI GPT-4.1

    +
    $2 / $8
    +
    Input / Output per M tokens
    +
    1,047,576 context
    +
    +
    +

    Anthropic Claude Sonnet 4

    +
    $3 / $15
    +
    Input / Output per M tokens
    +
    200,000 context
    +
    +
    +

    Google Gemini 2.5 Pro

    +
    $1.25 / $10
    +
    Input / Output per M tokens
    +
    1,048,576 context
    +
    +
    + +

    2. Pricing Comparison

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    FeatureGPT-4.1Claude Sonnet 4Gemini 2.5 Pro
    Input price ($/M tokens)$2.00$3.00$1.25
    Output price ($/M tokens)$8.00$15.00$10.00
    Cache input ($/M tokens)$0.50$0.30$0.07
    Context window1,047,576200,0001,048,576
    Max output tokens32,76864,00065,536
    Free tierNoYes (limited)Yes (generous)
    +
    + Winner on price: Gemini 2.5 Pro offers the best input pricing ($1.25/M) and + cache pricing ($0.07/M). GPT-4.1 wins on output pricing ($8/M vs $10-15/M). +
    + +

    3. Capabilities

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    CapabilityGPT-4.1Claude Sonnet 4Gemini 2.5 Pro
    Tool calling✅✅✅
    Structured output✅✅✅
    Reasoning (extended thinking)❌ (use o3)✅✅
    Vision (image input)✅✅✅
    Image generation✅ (DALL-E)❌✅ (Imagen)
    Audio input✅❌✅
    Audio output✅❌✅
    Video input❌❌✅
    PDF input✅✅✅
    Code execution✅✅ (analysis tool)✅
    +
    + Winner on capabilities: Gemini 2.5 Pro has the broadest multimodal support + (video, audio I/O, image generation). Claude Sonnet 4 excels at coding and analysis. GPT-4.1 + has the strongest tool calling (BFCL #1). +
    + +

    4. Benchmark Performance

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    BenchmarkGPT-4.1Claude Sonnet 4Gemini 2.5 Pro
    MMLU~90%~88%~90%
    MATH-500~85%~88%~91%
    HumanEval~91%~93%~90%
    SWE-bench Verified~65%~72%~63%
    GPQA Diamond~72%~70%~78%
    BFCL v3 (tool calling)~88%~86%~85%
    Chatbot Arena~1380~1370~1360
    +
    + Key takeaway: No single model wins all benchmarks. GPT-4.1 leads on tool + calling and chat. Claude Sonnet 4 dominates coding (SWE-bench). Gemini 2.5 Pro excels at + math and science. +
    + +

    5. API & Developer Experience

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    FeatureOpenAIAnthropicGoogle
    API maturityMost matureMatureMaturing
    SDK languagesPython, Node, Go, etc.Python, NodePython, Node, Go, etc.
    Streaming✅ SSE✅ SSE✅ SSE
    Function callingParallel, strict modeParallel, forced toolParallel, auto
    Batch API✅ (50% discount)✅ (50% discount)✅ (50% discount)
    Fine-tuning✅❌✅ (limited)
    Rate limitsTier-basedTier-basedPer-project
    + +

    6. Budget Alternatives

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Use CaseBest Budget OptionPriceWhy
    General chatGemini 2.5 FlashFreeStrong quality at zero cost
    CodingDeepSeek V3$0.07/$0.27Near-frontier coding at 1/30th the price
    ReasoningDeepSeek R1FreeTop-tier reasoning at zero cost
    Tool callingGemini 2.5 FlashFreeStrong BFCL scores for free
    Long contextGemini 2.5 FlashFree1M context window for free
    Open sourceQwen3-235BFreeBest open-weight model
    + +

    7. The Verdict

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    If you need...ChooseBecause
    Best overall valueGemini 2.5 ProLowest input price, 1M context, broadest capabilities
    Best coding assistantClaude Sonnet 4#1 on SWE-bench, 64K output, analysis tool
    Best tool callingGPT-4.1#1 on BFCL, parallel calls, strict mode
    Best free optionGemini 2.5 FlashFree with 1M context, strong capabilities
    Best reasoningo3 / DeepSeek R1Reasoning models outperform standard models on math/science
    Most mature APIOpenAIWidest SDK support, fine-tuning, most integrations
    + +
    + Explore More: + Best AI Models + Free AI Models + LLM Pricing + OpenAI Alternatives + Benchmarks + Coding Models + Agent Models + Cheapest Models + Price Calculator + Interactive Catalog +
    +
    +
    +

    + Data from AI Models Catalog — 4,587 + models across 95 providers. Updated continuously. +

    +
    + + diff --git a/site/cheapest-ai-models.html b/site/cheapest-ai-models.html index 5ab40a6a..333ac392 100644 --- a/site/cheapest-ai-models.html +++ b/site/cheapest-ai-models.html @@ -1514,6 +1514,7 @@

    🔗 More Resources

    🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 Benchmarks + ChatGPT vs Claude vs Gemini — star, fork, contribute
  • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index ff91a883..e0344757 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -864,6 +864,7 @@

    🔗 More Resources

    🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 Benchmarks + ChatGPT vs Claude vs Gemini — star, fork, contribute
  • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index 5508136c..b78635f7 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2786,6 +2786,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + ChatGPT vs Claude vs Gemini — star, fork, contribute
  • From 7d5b17b674cff5d19613b131ac4203caf3f4cf97 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:42:38 +0800 Subject: [PATCH 286/311] Update index files: 20 SEO pages, add ChatGPT vs Claude vs Gemini --- README.md | 20 ++++++++++---------- llms-full.txt | 2 ++ llms.txt | 1 + 3 files changed, 13 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index 95fe1752..61ef2ad7 100644 --- a/README.md +++ b/README.md @@ -571,21 +571,21 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 19 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html), [State of AI](https://i-need-token.github.io/ai-models/state-of-ai-models.html), [Benchmarks](https://i-need-token.github.io/ai-models/ai-model-benchmarks.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 20 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html), [State of AI](https://i-need-token.github.io/ai-models/state-of-ai-models.html), [Benchmarks](https://i-need-token.github.io/ai-models/ai-model-benchmarks.html), [ChatGPT vs Claude vs Gemini](https://i-need-token.github.io/ai-models/chatgpt-vs-claude-vs-gemini.html) | ## What's New ### v0.2.0 (May 2025) -- **19 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal, State of AI Models 2025 +- **20 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal, State of AI Models 2025 - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities diff --git a/llms-full.txt b/llms-full.txt index d2a3943f..4ab194ad 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -409,5 +409,7 @@ npx tsx scripts/export-csv.ts - Multimodal AI Models: site/multimodal-ai-models.html - State of AI Models 2025: site/state-of-ai-models.html - AI Model Benchmarks Comparison: site/ai-model-benchmarks.html +- ChatGPT vs Claude vs Gemini: site/chatgpt-vs-claude-vs-gemini.html - State of AI Models 2025: site/state-of-ai-models.html - AI Model Benchmarks Comparison: site/ai-model-benchmarks.html +- ChatGPT vs Claude vs Gemini: site/chatgpt-vs-claude-vs-gemini.html diff --git a/llms.txt b/llms.txt index 5cb4f27a..06af5a50 100644 --- a/llms.txt +++ b/llms.txt @@ -409,3 +409,4 @@ npx tsx scripts/export-csv.ts - Multimodal AI Models: site/multimodal-ai-models.html - State of AI Models 2025: site/state-of-ai-models.html - AI Model Benchmarks Comparison: site/ai-model-benchmarks.html +- ChatGPT vs Claude vs Gemini: site/chatgpt-vs-claude-vs-gemini.html From 0d07da4d9e4237cfbad4f748063371ee751ec759 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:45:21 +0800 Subject: [PATCH 287/311] Add Quick Value Demo section to README hero area Three instant-value callouts answering the most common developer questions: cheapest tool-calling model, best free reasoning model, largest context window. --- README.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/README.md b/README.md index 61ef2ad7..8b53dd1c 100644 --- a/README.md +++ b/README.md @@ -31,6 +31,12 @@ Machine-readable YAML catalog of every major AI model provider and their models > curl -sL https://github.com/i-need-token/ai-models/releases/latest/download/models.json | python3 -c "import sys,json; d=json.load(sys.stdin); print(f'{len(d["models"])} models across {len(set(m["provider"] for m in d["models"]))} providers')" > ``` +## 💡 Quick Value Demo + +> **What's the cheapest model with tool calling?** → ling-2.6-flash at **$0.01/$0.03** per M tokens ([see all 2,350 →](docs/tool-calling.md)) +> **What's the best free reasoning model?** → DeepSeek R1 — **92% MATH-500** ([see all 81 free →](docs/free-models.md)) +> **Which model has the largest context window?** → Gemini 2.5 Pro — **1,048,576 tokens** ([see all context windows →](docs/context-windows.md)) + ## 🖥️ Interactive Catalog [![AI Models Catalog — Interactive model comparison tool](docs/images/catalog-screenshot.png)](https://i-need-token.github.io/ai-models/) From 93cf6d3e60f47e926ed28104a45627b38171ab3d Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:58:24 +0800 Subject: [PATCH 288/311] Add AI Model Comparison Chart SEO page Comprehensive comparison chart with 5 sections: flagship models, best value under $1/M, context windows, capability matrix, and best model by use case. Cross-links from 6 related pages. --- site/ai-model-comparison-chart.html | 757 ++++++++++++++++++++++++++ site/best-ai-models.html | 1 + site/chatgpt-vs-claude-vs-gemini.html | 1 + site/cheapest-ai-models.html | 1 + site/context-window-comparison.html | 1 + site/llm-pricing.html | 1 + site/openai-alternatives.html | 1 + 7 files changed, 763 insertions(+) create mode 100644 site/ai-model-comparison-chart.html diff --git a/site/ai-model-comparison-chart.html b/site/ai-model-comparison-chart.html new file mode 100644 index 00000000..7f4b3bec --- /dev/null +++ b/site/ai-model-comparison-chart.html @@ -0,0 +1,757 @@ + + + + + + + AI Model Comparison Chart 2025 — Pricing, Context, Capabilities | AI Models Catalog + + + + + + + + + + + + + + + +
    +

    📊 AI Model Comparison Chart 2025

    +

    + Side-by-side comparison of AI models: pricing, context windows, tool calling, reasoning, + vision, and structured output. Data from 95 providers, 4,587 models. +

    +
    +
    +

    1. Flagship Models Comparison

    +

    The top models from each major provider, compared across all key dimensions.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/MOutput $/MContextTool CallReasoningVisionStruct. Output
    GPT-4.1OpenAI$2.00$8.001,047K✅❌✅✅
    o3OpenAI$2.00$8.00200K✅✅✅✅
    o4-miniOpenAI$1.10$4.40200K✅✅✅✅
    Claude Opus 4Anthropic$15.00$75.00200K✅✅✅✅
    Claude Sonnet 4Anthropic$3.00$15.00200K✅✅✅✅
    Claude Haiku 3.5Anthropic$0.80$4.00200K✅❌✅✅
    Gemini 2.5 ProGoogle$1.25$10.001,048K✅✅✅✅
    Gemini 2.5 FlashGoogleFreeFree1,048K✅✅✅✅
    Grok 3xAI$3.00$15.00131K✅❌❌❌
    Grok 3 MinixAI$0.30$0.50131K✅✅❌❌
    DeepSeek R1DeepSeekFreeFree164K✅✅❌❌
    DeepSeek V3DeepSeek$0.07$0.27164K✅❌❌❌
    Mistral LargeMistral$2.00$6.00128K✅❌✅✅
    CodestralMistral$0.30$0.90256K❌❌❌❌
    Qwen3-235BAlibabaFreeFree128K✅✅✅✅
    Command R+Cohere$2.50$10.00128K✅❌❌✅
    Llama 4 MaverickMetaFreeFree1,048K✅❌✅❌
    Nova ProAmazon$0.80$3.20300K✅❌✅✅
    + +

    2. Best Value Models (Under $1/M Input)

    +

    Models that offer strong capabilities at budget-friendly prices.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/MOutput $/MContextTool CallReasoningVision
    Gemini 2.5 FlashGoogleFreeFree1,048K✅✅✅
    DeepSeek R1DeepSeekFreeFree164K✅✅❌
    Qwen3-235BAlibabaFreeFree128K✅✅✅
    DeepSeek V3DeepSeek$0.07$0.27164K✅❌❌
    Grok 3 MinixAI$0.30$0.50131K✅✅❌
    CodestralMistral$0.30$0.90256K❌❌❌
    Claude Haiku 3.5Anthropic$0.80$4.00200K✅❌✅
    Nova ProAmazon$0.80$3.20300K✅❌✅
    + +

    3. Context Window Comparison

    +

    Models with the largest context windows for processing long documents.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContext WindowInput $/MTool Call
    Gemini 2.5 ProGoogle1,048,576$1.25✅
    Gemini 2.5 FlashGoogle1,048,576Free✅
    GPT-4.1OpenAI1,047,576$2.00✅
    Llama 4 MaverickMeta1,048,576Free✅
    Nova ProAmazon300,000$0.80✅
    Claude Opus/Sonnet 4Anthropic200,000$3-15✅
    o3 / o4-miniOpenAI200,000$1.10-2✅
    DeepSeek R1/V3DeepSeek163,840Free✅
    + +

    4. Capability Matrix

    +

    How many models support each capability across our catalog.

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    CapabilityModelsFree ModelsCheapest Paid
    Tool Calling2,35054ling-2.6-flash ($0.01/$0.03)
    Reasoning1,30618qwen3.5-0.8b ($0.01/$0.05)
    Vision1,48735ling-2.6-flash ($0.01/$0.03)
    Structured Output82924ling-2.6-flash ($0.01/$0.03)
    Open Weights52781Free
    Image Output285Various
    Audio Input11812Various
    Audio Output348Various
    + +

    5. Best Model by Use Case

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Use CaseBest ModelWhyCost
    AI AgentsGPT-4.1#1 tool calling, parallel calls$2/$8
    CodingClaude Sonnet 4#1 SWE-bench, 64K output$3/$15
    Reasoningo3#1 MATH, GPQA$2/$8
    Long DocumentsGemini 2.5 Pro1M context, best price$1.25/$10
    ChatGPT-4.1#1 Chatbot Arena$2/$8
    BudgetGemini 2.5 FlashFree with 1M contextFree
    Open SourceQwen3-235BBest open-weight modelFree
    VisionGemini 2.5 ProBest MMMU, image+video$1.25/$10
    + +
    + Explore all 4,587 models: Use our + interactive catalog to filter, sort, + compare, and calculate costs for any combination of models. +
    + +
    + Explore More: + Best AI Models + Free AI Models + LLM Pricing + ChatGPT vs Claude vs Gemini + OpenAI Alternatives + Benchmarks + Context Windows + Cheapest Models + Price Calculator + Interactive Catalog +
    +
    +
    +

    + Data from AI Models Catalog — 4,587 + models across 95 providers. Updated continuously. +

    +
    + + diff --git a/site/best-ai-models.html b/site/best-ai-models.html index 885f3db8..7c306ac9 100644 --- a/site/best-ai-models.html +++ b/site/best-ai-models.html @@ -1062,6 +1062,7 @@

    🔗 More Resources

    State of AI Models 2025 Benchmarks ChatGPT vs Claude vs Gemini + Comparison Chart — star, fork, contribute
  • diff --git a/site/chatgpt-vs-claude-vs-gemini.html b/site/chatgpt-vs-claude-vs-gemini.html index 9d9afa54..45c044e7 100644 --- a/site/chatgpt-vs-claude-vs-gemini.html +++ b/site/chatgpt-vs-claude-vs-gemini.html @@ -615,6 +615,7 @@

    7. The Verdict

    Cheapest Models Price Calculator Interactive Catalog + Comparison Chart
  • diff --git a/site/context-window-comparison.html b/site/context-window-comparison.html index f91c8533..9c8afd26 100644 --- a/site/context-window-comparison.html +++ b/site/context-window-comparison.html @@ -1597,6 +1597,7 @@

    🔗 More Resources

    🔓 Open Source AI Models (527 models) 🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 + Comparison Chart — star, fork, contribute
  • diff --git a/site/llm-pricing.html b/site/llm-pricing.html index e0344757..7395193a 100644 --- a/site/llm-pricing.html +++ b/site/llm-pricing.html @@ -865,6 +865,7 @@

    🔗 More Resources

    State of AI Models 2025 Benchmarks ChatGPT vs Claude vs Gemini + Comparison Chart — star, fork, contribute
  • diff --git a/site/openai-alternatives.html b/site/openai-alternatives.html index b78635f7..0e5b51a3 100644 --- a/site/openai-alternatives.html +++ b/site/openai-alternatives.html @@ -2787,6 +2787,7 @@

    🔗 More Resources

    🎨 Multimodal AI Models (1,548 models) State of AI Models 2025 ChatGPT vs Claude vs Gemini + Comparison Chart — star, fork, contribute
  • From c0ceb6b4641ede46419f70cce0e0b79bec0a73b2 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 10:58:56 +0800 Subject: [PATCH 289/311] Update index files: 21 SEO pages, add Comparison Chart --- README.md | 20 ++++++++++---------- llms-full.txt | 2 ++ llms.txt | 1 + 3 files changed, 13 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index 8b53dd1c..f98c3858 100644 --- a/README.md +++ b/README.md @@ -577,21 +577,21 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. ## Ecosystem & Integrations -| Integration | Description | Link | -| ------------------------ | ------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | -| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | -| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | -| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | -| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | -| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | -| **SEO comparison pages** | 20 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html), [State of AI](https://i-need-token.github.io/ai-models/state-of-ai-models.html), [Benchmarks](https://i-need-token.github.io/ai-models/ai-model-benchmarks.html), [ChatGPT vs Claude vs Gemini](https://i-need-token.github.io/ai-models/chatgpt-vs-claude-vs-gemini.html) | +| Integration | Description | Link | +| ------------------------ | ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **npm package** | Install models.json via npm | [`npm install ai-models`](https://www.npmjs.com/package/ai-models) | +| **jsDelivr CDN** | Fetch models.json from CDN | [cdn.jsdelivr.net/npm/ai-models](https://cdn.jsdelivr.net/npm/ai-models@latest/models.json) | +| **GitHub Action** | Use in CI/CD workflows | [action.yml](action.yml) | +| **Hugging Face** | Dataset on HF Hub | [huggingface.co/datasets/i-need-token/ai-models](https://huggingface.co/datasets/i-need-token/ai-models) | +| **CSV download** | Import into Excel/Sheets | [GitHub Releases](https://github.com/i-need-token/ai-models/releases) | +| **Interactive catalog** | Search, filter, compare, price calculator, model picker | [i-need-token.github.io/ai-models](https://i-need-token.github.io/ai-models/) | +| **SEO comparison pages** | 21 curated comparison pages for discoverability | [Best Models](https://i-need-token.github.io/ai-models/best-ai-models.html), [Free Models](https://i-need-token.github.io/ai-models/free-ai-models.html), [Pricing](https://i-need-token.github.io/ai-models/llm-pricing.html), [OpenAI Alt](https://i-need-token.github.io/ai-models/openai-alternatives.html), [By Provider](https://i-need-token.github.io/ai-models/ai-models-by-provider.html), [Context](https://i-need-token.github.io/ai-models/context-window-comparison.html), [Coding](https://i-need-token.github.io/ai-models/best-ai-models-for-coding.html), [Agents](https://i-need-token.github.io/ai-models/best-ai-models-for-agents.html), [Reasoning](https://i-need-token.github.io/ai-models/reasoning-models-comparison.html), [Cheapest](https://i-need-token.github.io/ai-models/cheapest-ai-models.html), [Tool Calling](https://i-need-token.github.io/ai-models/tool-calling-models-comparison.html), [Pricing Calc](https://i-need-token.github.io/ai-models/ai-model-pricing-calculator.html), [Image Gen](https://i-need-token.github.io/ai-models/best-ai-models-for-image-generation.html), [Vision](https://i-need-token.github.io/ai-models/best-ai-models-for-vision.html), [Structured Output](https://i-need-token.github.io/ai-models/structured-output-models-comparison.html), [Open Source](https://i-need-token.github.io/ai-models/open-source-ai-models.html), [Multimodal](https://i-need-token.github.io/ai-models/multimodal-ai-models.html), [State of AI](https://i-need-token.github.io/ai-models/state-of-ai-models.html), [Benchmarks](https://i-need-token.github.io/ai-models/ai-model-benchmarks.html), [ChatGPT vs Claude vs Gemini](https://i-need-token.github.io/ai-models/chatgpt-vs-claude-vs-gemini.html), [Comparison Chart](https://i-need-token.github.io/ai-models/ai-model-comparison-chart.html) | ## What's New ### v0.2.0 (May 2025) -- **20 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal, State of AI Models 2025 +- **21 SEO comparison pages** — Best Models, Free Models, LLM Pricing, OpenAI Alternatives, By Provider, Context Windows, Coding, Agents, Reasoning, Cheapest, Tool Calling, Pricing Calculator, Image Generation, Open Source, Multimodal, State of AI Models 2025 - **Interactive catalog** — 25+ features including dark/light theme, keyboard shortcuts, model detail modal, price calculator, model picker wizard, copy as code, share button, j/k vim navigation - **95 providers** — comprehensive coverage of all major AI providers - **4,587+ models** — with pricing, context windows, modalities, and capabilities diff --git a/llms-full.txt b/llms-full.txt index 4ab194ad..6a0113de 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -410,6 +410,8 @@ npx tsx scripts/export-csv.ts - State of AI Models 2025: site/state-of-ai-models.html - AI Model Benchmarks Comparison: site/ai-model-benchmarks.html - ChatGPT vs Claude vs Gemini: site/chatgpt-vs-claude-vs-gemini.html +- AI Model Comparison Chart: site/ai-model-comparison-chart.html - State of AI Models 2025: site/state-of-ai-models.html - AI Model Benchmarks Comparison: site/ai-model-benchmarks.html - ChatGPT vs Claude vs Gemini: site/chatgpt-vs-claude-vs-gemini.html +- AI Model Comparison Chart: site/ai-model-comparison-chart.html diff --git a/llms.txt b/llms.txt index 06af5a50..4e0e67e5 100644 --- a/llms.txt +++ b/llms.txt @@ -410,3 +410,4 @@ npx tsx scripts/export-csv.ts - State of AI Models 2025: site/state-of-ai-models.html - AI Model Benchmarks Comparison: site/ai-model-benchmarks.html - ChatGPT vs Claude vs Gemini: site/chatgpt-vs-claude-vs-gemini.html +- AI Model Comparison Chart: site/ai-model-comparison-chart.html From 8c6fa05bc8ddf898c7aaa27f64172635add9aa65 Mon Sep 17 00:00:00 2001 From: "dependabot[bot]" <49699333+dependabot[bot]@users.noreply.github.com> Date: Fri, 22 May 2026 11:09:48 +0800 Subject: [PATCH 290/311] chore(deps): bump actions/checkout from 4 to 6 (#1) Bumps [actions/checkout](https://github.com/actions/checkout) from 4 to 6. - [Release notes](https://github.com/actions/checkout/releases) - [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md) - [Commits](https://github.com/actions/checkout/compare/v4...v6) --- updated-dependencies: - dependency-name: actions/checkout dependency-version: '6' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> --- .github/workflows/digest.yml | 2 +- .github/workflows/pages.yml | 2 +- .github/workflows/publish.yml | 2 +- .github/workflows/release.yml | 2 +- .github/workflows/stats.yml | 2 +- .github/workflows/sync-hf.yml | 2 +- .github/workflows/sync.yml | 2 +- .github/workflows/validate.yml | 6 +++--- 8 files changed, 10 insertions(+), 10 deletions(-) diff --git a/.github/workflows/digest.yml b/.github/workflows/digest.yml index 2d45b958..d441e1f7 100644 --- a/.github/workflows/digest.yml +++ b/.github/workflows/digest.yml @@ -13,7 +13,7 @@ jobs: digest: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v4 with: diff --git a/.github/workflows/pages.yml b/.github/workflows/pages.yml index 2e9f8dcf..156b4401 100644 --- a/.github/workflows/pages.yml +++ b/.github/workflows/pages.yml @@ -26,7 +26,7 @@ jobs: name: github-pages url: ${{ steps.deployment.outputs.page_url }} steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - name: Setup Pages uses: actions/configure-pages@v5 diff --git a/.github/workflows/publish.yml b/.github/workflows/publish.yml index b22bee48..3bf9cc3f 100644 --- a/.github/workflows/publish.yml +++ b/.github/workflows/publish.yml @@ -11,7 +11,7 @@ jobs: publish: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index f623719c..656a236d 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -13,7 +13,7 @@ jobs: release: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: diff --git a/.github/workflows/stats.yml b/.github/workflows/stats.yml index 28b65cb0..6a731338 100644 --- a/.github/workflows/stats.yml +++ b/.github/workflows/stats.yml @@ -12,7 +12,7 @@ jobs: stats: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: diff --git a/.github/workflows/sync-hf.yml b/.github/workflows/sync-hf.yml index 034650a7..c974da28 100644 --- a/.github/workflows/sync-hf.yml +++ b/.github/workflows/sync-hf.yml @@ -9,7 +9,7 @@ jobs: sync: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: diff --git a/.github/workflows/sync.yml b/.github/workflows/sync.yml index a5a0fe5b..85c2f12b 100644 --- a/.github/workflows/sync.yml +++ b/.github/workflows/sync.yml @@ -13,7 +13,7 @@ jobs: sync: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: diff --git a/.github/workflows/validate.yml b/.github/workflows/validate.yml index 900b2b97..16e649bb 100644 --- a/.github/workflows/validate.yml +++ b/.github/workflows/validate.yml @@ -20,7 +20,7 @@ jobs: validate: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: @@ -41,7 +41,7 @@ jobs: format: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - uses: actions/setup-node@v6 with: @@ -56,7 +56,7 @@ jobs: stats: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v4 + - uses: actions/checkout@v6 - name: Count providers id: count From 8156ec3847ae2688198a3c6aed656cbc5992a506 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 11:12:41 +0800 Subject: [PATCH 291/311] chore: bump dependencies (upload-artifact v7, gh-release v3, create-pull-request v8, oxfmt 0.51.0) --- .github/workflows/release.yml | 2 +- .github/workflows/stats.yml | 2 +- .github/workflows/sync.yml | 2 +- package-lock.json | 168 ++++++++++++++++++---------------- package.json | 2 +- 5 files changed, 92 insertions(+), 84 deletions(-) diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 656a236d..d597af4f 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -35,7 +35,7 @@ jobs: run: npx tsx scripts/stats.ts json > stats.json - name: Create release - uses: softprops/action-gh-release@v2 + uses: softprops/action-gh-release@v3 with: files: | dist/models.json diff --git a/.github/workflows/stats.yml b/.github/workflows/stats.yml index 6a731338..d37890ea 100644 --- a/.github/workflows/stats.yml +++ b/.github/workflows/stats.yml @@ -51,7 +51,7 @@ jobs: git push - name: Upload artifacts - uses: actions/upload-artifact@v4 + uses: actions/upload-artifact@v7 with: name: catalog-data path: | diff --git a/.github/workflows/sync.yml b/.github/workflows/sync.yml index 85c2f12b..5977a609 100644 --- a/.github/workflows/sync.yml +++ b/.github/workflows/sync.yml @@ -38,7 +38,7 @@ jobs: - name: Create pull request if: steps.changes.outputs.changed == 'true' - uses: peter-evans/create-pull-request@v6 + uses: peter-evans/create-pull-request@v8 with: title: "🔄 Weekly model data sync" body: | diff --git a/package-lock.json b/package-lock.json index b104fe46..2d556cd9 100644 --- a/package-lock.json +++ b/package-lock.json @@ -16,15 +16,15 @@ "devDependencies": { "@types/node": "^25.9.1", "husky": "^9.1.7", - "oxfmt": "^0.48.0", + "oxfmt": "^0.51.0", "oxlint": "^1.66.0", "typescript": "^5.7.0" } }, "node_modules/@oxfmt/binding-android-arm-eabi": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-android-arm-eabi/-/binding-android-arm-eabi-0.48.0.tgz", - "integrity": "sha512-uwqk+/KhQvBIpULD8SMM/zAafMRC/+DV/xsEQjkkIsJ/kLmEI/2bxonVowcYTiXqqZ/a0FEW8DPkZY3VvwELDA==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-android-arm-eabi/-/binding-android-arm-eabi-0.51.0.tgz", + "integrity": "sha512-Ni0sCqg5CIHaLIYFGj+ncbcumylvNC6FE4rfD0KfdmnWHbPJ+zev0qZCXKxy2hFVa0fYRK0yPzf5nzPbkZou7g==", "cpu": [ "arm" ], @@ -39,9 +39,9 @@ } }, "node_modules/@oxfmt/binding-android-arm64": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-android-arm64/-/binding-android-arm64-0.48.0.tgz", - "integrity": "sha512-VUCiKuXK5+McVssgHEJdrcGK7hRJzrRb36zm9/jwzMholyYt4BgXhw5Nm1V1DX6Ce717Zi/1jk432b/tgmQgtQ==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-android-arm64/-/binding-android-arm64-0.51.0.tgz", + "integrity": "sha512-eu5lAZjuo0KAkp+M24EhDqfOwA8owQ8d7wyBlOUUGRbDLHpU3IRlDHp8Dif+YqGlxs6jra7yS6WQu/NkPhAxeg==", "cpu": [ "arm64" ], @@ -56,9 +56,9 @@ } }, "node_modules/@oxfmt/binding-darwin-arm64": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-darwin-arm64/-/binding-darwin-arm64-0.48.0.tgz", - "integrity": "sha512-IkKp8rnIyQLW6Jt+6jragCbUVYSayk55lapiprLjIVvt4NczLyO/nwX2GgefLQ5iaBdfS8UEAFgCs/pLO6Cl0w==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-darwin-arm64/-/binding-darwin-arm64-0.51.0.tgz", + "integrity": "sha512-6LsUNIdURhhcIfIn8+xsOb61mSTa9msAHTeSGx9Jf4rsP/gN8PGCF+SKWPAQZbND2w/WBkqQ6303jqEEIXzMdQ==", "cpu": [ "arm64" ], @@ -73,9 +73,9 @@ } }, "node_modules/@oxfmt/binding-darwin-x64": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-darwin-x64/-/binding-darwin-x64-0.48.0.tgz", - "integrity": "sha512-+aFuhsGIuvnoOjXyKVHMhPKJZR1kQkAl8QyrKoMlA7yJsSTC3N0Asl53La8TChSHhW8epToQ/Q0nvLmEmfNmLg==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-darwin-x64/-/binding-darwin-x64-0.51.0.tgz", + "integrity": "sha512-9aUMGmVxdHjYMsEAW1tNRoieTJXlVNDFkRvIR1J7LttJXWjVYCu2ekclLij2KJtxBxSQOYSHd12ME/adVGVbZg==", "cpu": [ "x64" ], @@ -90,9 +90,9 @@ } }, "node_modules/@oxfmt/binding-freebsd-x64": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-freebsd-x64/-/binding-freebsd-x64-0.48.0.tgz", - "integrity": "sha512-fbqzQL8FjI9gGnktI7RIo0dksDziTAYBy7xlI7jU7eID5fxLF/25fS4Xj6GydD8Y5oWHL83U4NK160QaOAxtyg==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-freebsd-x64/-/binding-freebsd-x64-0.51.0.tgz", + "integrity": "sha512-mkY1nhZTqYb+NHaAWxOCKISN6FwdrwMNsu17vTUA3wzUV2VJ+Paq15ZokRcsMU/2PUdHO73prxyeJpjXQ3MPpQ==", "cpu": [ "x64" ], @@ -107,9 +107,9 @@ } }, "node_modules/@oxfmt/binding-linux-arm-gnueabihf": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm-gnueabihf/-/binding-linux-arm-gnueabihf-0.48.0.tgz", - "integrity": "sha512-hn4i0zhAyTiB3ZHjQfYUZkDvrbVkohw1S7pySWxWUoZ87HnkDoTFThj7QTxk40hNPOTUP0vHbPRNamFIv1HBJQ==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm-gnueabihf/-/binding-linux-arm-gnueabihf-0.51.0.tgz", + "integrity": "sha512-wtFwNwE4+YCNuPaWoGDZeGsKvD6D1YSUNBJNn/rJBh7CrDBThFE+TBI5kY7vRW9rIOQRsbW2IpyyL3Du4Zqwiw==", "cpu": [ "arm" ], @@ -124,9 +124,9 @@ } }, "node_modules/@oxfmt/binding-linux-arm-musleabihf": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm-musleabihf/-/binding-linux-arm-musleabihf-0.48.0.tgz", - "integrity": "sha512-R4WBD9qF3QM9hqgdAa+fBGXmquTvDUujrPQ36t2Sjk8RPOSKGHDeN7l/khr10hqbQaOq9KCgPHG9ubNET/X/RQ==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm-musleabihf/-/binding-linux-arm-musleabihf-0.51.0.tgz", + "integrity": "sha512-rnOaNx86G7iRKM6lsCIQMux0SMGNC/TEbFR+r7lpruJ12bnrIWgxd5w1PLqOvgR9r8ZJbpK/zfRKctJnh8/Jfg==", "cpu": [ "arm" ], @@ -141,9 +141,9 @@ } }, "node_modules/@oxfmt/binding-linux-arm64-gnu": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm64-gnu/-/binding-linux-arm64-gnu-0.48.0.tgz", - "integrity": "sha512-5bVdwSwlm1M8wbYCorLOxWxUBw/8tBvHYyQNIfwWVPwOJaj5vg1APSGJQVpwJfV5VNE9PSrR91UKEpoNwHhqUA==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm64-gnu/-/binding-linux-arm64-gnu-0.51.0.tgz", + "integrity": "sha512-jOgDzSqWcICGRjsp4mc08FxKMN8vzP2Kgs4E0d2HUP99F+nJDQKklRV4Zuj+0gcBgjrzx2CbpqaIdUVPepCojA==", "cpu": [ "arm64" ], @@ -158,9 +158,9 @@ } }, "node_modules/@oxfmt/binding-linux-arm64-musl": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm64-musl/-/binding-linux-arm64-musl-0.48.0.tgz", - "integrity": "sha512-vCS3Fk7gFslTqE1lUE2IlroyVV7u/9SmMA/uBqDoshuck2psGWcjW0ePyPZI3rM3+qtf2pDaMVIKMHozraifuw==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-arm64-musl/-/binding-linux-arm64-musl-0.51.0.tgz", + "integrity": "sha512-KBUCdrH5bwVrAvI9gU/1S55oH6fzXjr++J/oVocdu7bYTks1l7DNNT+rLd/1TDdAEjObGwmfWamn7LC1m8A0DQ==", "cpu": [ "arm64" ], @@ -175,9 +175,9 @@ } }, "node_modules/@oxfmt/binding-linux-ppc64-gnu": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-ppc64-gnu/-/binding-linux-ppc64-gnu-0.48.0.tgz", - "integrity": "sha512-gKtfFfueUClXDumyoHUbymqRf7prHejOOyzJK0eIJn93GF9JBdFHdo60TM1ZBHxkEwZvjuOgHmKtneKbEOc/Eg==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-ppc64-gnu/-/binding-linux-ppc64-gnu-0.51.0.tgz", + "integrity": "sha512-NapfjYsABFqTJ1Dn9Efq6sN5esaHconVKwVLbDGNQLrwpOx/g17mkwErHzU72PutL67nf3wNAkbq122H+zLxag==", "cpu": [ "ppc64" ], @@ -192,9 +192,9 @@ } }, "node_modules/@oxfmt/binding-linux-riscv64-gnu": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-riscv64-gnu/-/binding-linux-riscv64-gnu-0.48.0.tgz", - "integrity": "sha512-SYt0UhOvZD/UwZz9sXq6J2uAw8o24f5VZpLB2DH01f6MevshmlgakQlZe2lwek2sZJkd07eLu7mZa0g7yeiw7Q==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-riscv64-gnu/-/binding-linux-riscv64-gnu-0.51.0.tgz", + "integrity": "sha512-5dlDt1dUZCVi6elIhiK1PWg9wpTzTcIuj0IZnSurvIoMrhOWqqTcc1dSTxcSkNaBZhfsNqRZdINI1zAgbKkJNQ==", "cpu": [ "riscv64" ], @@ -209,9 +209,9 @@ } }, "node_modules/@oxfmt/binding-linux-riscv64-musl": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-riscv64-musl/-/binding-linux-riscv64-musl-0.48.0.tgz", - "integrity": "sha512-JLbrwck2AopG4ud/XklZO5N+qxGC7cS7ROvXZVNfx0MCLDDL2kGOLvzuWORkVjnjAM0CMAfIMU2zNBtQbM+4dw==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-riscv64-musl/-/binding-linux-riscv64-musl-0.51.0.tgz", + "integrity": "sha512-pgdWUJn0S5nulyiVdlFV8DzCUnGXkU99W5PSkkmbaZW+LrZBPxpezun4G0DDHbQaVYuJeCuKsXsGKGo77CkUTQ==", "cpu": [ "riscv64" ], @@ -226,9 +226,9 @@ } }, "node_modules/@oxfmt/binding-linux-s390x-gnu": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-s390x-gnu/-/binding-linux-s390x-gnu-0.48.0.tgz", - "integrity": "sha512-mdxt5L8OQLxkQH+JVpdC/lknZNe0lX4hlO3d8+xvw2wToo+iDrid9tiGOd5bmHfUVd5wVhrUry0qlu5vq66NkQ==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-s390x-gnu/-/binding-linux-s390x-gnu-0.51.0.tgz", + "integrity": "sha512-2XTFUe97CbDGAI8vjwDfZ1HdakO0XIADyJ24idEg64SC4/K4in/OisXVnrW4NMK7I6TgC7EqRhC0Ln/nKhAemA==", "cpu": [ "s390x" ], @@ -243,9 +243,9 @@ } }, "node_modules/@oxfmt/binding-linux-x64-gnu": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-x64-gnu/-/binding-linux-x64-gnu-0.48.0.tgz", - "integrity": "sha512-oEz1BQwMrV7OMEFx/3VPDU3n9TM0AnxpktDYXjEg5i6nTX87wo18wSfBvkl4tzAICdKtoAQAdBIl7Y7hsPlx5w==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-x64-gnu/-/binding-linux-x64-gnu-0.51.0.tgz", + "integrity": "sha512-kQ1OuCqqt/yyf0ZN9VFxW1/JnlgJgii3Dr7pWf9vNBvrX1hv6g39/+mc5oGRHRGJFZtl3zsGDWR9c5N2B/gwBw==", "cpu": [ "x64" ], @@ -260,9 +260,9 @@ } }, "node_modules/@oxfmt/binding-linux-x64-musl": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-x64-musl/-/binding-linux-x64-musl-0.48.0.tgz", - "integrity": "sha512-g2SKTTurP5mWjd8Ecait0erYqmltL4IqW1EwttM25BxM6NiTt4ubobJYMR1uox1V2QgG4UfHH10CGRvWlUixjw==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-linux-x64-musl/-/binding-linux-x64-musl-0.51.0.tgz", + "integrity": "sha512-ARTYqxHF475o96Gbn41hvSWSSRygPlRDXZZgZ9I2scU1y0qiWpCQyZCoefaQa0mwv+wwtZ+luS4YOzsRzM/izg==", "cpu": [ "x64" ], @@ -277,9 +277,9 @@ } }, "node_modules/@oxfmt/binding-openharmony-arm64": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-openharmony-arm64/-/binding-openharmony-arm64-0.48.0.tgz", - "integrity": "sha512-CIg24VgheEpvolHL2gQuax5qcQ602bRMHrJ9g8XsQr3iVj9aSPgopigBKuMqrXsupwkrU+RQCn5cG8PgFntR6w==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-openharmony-arm64/-/binding-openharmony-arm64-0.51.0.tgz", + "integrity": "sha512-QiC1XrCl6a6BmqMzduO8hdIRMf1m44hCkt2Q68KWkTvUB/E7fd2iomyNh6KnnRca5w6eBrRAAtLFqTh+xjsjJA==", "cpu": [ "arm64" ], @@ -294,9 +294,9 @@ } }, "node_modules/@oxfmt/binding-win32-arm64-msvc": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-win32-arm64-msvc/-/binding-win32-arm64-msvc-0.48.0.tgz", - "integrity": "sha512-zeaWkcxcEULwkGF3I/HgEvcDPN8buYDrxibBUa/IFh5Vmwyge+KpLO+hEwSovW349H0O/C0Z2kaFmEzEDm00/Q==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-win32-arm64-msvc/-/binding-win32-arm64-msvc-0.51.0.tgz", + "integrity": "sha512-NC/hJb9dtU23Zf8L7IVK95xnFjiQ7AfcLO2l5pb69TDEr958qxrtnB2CveeeNSCBFNIkgaTCfd/vHNSoG78l9g==", "cpu": [ "arm64" ], @@ -311,9 +311,9 @@ } }, "node_modules/@oxfmt/binding-win32-ia32-msvc": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-win32-ia32-msvc/-/binding-win32-ia32-msvc-0.48.0.tgz", - "integrity": "sha512-yiEKnIAGvx5CyZQOlMaNlZkAbwT7/Quk0j3WLt+PR5hK+qYjPTRRJYDfD77wCBPLvEYAG41v4KG3iL0H+uxoxg==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-win32-ia32-msvc/-/binding-win32-ia32-msvc-0.51.0.tgz", + "integrity": "sha512-2C45za4Rj36n8YIbhRL1PQbxmXJYf81WEcAgvj5I4ptRROG+A+81hREEN5bmCHADE1UfYaN312U6tkILoZZy6w==", "cpu": [ "ia32" ], @@ -328,9 +328,9 @@ } }, "node_modules/@oxfmt/binding-win32-x64-msvc": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/@oxfmt/binding-win32-x64-msvc/-/binding-win32-x64-msvc-0.48.0.tgz", - "integrity": "sha512-GSD2+7t2UoVMV2NgxXypa4bKewflPMAjYnF0Xw9/ht82ZfafAHhb8STwrEd7wlH2PFogt5zw3WVCxYJaHUdbeQ==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/@oxfmt/binding-win32-x64-msvc/-/binding-win32-x64-msvc-0.51.0.tgz", + "integrity": "sha512-73RqdAuVKQTkjZIDw08JaDHUM4lav5Qu+CaPwg4QbbA7k8o7LEW0p3UsfZ/F8dsO/pwVYh3RzFcanwLRTTahbQ==", "cpu": [ "x64" ], @@ -874,9 +874,9 @@ } }, "node_modules/oxfmt": { - "version": "0.48.0", - "resolved": "https://registry.npmjs.org/oxfmt/-/oxfmt-0.48.0.tgz", - "integrity": "sha512-AVaLh+7XeGx+R1zfFV+f6VV61nT2MWVJXVUDhbTm5LBWGyNt64xAyh3NYYyjeY2WykNt9AvqSQLPHcbWquYF9g==", + "version": "0.51.0", + "resolved": "https://registry.npmjs.org/oxfmt/-/oxfmt-0.51.0.tgz", + "integrity": "sha512-l/AoAnaEOV7Q5/Z9kHOMDehVJnCgYN7wRoooWCTUMBMi16BJhLZqd9cmCnwcVFfVlzkt53zK2KLPFNp8vSsoDg==", "dev": true, "license": "MIT", "dependencies": { @@ -892,25 +892,33 @@ "url": "https://github.com/sponsors/Boshen" }, "optionalDependencies": { - "@oxfmt/binding-android-arm-eabi": "0.48.0", - "@oxfmt/binding-android-arm64": "0.48.0", - "@oxfmt/binding-darwin-arm64": "0.48.0", - "@oxfmt/binding-darwin-x64": "0.48.0", - "@oxfmt/binding-freebsd-x64": "0.48.0", - "@oxfmt/binding-linux-arm-gnueabihf": "0.48.0", - "@oxfmt/binding-linux-arm-musleabihf": "0.48.0", - "@oxfmt/binding-linux-arm64-gnu": "0.48.0", - "@oxfmt/binding-linux-arm64-musl": "0.48.0", - "@oxfmt/binding-linux-ppc64-gnu": "0.48.0", - "@oxfmt/binding-linux-riscv64-gnu": "0.48.0", - "@oxfmt/binding-linux-riscv64-musl": "0.48.0", - "@oxfmt/binding-linux-s390x-gnu": "0.48.0", - "@oxfmt/binding-linux-x64-gnu": "0.48.0", - "@oxfmt/binding-linux-x64-musl": "0.48.0", - "@oxfmt/binding-openharmony-arm64": "0.48.0", - "@oxfmt/binding-win32-arm64-msvc": "0.48.0", - "@oxfmt/binding-win32-ia32-msvc": "0.48.0", - "@oxfmt/binding-win32-x64-msvc": "0.48.0" + "@oxfmt/binding-android-arm-eabi": "0.51.0", + "@oxfmt/binding-android-arm64": "0.51.0", + "@oxfmt/binding-darwin-arm64": "0.51.0", + "@oxfmt/binding-darwin-x64": "0.51.0", + "@oxfmt/binding-freebsd-x64": "0.51.0", + "@oxfmt/binding-linux-arm-gnueabihf": "0.51.0", + "@oxfmt/binding-linux-arm-musleabihf": "0.51.0", + "@oxfmt/binding-linux-arm64-gnu": "0.51.0", + "@oxfmt/binding-linux-arm64-musl": "0.51.0", + "@oxfmt/binding-linux-ppc64-gnu": "0.51.0", + "@oxfmt/binding-linux-riscv64-gnu": "0.51.0", + "@oxfmt/binding-linux-riscv64-musl": "0.51.0", + "@oxfmt/binding-linux-s390x-gnu": "0.51.0", + "@oxfmt/binding-linux-x64-gnu": "0.51.0", + "@oxfmt/binding-linux-x64-musl": "0.51.0", + "@oxfmt/binding-openharmony-arm64": "0.51.0", + "@oxfmt/binding-win32-arm64-msvc": "0.51.0", + "@oxfmt/binding-win32-ia32-msvc": "0.51.0", + "@oxfmt/binding-win32-x64-msvc": "0.51.0" + }, + "peerDependencies": { + "svelte": "^5.0.0" + }, + "peerDependenciesMeta": { + "svelte": { + "optional": true + } } }, "node_modules/oxlint": { diff --git a/package.json b/package.json index 611709a4..e7737f43 100644 --- a/package.json +++ b/package.json @@ -109,7 +109,7 @@ "devDependencies": { "@types/node": "^25.9.1", "husky": "^9.1.7", - "oxfmt": "^0.48.0", + "oxfmt": "^0.51.0", "oxlint": "^1.66.0", "typescript": "^5.7.0" } From f549e1deb6cf64d5a39234e0489704dbd2a043b8 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 11:51:44 +0800 Subject: [PATCH 292/311] Add interactive catalog demo GIF and updated screenshots --- README.md | 7 +++++++ docs/images/calculator-screenshot.png | Bin 60967 -> 25594 bytes docs/images/catalog-dark.png | Bin 16811 -> 25706 bytes docs/images/catalog-demo.gif | Bin 0 -> 27756 bytes docs/images/catalog-screenshot.png | Bin 17791 -> 25594 bytes 5 files changed, 7 insertions(+) create mode 100644 docs/images/catalog-demo.gif diff --git a/README.md b/README.md index f98c3858..cc2aced0 100644 --- a/README.md +++ b/README.md @@ -41,6 +41,13 @@ Machine-readable YAML catalog of every major AI model provider and their models [![AI Models Catalog — Interactive model comparison tool](docs/images/catalog-screenshot.png)](https://i-need-token.github.io/ai-models/) +
    +🎬 Watch demo (filter, sort, dark mode, calculator) + +![Demo: filter by free models, sort by price, toggle dark mode, use price calculator](docs/images/catalog-demo.gif) + +
    + **[Try it live →](https://i-need-token.github.io/ai-models/)** — Search, filter, compare 4,587+ models with 25+ features including dark/light theme, keyboard shortcuts, price calculator, and model picker wizard. ## Why This Catalog? diff --git a/docs/images/calculator-screenshot.png b/docs/images/calculator-screenshot.png index 5eb8058d968acad958f8aea7e09532a396914d67..de1afc20c158d943c6276e4ca26612675d08128c 100644 GIT binary patch literal 25594 zcmcG#XH-*Z-#+S$uVSI-D1stn98^TQ2%%>NRE8P>X;Lx~5UBwvp(S}71XL6hq&Jb? zd#Fj22uMeIP3SF@00{}HXZxK0yVg11&X==4=NGrITJqmaz4xf7XA?k}hd0lShDu8TWn z8HFX5n>)_8>zjzR^5ipeU ze}sYM_po0DfaRR-i{rrZD*MYXz;X{d4}AWAxeS-^KfsA0KH};TAwE7%_MxG*NZKK- zzM(-KtagK)`QM+bLKhSixOBZ6Tzy_qQGvtZx=9u`Ha6Byl2>}B)ymi`Rs!mlm9zbS zpS;TR!cFtdogLvI>6G3Z;btLJs{BS^*d%qKLP4QhZ~5xaho{~adv2hf;)Os!4_{O7 zEo`(W-ekZ3zi;-up&=+sQnFj`Vz#1=N3Dpaz|TL}){~T!vWq+L^|lIW!69#t_|*RQ z`*S$kA*riXWW=|eZ0npJF+7u<%)aDp3|w+7xL@vu-^V0Z-hx?@4DNt>-AF_tfLs7V zY;W}5{CPs_e9~_kmL>V0QTB(}nD)T?z14x1KVSU0I=Ml;raYMdu7Ze6H3OIVLOM}a zwoLx)g{X*Jd8?qV3j!By8ZRbrvM8lJx05d#Yg=9m&rcxjWIF~k6o(zAlyht0mh|-v zR8~dCCDPo?k9xX}kg78|$2uWDXvJ!32fMBNw||RIe;rD)e)?OC|M+GWn76na4>&Sw_fy;Fk%?eecZP*ZFw z^ZzN^-{9TZ?w)@jImI<1EguNxhWYXT3H*uByE+pm8ir1@tSAx!; zT6x0xXmlV9ah?5k+Nkv5-)D5+b=URE*_Kxhi0(@^#~mo?ea_E)ulRB$=`rs4ii?OY z_UY6ceYmaI+y@ZLqCYjQV?!;Tv$=q1bC11jeA_dg*9*H?1ZrjRE7}-_4v_DeCVEoF zfBA=P=MlozDZ5jv0R?voy7=7w658sGP)5;!uQTiKeM=4MKBlS|BylsDjOxEY=6|mw z`wtZGZRDlpcvNl|fsxW3tNNUOLAYh#%38WekaTBznGwy8_NGz69lZKY%*NmNqYfq4&VJErk^l}@g&X{ts^L36XmNbumV_D)WctM(1g zJG|QbUF(|Pna!nSsjM;$607+2{@p#}=fKVuLDHe2zZ%|F^XRN+a61Iz^$t@vTUdb! zw`CRV3I4=6@GOq`;YKdMPd9Md-(6Ds`Es!%OdWQsz5zQM&X^uwhpMB(}{^ST{mJ!OUc-~(sf1a}Fo&?qd)~106QBi_QTlMsOIb0Ti(^*-*HHJ-#CUR{ z@!2Uy52O1`8#n8@%WIi)r18TQXav5Yz^y1ALjQ&o;@)wx)H<9G7}sM_mhYcB?b73q zm|6WBJzcU3buM@y={G#oa(>g2@tsXgu^<#W! zJUhbou^yy3pH_K&3YRGE4Rv!{5k8y~zj(!vTMh4)gK8yeE)Nl(99OopFjX$a> z<39Cw38nWLT6 z2D8{*%pw~!MQYaXx*GVv=)ReMiwk$r5YLq{5IL}CPa<-Trn+L++KY=p4v>vcUq?R0 z&GKJPt?`|J*cb^NZCn1@MZ6HyaPxS0gsH~jB6TrVH68-2R*ZU|rX`5EBFS<}dEt|s zo;BEGXI^9y`;<2V;dLn_bH?wuCBbXnW`}!n^CL$JR&YKN5@K_$tQR@aDO!QpA-$)W zydoD7O|Z^=W^Iw%Tq1MwwCBy~JWshE6B855q?L`H zp8XnW&jF34AS3$6_wp=o*~B1znbRC=6Giwj^LVCJyR}XiH&wZVck@bb|MuRn$2ENL zwk$5MHtgV8aXu|NE?K)4C-yO!`y@sQE+Er7yM6G_4(7`XL&<=x5$MjkmI%9_gqePu z#+VIq-=D5?Jt|C&JAb4iO$rERzhkKP!8m8MP+NH6t8Q*?c@)nPqg?G2HfcJJGTmBe zldm^2%5_xd%2;_NB=8hAn;myddS@^{NNQDjC^JLk%=AF7VWHI#XT7IXt)Kv>J%T{Lwb%D~1VvN=jp(cZ4DvV{QfX8KRwZT%|9I7su-^w?(HJA`VnULfK+6S3@`E zQ{2>~_5VE0*YCFvUbF5FHq}1D<;rvRl8T0b>~1TCws>_GiP_AIqt8rQd++-rBr@@g zgA{MnZ;1ts7_Xk!yYdy~Caep?KRUkD@T2(3OK%}F`>&&6%A5a5a06U7yZUu00erQ*p_vb~;G4~)2pEk0rVAa8>0fEY70!98^J zHP@=E&!=_n_YB7T;ui9L01HY=!hrj7wzV6|7b$L{Ftgw(TAFH#5)2!8*<5q`Eiq2x zPLeT~NZ_!=AZ&<&LVt=Qx?}6OuwbKnsP7LE0%Inbjg12R9Ycl8aU=V={-9Kh%YlO9HJ63sPzcG?Wuzd0digV_q;8{zg)B!AN+ z>F(QN^eU5B*!KC?F1@*&>xsrVl|+8LKFrYS;JAcCm6tcW%!m$4*W?{f^sGeeghZeM z0+5DGv76|r**(0h8Z29dhqyV$ePYGtO3Y|UWawpg<7)EXPc6VtED6uF9Sm;@pl!ad zT@jmg;ymv;z5Dm!Xs7$`E%Ud!%yQ-TQf;W=AGQ}`7Pl#vd{Dd}%A=si;FgRD!R%{l zZzqYLS~Q$2EdppTXNVp6-$jF{7bs?9hsywdJ})vzdS>KD%i`;=Ygs_ZJy~^B8}&#l z;ERMF_Q-i|8{SWgy z_T1EFi|eLr30Kj)Ukjo}-*W9cZRBC_?ELW;dD@(yoR9B}9$d}2f(-NQIV#ymaqR-l zIayBcw^I`GhbdMio%n1ygQQc$uElq030)?}Y~;sVvN+AGP<7Lf$xblKX~^5b!|b`a zz3%H{QY{a(aQTgFoURkbxy?!soY#5)3uq24Sb+?}&Hzi;`US7IO`1};ulGqmh|@WInTSSl2MJty)#!wX z_Fg_U?*;<37Vb&eMEmt4v%~iH&MM+vmy0bHI!zP_&sTaxZ82;KuZhajBi&!l-x3m% zz9J!k@uOR3Wo02SjPMvC30~eVj2okk5Q*0WGIvK&s!q_fNY&9m`_W)_$bQMStn1bhVx0^U|9lK3<@M?~x`W zs+&>{7Jr{H=`xk#GCvzU!1;blJzv)n{y^t&Xum{A)N5Q4&0H8suKXL>u#(^djyorO zGQRZIPE`$O$Lo#DM9b*_ppX?2PDPs7j+A1@21nx{zrV%FfEd_AWd`sb?Se$|Xl z%4YF#^D7a>4!omR__$Tq?|PC#Y@!c!>VL5S$qT=Jy%f`T{kWKsta+gC^jN?NC}cab z<298?;xGha{^vqrJZEG#K0UxVwKyl@VaVqq46BB^OM~jz&$>LJM?x?(?F_Ne;^RTtjzZ_W!w{1b*#0D~H6KV?Re<*8`x{P{O_raSHN>0ns*Qz!Jj>Zbz z#hjNCuIiT5al>jB=SB^@mg}WMU1@yTA)dV121bRgo2vjFj_*;1l5R_(0==>?C};#) zF0QWPEp2A$#BF^}6Pb;Un>(}6tGq~=rJ9bY61cPxjP_Hf zmtH2CU)Bn+w&sh=1jf8q(4)in>eTi-01m-kVBlLlJKj)t=M6;2n1R_=Ujm)Luj30I zb6Z&vAz=`wuEMeT{KfJq5Pv>h*}~HDPFUZhS0!t8aqW^q$YC0tG7(gd(3|;ERmq-8 z=opOBsO4uK?}DaMXFo@ElkTFPqYMl}D~A$qmt3)~JFu$XF=5~Jd)auff1Pdz^IeHH z%ZrJ>GV&v$zd}R@9NvQfp~C{9?w-UJf8^|k+ZVI7R_ax7k`uEL$E9Qz2I$?GnPtpeWG zIgMyD?KTYueAOCW$XAqENDT~jn?B@Pv0)ed68imgl9*S=j$nLPZp z>-&_4-&H*r_%XVA9lMh~b zDrG&TtO9^r!7y9oE(XoT7nEi2S0%k3>|#vm#hN9iMwsczS&fqy&lqVreW%w~HX)~o zjl2n@4VP8f-~f44xcoJpwZvi#^dDxYZR^T$ObU3pwX05P2|im<7M5WymRF>ja^Hij zZ;dVoES-&l?3{pE29NBFzx${2ZJMOyHUNCJqEphRryIP$BI)Cm_L67r-{#~Gf(mM% z9Mu|v&{SI8facOxz(lz-J#W{^NzKoENRGtm3i$Qy)hU8N*VLMmACuyoV~!xn{GsNz zrH>d~%~~{kw*#42F1416h3OrPCN2_I#_j zHUep=)h+g1iRht={0AlwcY-qEzKbu*^9^xxpJ1L>6VYwl5;=S6UHR<{9RFia4CuBK zzMOxxRYdo$$4>6Sni8K@~V=fD5T2NxA67)sX{ic}T?ik=}$(AotP>m*uUc9KMl> zIX6BgvEt>KhmfQBbB5R^oIfusGRd7E%kwxq0TNOw@&5_}P#&|HP4|TD?}zm%NwnQ> z7F!yvpUeF3GsT>VA1{zJy?eeESQAZddzC#fo$hF51U1}MY?G1Ctcl6gX18N*RY}@g z+KVN88(KnQOW@(R!D9U6G{45}> z$b35IMp>)gVYZp!)rM;CEx1vko>7voo}B)*oQtdkeJgTj(QfMoJ|2S_TZ(<$$j~SW zbg@4|5()ZnJJ_HbPzy|r?H#4F5*xDv<~2U@d)yiWIuai_)Uu!UWi~uN+ILD*rE|J3 zXlZLlgT002grpxF(&?hkA;NJK3~vPHgqhpoiZoFdO_W+_`yg}`Ms6#GKy4PHOb89A zK3V=gf#a={xvUfsJwNid*NLRPYz3vosV09H=5Db?o^9niCUoJl0sFhyez%ECmhmp_ zmm=uZP*%5eZ?C0p#oR{3MEZD5Zp-y5akiFUAnma7G4_v%af=_*79V7@G+VS_w;nyR z2@VXd!1!*=Njfc{YImC5wnrE_nrLQzL`xd#IP2)FZ($iI@f{RrO`)WoI15*11&A#8aDbX8}ntDTH~N;!%#)6CaU z34G`@+U#$y;&(CLx^ z#S}5z7_^-7_JKL$&{y zQY{$;;6^OksTuVJCwBSV2Pu&j91S6HFxTM-g{7rxstghw^bqE zvLEPH23HP&Lec8h-F!<2DWM-@dmH-bC->=OS7Ryjg2(!2x(}V-w`p9Ef#MIa4oEEr z;Baz66$_eWTuD(x@2tDTV^&LK^rtme>Q&!#RS0TPKSd7eF*s%TqlOcMdLz@*=|HeZ zedgtbV-EKcFm&sMzeu!kg*CXkT3lJFyjc5RT2{kNc$Jji3JLJWiv?*Gxl`1-B+1KIHSJAl zPlRGh&uVKJOT9iQ&U|=bXGXhYy;ATkQ3uLpf)aFV*P{o;jvL;RO^gpWQ7Npivajxc zGF$`Sdo|S+^?G=1&{d`QGde)&mejqG8Is+Wqyy-W523t>KQcf8OT&vUH@WWO&0*_} z?WYJDPY{py|EX1Ycd*Oi8a|Dgfb126$KPhDg&4=3Gmhi0D}W@{J{bp4AR!*Eurq|{LB z4KAKt{pJAnCg$Vlgs@k4dIn2g1{YeUUC-Je4`=OCAn$c%tn8nGX5M@6X678tmJeWe zZ62Y@>p9a9(&J*I0xkWmzghe*ph?1~hdKEziTa^>;MV%BwE{Q7)l{X#Ua71$4S=;# zQ)_5_E*#q>gZy&+aGgWe`Wxno`I=$dBHGCt5P=fj195FTQqqkvA z=4`f-7%~{%1ZIL?eUO9N`eArmPOdT4ol*IUt$!eV6~%iU1Y`dwrT5QF9=YwWw;*Ba z3l%Bvc6@ea0~?brK28S5wXi<~b<1_*=OU$(NO#L~y(g<1a8tFbtXk|0{5U!h+Kqq!3kB&aokEN_TlBVWt!Sb@Qaxb&%$8MMqZCG%F z;H%HN*RGx=W_xbW9=BydNu|Ey06JWS^r|gEP zT0s&W<)qLtW3^$r@~XRIqcaM#&WGFTI+;q)6GemxRgsYC>XH~=|7Q=DINCBGcCzby zb&3p{I??Uha~LR(KayL3@gt$yoG&K)3UO)ulFK#ZnvM&Wf(^7dgh+PS0UZ;(-Ou-X zhGn1F-M3%k1^KVN`e*DoL@2g^cf_@^AtAS2$Xq}Fu)emsU=_c&7<5r;Wv6H(ulGmC zaZJ0_)?L55N}<)_8k;w-YS0d;EAw||seRi{?=Z6?-kkGdN}PVaH}8+6E2w#opXmQ* z$yV6yq;17|gsHp3fl+A@81z8XZN*4SFo;!~=Dn*Tkj&%pe8^y{{f7Yj0<^COM@DH0 z)!W$mD#olg`Lx^0#7p?Y)N5IxcR7u%6og4H8YrH^wNwdJj78J#r4ky@F!$~^mvoGL z*r_QQDMSKCP|$YZKyvabpT&fvd`y1IW+CCsPFr48e*X0&V>;cZ+0dYA#S6wFE4$xJ z+}I#Rn7<$Ti??=x@9b=G9!qo2A@m}FFju?J6RM9~n`>eZ%n%7*V?2F5S$#5m)mb^6 zr+O2Z)vo}K6VD$p6mVxU;N1)!q}@%v2a}}K3E`NsH-)xQQ{L90-ODxB zT~}>__L`SOH2$%JIlz#;>88Ox*!L_BG$E-hBhfYibwy)8s-1Hf7ib?^TGm)uCSy&S z_|RXJtf09RR99q{?Kd-sL4Q->N-ql_S~l`k<6*TCV$(SRY_P!16MvwaqCGoQa9Yxh%My?jbUkA#52_9nzNf-EcF(7KF zdXtyq9WZcC`ef11l(owno$A;CaFVTg|GN^|%?g(%Ek1`V3{$69oey)M&NKWS!93Ok zipqA-S?xSRipmTZtgy2f;In8txkJg|vgRQI?GRyf&A=OFG+{udkL|@U8Y6d-8yY!F z=dWD_>}dUk5`j*G>RD&PsiPSNQK9?+S%Py*a^=CQ4Jw!))$?!Zx51t^00rnzD15Y*;ReU=5J(34QFT5_{CR8wM& zthw;LxBZ^|_IK_8_xryv{@@oY_qY(gbtKeQXXbBgD=s7}%YKdH)>~^h4haeY9~_QSRR`L`3l6EQq(@od{zCN0p%EEiw+FngdV%qqsLOisaw?$+2` z_7}eNuxBR@v%8IkCV1>@Rk?3;1 zgBw9sfe*(rI4>@rHQ=>M=ej+zwM7Pc5B!mM6KanQ>>jU+3dP#%)*3v2pq^odRq9Lo zMAo|XzgWOrDN67#?ypD>7Xd7O!oL6m?w-J|LBzNQ~5 z7i4T~Z3@HH8yNNAf?5gVVcf*5-XRiMajUJ^>49??oAt4%L|Jd}##Hk7>JBMFePU`P z!&f|fa9?P!+;?`!c9>GnPRyQIsT1N0vK#RCI0&HwcgqKb<$$D znSteL*<6oqgP#r4Rr8HtQb z5akFKCE-P$<^xcO=hDBeY-o4rdT2-OIjI0j%lUF9kpc9wpV3=;Ae{dS3NjGfDCE4P zF36kEJ8bpgU1addNiv)f$qs=kNidrkPx`pMvF_mnQ+qNzlmh0b@I`qN*;)~elg5XH z^_*xZ>gC-`E6y&hU;LCSc(>Jp zUUgh#*hZW4AEyu;H{8-RbrSMGPCd@Sm&F}Fc-dkB@t(ZoJ|5B#M%$ocj&hb958qf{VvS7qm)A1E%DX`Q zO=}{sQG2GJ0V3p0tH(t;)~9U=0+sFjP-!LyVzFA|L-@!y=J=udc%>+-e+OW8)hJU)I!A+SPZ8 zf0EF;P~srf8g>BG2_U9DH$OCzoh_*MJVbZF?fOrwt>-X%1gP*AY0J|&^^K@{ZCI@= zcrhK)#?2t00~`$b!lOvNA=O0o{U96{k;~Fg7

    I40_yUAo><%`f3v1sf z`!4;BX#4$E_W0e#z?)S%O2qqKBhC0|N95A9n%S7OTApD_{zz2%wBv2-TNf<>3^+6Nc${oZ>9(T~gLrdqNBDt${5T*ViJ-QfVsL8`9p~BedslMr&2qa{*KjL! z3x=nDxKhun>xHwI$n3HqV{FLRNbS^Fdbg@OjQYbvhJbkHwZ8oXYZUklTA&*ZsB79Dy!`Y|{#^kOQ3 zwQ^u0%iuiWkb9zyF2uD3u2ZI{GzA7`@6FG=`DD>3y9#2+?+EIcdIK3yXQXfVItNV1 z3}~#sfsR9U1FO#j0qi0uc{`quUe6rlY{AAM{zRrUNqO^&6|uG4XH0dF!+{wN>Q;zr zoKf0vO9oUr)>Zji;07T7n7j9w*sR?K**UW+GN!Yg4x9700vg=7jg0HvDd_Cky?aM9 z(26Y8&)U}e6AeVZPN*9o%%OX^?7q!17TWr7?qqvxasy|^b{g&PMXWj4bcgm$*wUXS z)>=zYK(NsO@`Zl6@rI}mVf$}BmP+c~oN8<@uC1+92R}EpDlLC9x^8N3pFkQ=`SmCQ z+gOFyFdOS+1x7Fb^B^vx-hD6!KzK7G24_MsvN z^;7cWeFJbokEQnTlL!bMA|)$jJW>Pi_Tle88N>-5J62^DAmbjF<`A|vGUcA^%Rcon z^65v(J2#rSd-xTJysK09eLBmzHdT+4wj7pZOhbGZ{&j_gUjEh4!pP0d(t6~BteLfq z+nqGmgV1wlj!zY9qz$e-kE$G)uAU;~RWu$XH60i9*W~4aMdKgczPN9K;D#-=LmtIv zt1Nssi8cJ~L@a{0uL@yL0BF+u*!}s*-oY$mx^1gu&bd70>KxNdZC|Rfs`n=RXNfhV zUG{9P(+$aI*oR|hPW{v+O?h4OYs;L7Umv`*07J*{&zuN~OffzfjzO>|Ejmm(g{{B4 zPv}H&0V^+W0U(NVB`Ey7tXM20w z>@caewzj3EMYK2N2K&wbk>3%cMXJh6eSN2q+7&VGTbo;xDTB=CQlEb+Ah(WRyZh1g zj+&ZUx8C{eTRL<2atR}T)Eha*kNE$p^c-8Xv}1%o&Nj~3gA7>x-6?HAUk0#L?U)Gw zFU`I+6?65!KU5`}>pwa^9@Y0@OA@E6>*G_w4k5q$|iKQa7o0B`iF=RZDw%W`-RWo`@r|rd7uWc51&lFCF$v06n{H;-hV< zN_X#dL4_+1-U4LCCU9(TY)~j?F+Xl@l_=}TDJUq&DOHu1Z*6YQj!rU%Wz6^l(Dk#qz3@_YHNe z@--7pDI6id7RoFO+iGu4twn?Hj*3V~VZR3$C9&EB>iDY7ITzb3bDOBp9m>aPha`&xKg^Eb|J7 zI>4PQru3W!FG$UYopg-+wo) z5fzp;Sv^%zTkAPD-JqSVkTf7zR-)s_zZ@f3&;MoQfe6KEqB+@#_At51M)Rgja@5hgN`^-cF zlP(n1en2VWW7vdKnil3BvJ{5)J^aCCf{U(bDn_R+Vzu6I|HuN%P~2nr>S_?QWA zZ#?~w>)X83YDW3v+0}45SEslVd)O{VR`D616+H{u@N}G2j_9*bA$YB3#`tnNL}s#g zrS&k*m%=$Gf94KE&dIp+^z=%}Y*1zG=O>(3=_PqyajBa7)KE4w6_088td)^5u7ZBt z+jF!nVI!sU+2ydtCEF3aSs9g)T!8=>Xz1h#cuhFGW->;KKLofzkR__ zO65t2Q#cXWJDt9EBmLFa&*Igx=d*!fH%5X2idx>2|v|{ILt?8X*BC>Y;1da zJ?7YTO?Lz>6~0To+lrq3xOv`13tWM4s%vuP*18nYC3WEKiDjF<5m=wE!y1QW2%d-O=3U*j=GCq`E|H{^*d=(4|0*ZeZmE8@@b6-;HoNWY zP4jd|OUqjH+%%c9&jRori@XPfg<7!h-aLUnhe(!3D^8uQjG)ZeR5%{R=5SrZRol2u zyEsbR%np}0*S)h&p_GxiBm3KdC_3c1B01qQzND?T{JK-NIsKC5%7#7?`|wu(-A1G> zcoa-mvmT!SH)0vIrxK+SC*!O_Uo1g&LI++1$d#3*XJh~*@am00y3%C80KsoKzuUmw z%X2z#TLho7Wk!^^LShy+>g-MC5!n4zSmwC-L~m?O%DDM8*yQLmE;hy@|Kuu}C5z>~ zyM-<*EoHTX9WCwc=@nHNx7ctXGl)d26XDSh!T5uUDw*)Fq;QCsq~B1R=`eeU7%xV7 zU}5OakQUKpm3*6JiNF2u=B zexXwWeL!(DN!eN&OvWwz8>@605f8=eLmES)Q$BB_T*t=70MN|a)7vzdoQ?Y#amQwg zUP!-<2#(nM_i!NlkQCw3()#(crycS8z+rOS;bHb+;k&a^JF`tK=>08k`?;2PPQB@D zn@V!3`AT?WKyDXz?P|!E7XgVcu2C+LJ;hZ%SK{>qp9ml}*4J;Dw{1J^mQ;Zb=eZ0z zoi63SY_W^x=S1*E_;Vv87oz`&RE_CRf=grUy)h0k9a5LjwHd)J_>c=bdOEK9N5 zb-3jE+YcjfrbYV4d@IqhwDdr9gpUy_G@(Rcqqp$(NR;>C36YRa`?aeP>+nH6?r2Mp z*P2uywIBs{#4RJ#d9J)^@UmHp(FwB4-Jm%W7NL{BM|4#*_T}(RYZ2Q84T;+n%Eq>d z(6`vpFQu2CnQ$eglZ*=l@V7{I;!Lqs^5S{yk;M7zz;SflAYol-;lxF3(0VWP^aY|a zkCnp@Hup+YvrgkPi%BUxS_$zOzP_H;-35|hRl_e{IjI8|tPboBn3+AT zzOM`$StaB;`VC*JPQuRF@Wn-#crX6434a@*NYyxYB0C`6Gdd*LVGk-kt;Agb? z6e-foP0nZx*?jomlG5Y50L%G0@uq@9OJK)-Kk-Le-63WDnUbt5pp#-5Fvt^AQ}$L4 z5}nKNP{qg9pPIDVe?@rt08)LoJw>Y`GZC_o9ImX@%dZ~mK)KPDS`aH1pP`(ao4X)_ zoV)Csrs6fu+`9nePkns1Z_&4H>*k0`TwVAh)KuGPwJcn7^1y5=t9D?AR3OpXsWvKD zX4)9&79P+&*-~3H5$Nr`J@03GYT(@2h{Q;3m*dK>2CZT4`&+NF;vMbD=cy#Ds`nrTt<3TtU?M0QaZ4(5+D zbp6OUhgmfBT1@MUJBe;=NK5)@yBPX}`xjj?>7L+i}M~?LQm+%L0O{One9C z1jlyiY*8no7XD&d>Ak(t)7vYeyG1=OZ5X9Q%-Mz4ZQBxPQ9!&LQX?H@Gh*Nd;!;9z7T#x#-n%mac$N=cHQ!f7jq4?>!&?_ge z!L|KD>Z$j-azciE8fWY2)t;bJa=DeII@Z4}KH&F!7h^F*i_fG-IDKqG!fL-5k945VK9W@NcgY5)!COF&z@<&Fa_Tu|*o(^19a6)<%udTbC!<(r>3q z%e>nE&|*z45uY9`E*vY$^0x`SfB(A?d_jjjYGqY4sKMZzC0d?Aq*_5C2MhwY-z8oz zf_c8`tVADyM#z!V8|(#F=z3Jt5M$Ax;n4f(GLQB>Cb1RV-4erHJ^1hMafApa%^mD9dHp& z2R()Km>n>hMxBoG%DHV6)wp}sc|jO0joz!vk#|()GZW28ETbZ`j2iM5m-tPkE34YO zd-0qG*6xmb4)W`42XDn>xJeVNSOvAza++3Fo| zItKP;xb{7)hp44{$h=@+z~PONFSG6$f&v^7h(R^-QE)0A;M<8;j<$;Q&KE*8xo@dk z*Zf}A`}6K2P7zMQ?Oz6k^&|1331-alvr5vmIM?5YUV^Ua>{=8QWMxS%h#jOAWy@y| ztiF!+QhBd$lgpC&jq&eGV_8jdmZ?(t*42UnGU`7t&!lL#$6 zfD zM(8gp1}6DRD}`r16G0;#zkC4$kQa3=glTh282*X9K;Tdu6&kfY>-Hgh1w0xZ4 zN!c=TAEso?B$$>q>%*Q3>JcE$BjMB(QP4$%?Ys1nF|J z>VV$blza9Wx%HpL+RS_6xxmhHdpHk}FS@Tw?|u#2*xcj{^U><%E6TqM21&QAR+9iF z^e=0#K1egT4>pu9^{=tQ!orE_hoqgj)kk<7F0TydW7TpqOIFsEoe%*taWFJ6v$usp znT_-pCui92P6fGyL$9>_>E%EP6?1}y~pkYYON3+g0wbY4sz4kz)y)(z+}SQIpN z1j*LBKsqgx<2SS+Gx*GRNB(AS^~Khf)>!xMipol79JNTm6fqMr|IIMlE)50$?Tm$V zgklzI=jeEfka136j|Xzje?}Kc7R8SZ&+2H*bfh?6x=p;ad$5hdzPxVhV9S_8V5HkH z58|$h>Ou7**<~C8A*4aI!pD?p| z@tBzN^Np4UuVCf7x3}Z3I4y{9sR9f@-0+xb#VZdOltatCbs?taHhWl&6q>1FaUy~ZC??C^yL+cXT>E~oM3Yg9aPFoBSNh2`?yU}H}_dRwI3Ri#@y0A1SP4hxJhr8 z8iGfyZM{R$FPaZlqizh1k55lFsmQ3lgZyIBSMmrM4qR&6F``$RqWG`-m`MhY?f81YI_iG9^>R&nylX||s04Ijn*++`B zx>Hsb1VHN&8t?9(IsvdjjiD{8l%gQCeO9In|Bb}h+scjs(g-}GkB z};277vX%XuV44e$#Og2Ja?|_zgZ88e>#?LEca>%{4wI%kaMVE}YzqEGb z=nDCV#yT=-iO^`V%g)rqc*rFwnOS5k=s@(q-`xC6)?D-T#Ga?H>i(nRym%pu7TJHM0)QmUP9Bzut9vt> z_Aa~$_Z8#Pra{8yV(}JA!s6|n{OS|??vSG&0cXW-9+%;S1418Kej-$I7Wuok#x}?U z(z5;G@R9?8eWCxHwYBxaAPYM_{?(yf{CA!E)Mkl{Iz(wFBRir2cft1r96X~U_dH{fkFya@W1_bpzhE&fp^e3rSlKo)+W z9Vtftn?K6wD-oIzm!@+S8J~-Yx zuLH;ZY{D}`_C30!FztIX*%wbDCL$O!a{v`{2YEl2|0#9o*g=_ky}a|7ovn3^_jaKh zMcqt7yY2n5;Y12l+&TI7b-J}%;sYBc`zYSo=8=K}Q;{;AuY0bosDk1Z@L+Ipweyo`-a@W6n~ zC4#9?W`OIuOSBSkUWaqmp-qXf*bm>L)e4?&kG*1Jl_b(2oAlAm5*Wl=4!BXG?{U-W zmXdt)o@}vCIy0dKBU|})dVNbGZik+qFl5sWK)&l zPxgw0WAA$3R?0s6Pq8&H9({a*ek)t0hB;p*V1D4wM=`b_YhQk1JmUYOw?A~;V3+=6 zy!T+Rbsq?dqBn&z!-q82U?I%&h9scQ>c%HQ_tjQzoK2~;ulM$`lZGle)3*}Fnj>*7Xa&yo7-uu1J`#jHGRw90Sv%2CU7u#-Q`Kw%l zmM&f##UElTwX0VQy!$?CSTLv7{CyIPIm9j&z2?S4mhUUnpsH{d+-ZPjh^l%u1F2_g zHAA=NxS{ZR8LDblpbFdxnDV%590qoS4zb^*Qb6YeSPQ zKaAjZYAOut@~1WVQhlC@uw!?&lmg4AnZM*!onk!6BaJ)uu3-DCDRAd4DACT6tn;l_5MIkY%l*^ zr$|9VW*2+gVN7d!cZMny7u48)i*3Q@x#ncv=xNTuX~bHlnFjprk%MRnN=IPqw~ZyzAZ>ih?HG+%@l6@?$cG z<&4&TJ{(tkS*7s`@@P!s1?Oj$G0)X^H1m~6WA?og)0(aT{fhlHpTwu#{ggJ+-@kpK z!f!sks4%FwkFlC8_gJ!uMTb&F+!1zN1Xq!E_+W)>Gx`3a1;i@IV81+E%RLMQYprE{ z^9@Xl?K1=*9q>E>YK9(zHU;dWP?kv}_4VZ1jfHuC*Chi=ar zt|h5>f~u3P4O{cMF}`F)GhsfqqnQ}$J^@>)x{()VX>;LBD>~w>e66S(B&W3eXb^Rk z6QfgH3?z?oXy%JI+f^*lXJk0lTQ!rQ!>nop5`d>!VVJRfX(EAuSyf%#W+r(88gJu{ zrxGaIzC34nn%Y!BzUHMWkboq&x5&$8)r4(jWPP(9pd< zuEisuwwn70TbQD6idpFz+b7DfhKTi7h1rz!Qn7Dd|FZ%uL;oM`CIzXDS;6`~BXPnc z#1upheeEG^aTMUZ(Q5T{5g_?LzQL&+L+}pDS-u4DVNleLGMFxouG;h{)0ECm+X2>IVAo0C!{Wx`f&V2u@ z;Ec>UO$S@s)B}1{?HGfvox8ilqjuJ@-51M(S9>m|8kLvP9-DZy&1}sTcQ$eq)MfRD z2tsKc;6sq|%^@0j1kBPnLf-kHu?Z-m>g&)6t;0?Z6;~paq1$#>+?V$bVkRIs77ax5 z_C}X9c%=J7KeYGSLbtuV-%upTLn$H-A-*aMN$RyG08()dF2wW>mr^Uffs61+Fg)Xj z)Gy3DwkLz?5$uk>J#LB=XF(T6q2qk*z7e)!^1Se?L&ko>$v_Jj`!Rb&kSNyaip~oh zwuwrf8?aUP0%Rme^tWHMI3XlarF@@j$NR+rSVap2#s5?ExSNxOOUOBzCJbn;A)Vde+T8LW&Le}+cqE{9WydrZok#=Wrm!v zC$5K;O(X-&lB0#VLm_!ye=Mt5Wj3OcuC`nN;|22P%~vs!#=W7av` z4ZoDUpjYVB4=201($c4JuOn)MfBSj!G!Arp@RF4%1)S`!1vTMaFD@X}#63!j1$lX& zZC)L3R)zZmgbV2Jt*qWE661UsxtJ=HlXZ4c5vW~#!(flo>@P-l%$}nyo(J2p0h>~a$z8?21-XwH#0K4;~e3W0qfL&5L6gAa_ zN-g}%GQ%@PTg(wK@gc%oJ&nCS3q^(b-UDR^`qiUE9Z@x7#M*5cWN&XK=s5`V<+i_n z@8_p7!dNZ6ffEP_S&=yr(m1Jn``CECPRlkIlSN2*x+|Wod*ip)+LE6-d+5Vzwc9~U z1u3*xIi7K%y7XcSi67bKc81QMVwkUxE3s-$1F+DS5tZ#G#>Sv(EmxA+cCL9k;88f; zMFBH};VzTOBfjQMP4}Mq^M@A5_~3%7Y>$}}nQey`Z>qfHtWi*s9LaM2=INDwfAPHT zn9>2^$!K~{TcLTFW%glySEuElbTr=YD5X9uagAFTFS!NT9e%Hx=b1V;yWXlMMN~no z>`_*ZHyVnzBw{lzM+Wr@C)K`&W8Uzpo#Z1HUys-zCPDMa5J>;WP=f5pbadEQJcZo%d97kg$ffkAt#J=1#Z%{66A z-6!4mRsL|qF8K@pN*re|OqbsK;G3?gFj)N*i%|KhtV}oajQhb^N|13x?Zb}IqB|t6 z&okm%raRkIuNvdl0>2E?Q8fL|XQE>4>09|b?}(D`MW~4;Xdz|G^Wi8g#GzE4^<&GaB1Xio&Ub1A_y3dHMCl!b~DPeePyfQh5U2qZvp3&>VY8 zB1KRLV=(2_=dp7fjg!h;(-JqZAR8i}ff1qOBAV|&d)bj{5*sqtzU*#oU*M2h6Z_+( zOMa^Uda*8LovRO4k|L|WPAn+ap+r= zP2b^eg9lnvxJ>qZ~w>T|HSC+bZ1P4C;T80avHEYS{oBN)j4aB0)u zE`7AzlJ2J2!bK#03G)fD1cnIbY{*Rs_nX}SA;rgQsP9!y&IJOQqv4GTcD!1AU0=Mu zIt>Enj^;YYrT4Qdgt2e|`VhmE>^1}CYu>(|o_^=xTbSNxLQqiJK8`BacD=>9 zD-2Mu+M@)`$BQB_RhDF#oov$IR{ zstO8pg(oryITV(zPvXLFbJx&ldYINHOmCV4teUzb3_A#2XUv?1i`VH>tgq&o)HhP9 zRL#W4@MrsH!qR_i>$Y**$1=$KSTosD&7A#!1%T-I{T%&1)_^w#xiORfYe}>blN&L) z5tF~08SF-IZ3NdwaQzo9${UHZkvJQPvynI(760#IQvXL=UX|ay4Frb@prq#rd;|~w zKic$cOv%4F>47EbFWI~U6vCU1N)SY?c9lQPZIw>g6`OUdad&rj4I13-O!xcj=Q{u3 zoDXk4gdrr8$z;u1w-T!GT>=#e9|;Ny3ROz-n-UZhEU@`b7~vi8BatE+3jLgqM|NHnG;&`SYXuCHwA_B!KI51GP><2Rk$2^xbwCsO( z+0ohL0KPq0GCC15|MSQL zV@y&Kcqvvi3-Nyr;2s9M^S`h6e24VEFBO3Pe;uRGyFdSPkRkFy|9ddKR9R4z|MS{y zEQSAllM6&@pZ~l2|9f(P5tjM?(a1r-2+$7M?DU~ckbZiA-!(P?@9tWpm@QXnZV054 z_q_W~xOjJ$(k3&1;@(Z*th7=WWZ^54@|QL# ziI?Z`x^#TLxB@GesM0Uwi3UEOtTsMR{!*;=Zo`LdY7(Shb38Eq{rUe~7BL|I$;rvt zSrtF&<*HZn3oMb|X8;ua2(z+Yeh$!~;wj#aI9Pj&#aL=khV*IUS*2JI=07oIMq z#>bS&>7~@YExIXUmi=hM&&MI~_MYEDG@BSp=6@O*AaMb|#hK6WI6`VwGLaox6(=+ z+^*%DWAfabh_#Hh=B=sC@mkp*$;g=X*6B$}Nf{YY;#cr-RFyg{YS`%;Net!MO*O0I zV`J9KMX`6*cQQoF`EVwOUBM)f$?DwNAs%iT^P%4ql^u^ISCi@67k%n__2w({i(Z!7 zz4C$&-=24{i%(wQsrA~QZ9%W*yNq=vXBqse=%Ti*b|;N?;@Rv%^_tgL6T}&=A|RMG z_Z4fG2m<&IrOjGj6WL4E1)oM)gVBjDf4MU=L=r5uTPB^OIc_^bCZP#FeE48FU&)z4 zdU0`~?tI1ojOrMf&UPKGJio@zW`=eMY)woncbS^C6Te+4*+~V~exItxbc0xaA8UpEjILn*BuEox4E3 z>{3l6K0C|vhyyvJV|F!@Xl!gOviIX*V3MwVVwMO< zi1aw=`PCyYH7zVORO7GxUQe`ug;Iym44cYOU&kxY{_ToRhSrIUdXYlcge684A00hC zs#kcBM*^)XB$G+_--7qXf!*0lH-91}B{w4IM9a67^6hB^!6_7W4+}9M^i57d_nNFQ*WYBlz0PtJ=^?8M%|v`PurI7 z6J%s`l+D?o-kF;*9QZ^JRM}9Jy47d*O}WKep@=l*c-%}@d6cF^{PwR0Q+Zy6?`{)g z9&jBG7N_!`h$;f?q@Zhu)`%lJUn-i!F{10vCbvc^I~b;EfMJ}Mb{2Mfp@l`oQ>pgG z7~~u(gqc969za9ovhJiVwd6Y_2!~P|4Lt9NWvl%u!ur}$rA%eWF|5zfj-k5B9K_^Q zGGuooLBfr0-*wm>+5&0U`nZMbc6hlr5^flqp8oPsXWn1o1k(%7L=tELG3GqqPNMDnaxxG^o6JwpeG@zh`{)S3qcJj6V=#+StuRob3{O<}J5%qe! zb;6~~e0O=7T$q9BzSHRK57*{?7nCn~Gx=ACm#y*cbnV#0XXN#@#i+%ZDJ(-!?;Q~} zFdgP^-tt0zakP`FcKi|W+6hAaNUI|BAW@PMbSZE%%IffQqCE}FiiI`7_w*IzzTf8A zj?SK!n!4z%exW48!NK9I>9`^v^cs5Vesc_Zdb#VrH&AF3kEB&C`DCfQ#k@Fc_Z6kt zBdg-=ekVF4B!qLryTv1#_-N`*@IFhvQkTNsJYN20HMFt=V$fF>5CG$J7Z)wqeg$cn znxJ;EoGS^88c(28L-^0h@q712`wyiPZG91r5-$VYWTnc|^j4vKe!BcHy%r@I^_LR0 zntY4oVUO_;>nHhLJ#E%CHe1wF)?6YyLP8Wbs916a)G20W>*YGTraxFIHaXNno!U+} zbL^J$Wk0G7?UkmpdxAfl`op8S3LvWof2h~IIilL$QLof(v)CJc_OxyJD4oQBx}{R9 zMsJnI<8;*d`qJ@LGd?kqQW_iP>B+WZ=oIz}oD1u@@p8A}G{Ij&(m42W&|;y2h!B&E zsJ6D|_G}}XkdWP-EuQtO$XWZ<_wHm?GlLO0&N3y~hliHcuhlb;VRV{*EmYgwZ?ZZ& zcP=hg8mymA7xdEGnvhWwOO=LQr}$f3de&i?nas{N-EWS>0)lN<+P!W{REdLRpAKs$ z2{_IDvOb?B(Q5^9A$alCTFCRd?}mU)O8e=43L!e4Rb`7t>g3THkDic#dfPrl9e4_p zb-b;6j6-+-S*=zsoG;e`kVMvK*^l;T!}+mGmUKGxiilyd26|Z77MC-QP=_K)BO6-2 zlhr0=Y*<^H5@CNh@7GsfVc`<;VNh;=WVJ6}cT<6KYh^+RpH0_dpr@R)w8dPrQ%7lK zRCu^(U)saVT{pQ-i<7cT$LkTD5kV*+hCT>Pc31&fxf_fnx5o{va+%1WQ>#>yx1Fg- zCuPr^^vFUv5BcMbqRe~inXknC>J6*NT&-F9D(4T$|IF75dW(SeouA<=oG+gHdn?cVwu1x?Zp- zQ93Tf%=J_8j52q=@aT%WVKv(4rJn|Fb6NBn{-sUxnEy&jU zCrcJ6ByI8tXjCKa?Ve|gHmunrzw=dL zT{V3>@4<@Zr(B)98m99nge5Rul*<)QZYA9?s99+?3U=>XZ}TX9_sQ?M+-3c6tuVU8 zG$`^D>Gf8DZO6m@H5nQ+y%NuaR-?iVx5fQ0<%X%}OmC&gMSpnI~Cw@ z>Nf~3c9ps&y55KRjD_-?tKph-nV_ve?_tej5U{4%+g~~@=rOuv2k;|r@;Durh{k8L zc06w*8#QF_EL#Pm5wUzJf}uEnez-bv!KYo}(NHk7q~<5rUteu1m>Jk`nCPe1nyo($ ziY9!R#gftgKanPu4+5Ize<03<<$n>GT~479!|g@@b*D z00gQ(MUK3MG& zT-8G3rE0GhT1KO_G68g0Ozn78(AL9BGAjYQ#qK3F_DOle8l2@ikxgRrjZjPY^f!JJi2sLzw39x4%Lir}G_Rsh3if zXmraR&z0Lf(FZ1Q-QZ{nARn9afCPQ;N>z#A!o*Qgs8F&}FjF|pilJqv6OWHd<$vp4Pd9YkuxHLUx_o&* z0vDK{Ho@m{nYMRw;Y(ENWiP$w9GnV$A7xtI<^`7&TK#Trf%g;q`TZNEdK?z}{~wl7+ad}&d$YE;{Ps&wkZs5eIA zA`SE{*lAC95LMVJQfCfWTD7kS(x-X9aYf2$>u<=oPIi$+;yquC4_sy+@moT#O z2L}ourTh~~20LKYeJLvrCv(2f6|u$|JbZ}FR?H_OC(or*t@LDCe=Ehi*m!e&j%Gi+ zN-{YSh#|no#QE?M50Bs@0j^@Ur&--bs7<*g?u{l!gIo%mqE7o1)j0j@9`2motQ==3 zCIvf%SUMf=tQ~?Qm_cWDDuX!1a;D>XdUjm=>Uy%vpVEWnlg9v2KZ^XPPjZyF*)q~S zo!h<>1@h{1jN{#VIHRd9o1M_A@!cVJ22+C5$NZ5*4yVse4jC^p$Qj_nv4!iiGd>Z_ zaxf9E^B>s)RD8BiOuw?Ic_hhpTWZ}qr172}AA3W`smS*Gw4#UW>gKk3g85zh$G$hd zdGFj^hBHdL#1qCT$Vy3Rv^bYPB2tG2-|V8i`a0eQ<*cg!C?fR!9*gSxpBi-2^hxHQ zw>Xt*?pZpmj9qJOFOZfJWOu0uQbd!HL{;=`(4@)dr@J$Cmo<%!a7XiO zkkI1d?cH4_FQ{GW1<9*>P!wJ8d77@WvDuGFuC`LSFzO%sex;!H3ALo{e}siM{`M{C z+qYnef&TtP+MB24QN#&rkkjQjRw0YSTfm_@0bF1Um#W}P=n!MZ6^He!oiy%#jxvfPveG-+Qk?9qV;cy`x2$4V z^pDF$-Y@NK9-Ow*yYvzxm@3C!cN>fi9z3Y!0WWzPJN=OaY@fV$BFD!ZZ^8=Z&spX9 z14p{tmn!8m!I@ra{IjiY4snmPG?4Z?-I>TUrw$mlN0y3{^)}(PX8IHkcAOdeP;&$N zUn^sr4c2RqkI?Tm8_et;ugz~JPmKn`$Z9;7Y6>Fp>n+xw!xHI9^Jhh=by&)W z4X}MY&e~fH+S~`128i8yWo81@$_t`OJjZlwlUd}mPWleIr-e_4KT@pvzkZF| zz15yo_4D_~rPs(=)>qu^Uv)37CIE9GFRWRBJ?>7Jn~Izlwsr7}B(f9v7w6#7@h)ph zKi({GkALJ;lP3T7-MRhy?w;#f4~jZtVmd<(LKYi~{4X9zKptHD(4^P1X0pG-6Egv*+37yTl+2&`>*fV&Xq^-fEKhSfP!UsI! zD2ok$`dPz#wd(kXL{=9nwF;L-Rr5WrC+0*8K2Ne;R%zWqMD&`MW?{73 zJs)}(^pezl>+w((TxgNu=$u>tP+7wD=G-44N5F~SVzyd}B^{0`Yn<*_aWaQ&VKBAb zH$H+d?_B?b`A4n$&AkUCmTtJYR$D4d5J1$$&7&#YSTy7fv0U z&Mc~` z!jnk${On)dE(~W1g9=*T+8>Ka=irdXoA9jd2F-s7;|)}#8ER`DkS#gpuodQh;dPl? z1BkeGr!2*Kb?p$_YRZ4ua&wjWsL=Gjf=mj z$DVxFAWiXg>mJK{MC<<6+zTq!N;ATbpRhO2&PYDs5n?4`VSRLF7+RS)&!X2|vExc> zjyT-_z2u{H1VE_?(%E0+eW5yAt?v0WzEEJTRfj+tqfeqNHo|M$p))VXJI$7ezCFVR zz6Hi)I-6TfHJxlWmYCG;@NiP`Q=rg8w$ zNaD0IcAq7VE8yhewdG7DiH5|}2h*_TxcU9^5Swr7_qFDn10jrUs)i|H|BR4f)slEx zRazb%{3(X%g%U<59uIlM^?2zE;KB(-L@W^QBQ8>FYi>T?OB(%knCbnHZ`c=BLO$I) zn;xG5sj3o*AV4m0AnDS$lj-tsGv8v1cRx*GKPX{pS75iFEh;2}&sb%)S#5}_UsnkT ztD_|bR9B@M0JP!8M@1HC7ks8AlSD>9LPo#4LdU?NXPRW>W3b@nc=%?3UMwWPP39=D>gYNjO+$yu1u z@iLmfH-RWYrB;L#tV(k%)@g}`H$X_8$VpbMr3_6d(j~d3Bo_g@7vu)C-93+T| zK4K{QQI(gVCV>4_u|pa*4V>!-@ZdQ6bEvN|N1kW>-><0Jz)j%Hvz5xsVqNS6V9bl< zf3uO!cdftd4aK6AIGd~dg_#Ladr<457xhaKV7Wg2O@YpZTB)!gf(236c6p^v;nk|X zOM;V#CPuCHS|U)>o3SuHi2vqSDk?(D#qY2kO``5Hbqu3e&JZj24-p&vJGs4LaDlKn zL&PT}*o2=5VVYKobRGNb4YZu|TJIr{6K!OxpHFB7s+E{(xtc#THH9^$zkgo7gXJd8 z7WRF0+cP9@>zMysxK)t5CPO1>g79;3PtT!v zi^6tm`-kh=At(e(i@IFs0poQ$-1reX-I#M3R`|uuyA5w;i3*+_uV67e3Hp>+-QB%j z{++RRM%^@l25^qI4zH`+XRu64LLWORm9r85@DI=_D=B*GM|^A+NxgWv`zW2myQ5TC zdl$bcSBzb$#)GVnu6&Ws9I4cd}pT^IWd`yCZ3(Ha@*qh+9 zq-34RDw}KT9t%%a@M2B9X;05Pa`JZ>ZbtIu76x7(iRLRcDneNunH?ZsF`-vDSaasB zS6y9d28Igz4~4iQOi5)>a1NknyWJv1`xdLE8b+G%#!gAsjBF@zT^B18WvL zlIZBlJuZf+SR!P~S3~M*e=Hl2zy+Kp~=#*;Hr!$%QOt zjPx~<`aQ1pv1ESUG(7zcINf8ahg?f~LI?7Ix!5{vl~c2kJ<%%`)70y&Sh zIV2?!JzOE3g7u9JlF*&!z}n?{BAhLkCMFL>##24NSpS1BdOoBA1MMb zE^UROWwkK~KVRBkAx);tnM!C1z8ZdbsVtkU`aYXTG&Hi1&WoR!oS1RSj-qSKweta;8_Fk`u^=R2~|Go~z?pH9^gzRNkC@Gu4W)8@W;{KohBJWH- z-W>C5`A)*A27lm@Z*(HgL@|nz^8LK^OK*Tk0x@O(5&sLXuH6wliEKKHn6}}3nJ&Nk zuB2Tkv&PZnu*>-IKM6C^JmlE}(kv6tW=gmq#Bh9_utzvPilw`Sw{O(>S?l=s_vbUrZG6^^I$49G5w)uddwXypyd+(X`fhNf67A0kb*9GF}cCgm4aq9{BJ|Mr$W<5V@%+v~A z_n2G5ez~GwNf!vX>Y6P+vH|oU=R50ZohU4Pj~j^f`ggO3+}o>HWphbjBpMNKm<*KhpJwFx^NO7sA;~#1dMC6oeTxgx1+U%f(4Uea zB>SVx$@_Uhc&=+sCHGSD@*Pj(yr|iqaB$Ry%_(c)!_vq5n@)69vyubL*jg zL61{sst!~2g24-NvZNmf{T{xbdrZh15{Nsi)5#8tdEjcK>`N8KIFZ27h&dvf>1gAU+O% z9UPFSS!>#wkW%sr;0RL;nOJ$=WH-LrC63`vJjGFR^JWNwyn7yRGUT5p6&rJ)A5TY< zYkWhj50K)bUYv$7vnvJC54e!<#Y<>DmH(vwuNFY!jj+S2GFkP3oU;8Uv4t+>q%Ef9 zpO#&%1z%p?WOerq`FV@tW=@TSX^N&&Hesr${zgH)U5g}Fy5;>JIz!**!~ z9+5k&*}%F-X7`tr(_ZtmOn#%iD1$Qjaha}?kk1t(BOSDt5ZhB=GL&V6`}*u4>c#IC zathY-tbX%>E(ylExHyWdT};w_1ZN&mE&VFbA4$A9m=cBLT<6dg2Jd{?x;dUllsMvz zGNp+nh6hCFJl9_cY<6#q#X}OpZeJxFnW1m}Om)bSzis zBr@_c6OmC!2#5@WlH=neBSVrCLqmb8U8QHBRkypfBL-O|L3-_jX>hc|KHH4oB(1NN zGcl8A?DNF{36oHAV&Vg2#g@@^`}O6C6n*yAajC{c9Z*ds^@mxP5bdJ-sf@uBsV=2E zh9y;n$+8vU8QCNiy@f3K-!D`sOSGB&{7K31uG!HCYkU&!Av*BB>H{m_T_D=ZFh;SCUqkqd1LXfcLOfsZecd= zXzJbS)dT@VAqY@o1*$+GZ(-qb&3f5f`L=A=!&c|96bQ#2x!YzJ{25>;VSC+HHa4d6 zIP6C>dvdyCyy{jPZnS&Zw!c0Cv?@%8j`8>Es`gy_GO!Yrhr#suTTbr|jU(ne>nC?A zSC}@~*n3lnn5vCi90&dIgRSIwGIx_I^d&Z_FCV3_lR3>}-<|a99kNrk~_&=*t zcPiV%g1Y~F?r-6T@qN4L+aFGsistY^&s{SK`vE{6<(A(Mz;Pm`czP_Y#NX%%Fkx(K zEInJdpOGu5>)xB}{kF5{sOU1%toyf2ySlrHBo{uPhg>PO8?i5T%8R*%01LjefPc%KZ0MV&d#CbK>rsbh1dgI*7*z|tj$Avq4r zob?U^XmQB0vFSWc&D|VL4ganWW-pFn>lWvNQe4Jt6iDk18rU!Rn*-YIify}M{hgSu zd-r?vv|pNEP-qU7UcS47(*aW>nK` zHguaMV|d%#uelu0v#vV6oA%a`=FJvSTR)?3Y$sq|fbAZcB8pGHP~bVeMJ zlb?_tm*!1M_Y^ogd1z%d{XOxeK=&L|G#FLcWy=1NZ~apNl@so^2Aa>POIoR0SD%>b zld7)~?|L0}qPc6iRPjDH#O6mwyF)Nq3>G&xclU=?d_OVKH|qf!PyMIA)(h9i;MUjq zA;9);xz&tQV$oO!aL!u2GSv@o8U@dXh&6{zrWwqU)sDC+hlx6Q2}UI)#GgkWZy{bQ^bMbjr*?dfyW1NM zz?q~GneN-6p9WNuT4hPh{&xApMG~?8T$wLtF!?f!ZY(}wvqs5{IxJ?<=>^0Ej_NTG zhM`vFU?&Xx`ic??Jg+Z%rr=ydifN!6njdv?TFHVawY>)K0NQlAQuhFWv@~1k122r8Ahd+5 zepQCoX=-|xRc~GqXYMd)2~OMf)wiS4^h|q^Op*$SNPnuD@_(cosk*^Mdp``SBegIK z#XhA{L!l)l(!F7^-u|F~KzK0;_#_6bCIeG7X-#49ydw7!o`wc@vN^Zu_kuKmbnkub z@hPlB3Hd!MZX#E{YO|oMH_S+U&I+vt8*OdES+p-$jwi5Vh9<*_)^jDQKoBOI4p?G! z^B8OHA0F6kreRJ`78idsStQBuUxAQ*S4+(l)LYDnAU~Zf)k=ACmW+1ZwyRtmv~?U$D$de~DjAwow$0;bSr&3UfL z@j0wquMeiud7T?mxfH&H#g2D%b+tWR3m2~dMgWht_E9yNVr#0%+Tm2Lgo-0u4f)}- z+qFXDv$!U+g~`SJVUnK8)Dyr#leT4aL@MuZj^=v2R7+G4L(%y=Os@9xaw4i3Ie^b* zxmFOV{ewWA&vmo2B!Q0q8oFGgL-!@`YL-Tsp|M%uKU4Yb>GJk-?7O5a9SLaJRZJX0 zg;J4k>0!d`!}m#H(;|Lf4M@GWwtN*8yAGcb znR!*USZNPP9pF4pYJmF!{#q@G+{ks$#B}=D}Y4|iD;QVJtjUEdXXL2dKtw^!ZWnE0BYrkG@ zR64%xT_4U+f9kjS_3OOltc}VO-4VK#6A+`qAMIOqk|zE!VGt~uFA#}Z@MLlWTq4gI zP^S-NB*8+7a-m2d5|EiH{GPG;*LtP7-ZF_%tG_F)mm+S%gOQTsuWT}lJ?Hu%d2_pv zq~zeyTxsN5JpeOTTGaou7I6^1`=?9%+unWq_@DOxHi6$q|8FEy==6vHn-{s3)izk7Yk zpd&w84kiI<7ux73+SdUr0$0}^F$wTMaCdWrS=36j@j8$N)I~)_^W}wKDR^8CSsZRp zEa6dpxv_C@;4~>WJDOjBMi%XMAb5i79&hV>fu6oHDQFZ=pWbj&4fx%!JM%BvaRQ!e zVGu;DR^1kKI03&~chR1cmG!FK z0Q0)z?KYKH&}GBJ_Kui;dU~2Tg-JLb=5)Cs^K~yNr64~qPpD}R3=R+PnV6U`!#rJT z^B9IF;d-J`uUN&B=ifVdU1gm}69nnnG?+~p_7*9k60l$1T>(?~#OPT^k`omR3+7f! z>$PQNtNnH_;)U%@yVsNXOrg1x1Sv_k$*_;#WsM>X8Iut2;??7<=}O<7*W(Jinedl) zzV4SKHlmV+qNxth8!R%;poMNtv>jurL6V>Eg5E}z@>Dy=Q&O8jWpr;!u z8e%`t+lvceK9PC|JnqIJ-fg^H$qR;p2#WA~!N-gpveMF+0#e_;^|tK{RO&TUlHc}`R-OZ^+5E}?>@yi!XW3J1cp3C)GobeU5rlTYhiu8w>z7l7q?>U zLrs)qz7=?4$V^KxsL-uP5}Zilbvl|=mgn1IfKha$RVni607x&j^t&H#`r(aPoOLG` z3x?HyWO2ab=~(XS$!NAAoIS3+HcYCt_Q6gub|Fc|^D-89A z3W|hTRije=S6JB4&``g4!VH;ZiA?XXc!}}X+v7?zh~7+xpmV18?Gg~R$tx&8@|xv* z6*1}^7$7h$*KAUHmRWTqulg0|Kr0h3ES;ju#W1 zBfF71&`F01N{t>79-Gto4|^N>v9w6taOP?J@SF`=$rCSee-g3OvO>t`KSAe3;%6K~9I5V}D z>o!1IG-}i2L4ab!VY?Dkx_J#u8l+rYKiqBRn%OxLD>dtXzPKOOIwUcKgI+$y`c7!n zn)1rm_`z#qL*p~NRJ7wW0I?FNzXqzEhmDit{2n;;x~mml9)|;>DiRxxSZH#t2iZ;p z9Gkrcv!2{3Y;2spPVqt;r$qv_%A;FkQKkQTc7 z)tl=Jff)o;`NQ{XKRSUNBU{j9s$qosjmKF>Gy?fWY_xd5NIZS`ar@)raR@6}#0GgZ z|9p|1k#SrE)%b4#6Y(CD8?YW|$nSA)L2ty5c2ylKxt59R^MIYoBPT6#%?iQ>j13|8 z{qBtXj>vhaP7lu!A?2lEDj9;#Pt;wCXUVH#l;?D*I1H{<(rydUhkSKiVUD4LuF-Iz zUt-GFo5C4>mrdr!YUngp+K}4A89RDHd$$e}&?AlP4^tpro;^rlnup3nM|TVu)w3Hw z;wwLS;sHz|mrg5xGF?yim1+T<(Ck?pg#hQ%rw~!V90THgtPFN* zHw)x0YGt~o3~1ETl2{t`Neo()nPRR!Y*vd3iCQP^f+k^GBojrp?Jqc?G|Kx|jO{{R z@xd5c#))Ov)PnU#v&=fk^0k42ZxQy?{1KnHr%mEh*e!!{&x2$>^9uaoncM3`MMrng zK1iaF*(HbNoEud;OuItU@PTI>{#};Q5Q%35vt6;zqwcBPyZqNOcpTZv^L5D-h??VO zh}kd0LbjQ;90b37fdT!?V6ksdwZAsg6iEOJ_C6Z%lWWjU&fst`iPx;7A>c{9B&-EoJF ze#LK7KV!}i0)o&?cCcO8mql&JE(HN^z_a@^5tn+}vlf6}>C$foW*7zEtnANLfMBQl z&E!Iq_z;lIEK@0-u1Mzv5)k-!9|P~|uHK=zjtiV``ZHHo6UD7b*>?NUUYxS>#D7&420_*kGht77dqrKg!i@Ocb>uM7a7wDF~fOwp3soDWm zF96AZ?i|SU#ALL+|K@;4#>YeX^K`l1%;6tG>5>;58cbqPMWX5k1AG#lkjn>ZYPwtR zPZP(%nMz(y%3P_-@`?JGR$Y-+1LV=^7^r*M@gA@!^2#py;C+y+h&0j)CJ}@3Cwi+zhhH66;IWkk zTasl_SLB=sQZ-l}fLMES^LYwdswK0S0D*)EZ^uvhQv5xbn~} z^N@4+W;Ud0-+f`tE&BrmA>~gW5Kxdbt3A0518WspBO|<@L#Y(TKOoq{lsjuE@T&mR z5&-KM^VSOFQe)VukBYt`pi04$2A>O7g|a0S6%{q#9AzKmNS-WKzdt-5wyew^ZA8I> zoimef97!?be^qiH>a;s7*ZSKz9?BBhxpx?|>K}~aThxEZl$NgBX0m3{yb~@8}X@N(@Lp} z?b1$A#Op(nURKRyAGu2`nu2r<9Sjd%av&6`3axs@^B)1+PAFRBmv_{XA%)lHqYe|< z&wsZ4bHb^hqPhtn*e+fsmAWAjL#e?^VXIvZ_3n}QMSlm$*+QoH)$^r4ZpFJa9=jpq zZg(Mj1+4lW<_&wL*AyO{-d()0s<{QV|%C z(x4CBn$GP2SE=8NV-2^2x4RX2HC$Rr)}jAVqN}F|i6$Zn(WO%5y~~8DOW%qp55h_u zPMfXdIv~da#%|6XGy}rcYLi0 z4>&P_?VH(?ymqAM?l%X_7We1GKx-=TgInEbgirzK^#KU~+if?n{>(}b_&KR3leAvg zVIPAAM09V^SL(>>NZoHvLeI50EQ5ASH0w)f1njrdrsSf&iqiG&)$lZ|y3BQ`LwhjZE1S}=ZD7@u`-nX0s_ zQao?^aMxNKo9xc@v3;mA+>y3ij#u%d7K*(*9OFMz+V=(^?vCu{8#Ft+OBmmtmArB_(QsP$zW`x-ic7pwb~I3V zH=IZ4QIF=uO0oIL^V1%`^2JYcQhx7MKAs9D5w8~`ZUZFM#tQ6nO!6~*X3_23+LK|( znxZSvWx$OE9VMFvyy`$Du`ihParRjQg*p2jph73kLYBs975-t6EZ{le?*70ix=Src zKn(_&e?$K8fX#Pdc!aUtECPvn=~01L;PJc%oZ*ENt>#wvOL$5wXn1e+v+zzdiC}JF z6ZY0Tdo!IXrn_Lc`7RiMh17D8N%bn7v$a}@sBbY@WP_%&BtW~PPLcdelwE-#aic6nX-#m+_nyE;9ftxizbOS z_T1{jl@ng$_2z@XvKImVo-6Dfpb52(+jQ)p5ni76`<{H`*uwr+59p1Oo{p6*9{Ko< z9s(_N>$tM;yC+WX;a1r!RFa~IK3HSh&f*97=onWMI_@*z3cQ~8oZV2uypsL3PYdsY zU11%kN7_Pi^vt~p{3!)+WYWB;RPqIBAMDJFG0vjHBHZ^CO>hiekJle;boVkGu%dad zpEjk!Cu973t}zfE)HezB4aM;J<{oS*KjV{1PrV@_L#1Sx-24qGwk1Htz7EMXa9QPW z!Jcq$%BF=!Z{@bjnFQ}_@#~a0nrnZz<}aF^3P#9L|H$96xLE6A3nB%w?3SAFdpkfM zZ{{z(dwQnlvt4VOv3mYwst=O|-sxKM1}M5gN3QVoXObsS?IVtQ301doqIiy_i6GlHRoJ+y4iejQ%54&y(d4nChYNuXM%3=j7J+=AlJ7|7rnR&q!@JfX;KF=>=YMDU{C?@R%=CzfsF#wYQ;iKfX`{ zdSquY*bY`i(~w4e=EM)Lvm8;lA71xMD<_Xmr)t3)%{Q0@9yBqNNL?F5D-ANLG}f@x zw&Dtaq@BPae!f(g*{!>rFY_;0xY?BQ$ud!L2P&l&BY%2tb|~(AsoqjO1WCE>!vs2| zeBz3v5?hT<(4l7`Lv24yJYU&>y$4{89~4R?@nMpvqO0L^=i{{AcO!>e`SbB_k_MXS zp-2a6XAI!6?Zi+@{#sJFjZirZ`642^cO<9}wVb8UG>9O{OW0hZ*I_oW4l~PpU_2wO z!+3;`PEffxyruGzP3mYA=(tqnR=etIJDtbqNY{+Wl>-fQvMjg}2o*&xI^irZmMd$N zb2YB^MQmS2cmnNfycQU=f$~;`kWSoI7HnuD1IPI_fw_k(QtGmV8d~!foK>e^zyQMZ za2*bISlsF#HaC{$xy088Dz-mYIjh9P`;tN`BzSH<1I3TuRIe%$bkj0sJ~DFI?;<=3 zPol!q)&jssu1{ih(!;MT0MGe2ipZ4=uhIuf=Lf!yzN8e4E-eED-wX9GP_bLldFz{z z7oa&-@mlkXG|c+`QP}x&c9pNFl4dSQ!C8LO1p1ibnVbay`s~9l<+>cxs#InsoihEG zn%a+?0%n^+$G`j^t{SX1Y#fIN(G;@LWxQpRJB!ig11IrKYENaLQ7)=B!I@KZ0Xzp4 zkUv5f3c-*aDGG_-GL|b{{_uMq%9P~8+=d}V+f*II!AhiYMdFe8u6m!gHn+}aXXHz9 z-GP>{PV>=~4=4H@hxPL28DP*xgdp;F$O;b`9LX>DOJvo}HptCq+w}SLvMmxmT36)G&$cTv(LkKCN_HQe#JbXFM=dr)ZsVK**A0O$ob zcg9M1{9|(T2J5A5&p}b7f6cwIRL-;6rmOoWNDunE*Vtf^L~HVot+)oagg|N640b!% zGoUqo2^s6rgHMREn0fyMl>X)VMek{nf@g%;fN=)eTQlnx9d|zkESxqZ5YUKLlPH@r?~1 z8U2mCrEsKyAe65IEn$x8rR({npr>vTO9cHQG@5r)WF(MAh);3e*y~nss!bDYx^@Ut zO2&PcXV}qCP4N{`8Zj$$B0lQXtbL+OLBB;ghi|E5{@266C7@+y7g|w9EGH5O05^zjOd+(tG&HXmZZaI=h}XiPEMq zAD}fT0@r^f5yUv#pTntOJUZ@v0e6W8=L8_UOdVNq#D`G;|877)sBC$Em780egsm!a zn*&uZ$$ymMsG)4K1L`nb*7u#n{7!Ih>(1@ULStqUgKvY_qijqDzlZohZ)aYIwmb}t z;3UDEC`p38kLf~Ft!c*tgx7^a`<>ielML?MUSS|IBehSPThaV1i5Vl$ZRj^zq9{y_ zBiRYj&|pq9r4vAURLH@2MP_V+spUJV$k^D)$ea-KVwRFMlE0io=Q%2e%@#CZ_=7yI zUB$m+5^h*YR@-oJ7S;TS0rkgPu}{S^puo1Jbc)!%R;e$nfRGR160l4U>Ojs-O|3%P z?TN@m2=v{h&K59GnDE&u+b@TKT3Qes7SC6~$8%A~gKhS+)|lV%&^30Om`<*RGBc1} zbO9l9^izUXqshjHlW<0hC2_{o!MUb+yZ?u}w~VQ?>%%=;ic{R(ihGN@LveR46nA%u zyL)kWx8f~MDehjZI23m{i|2XY3fN03%BVnO>t!M3?_}p&s`!Ti;YpR!|Inl1AEJ>Ii2;Edk2X*PAlGH zeApF~@xZEvumXKn!oNHBoq3~Fa`TIjmCiC$O2%a@+iHyx@!yInZW9ZZw)}aTNT&pE zjzM~7Of%NCu7^8I3>>Z>^RKKs3(j|86m9k8x(P{&_5NN!wHP7-H59})R2^vVHJk} zEm#SQvc;cd8UxWpVlZh9*h7A=UiS#pV6yEt-VxS|GQI9xktLfzCv5qsyI__j!I~_t3A0`TNyVNB_87xma$hNZuC+-;8njy>6JF zHaC4!Kfu7j@|=U_P1K^tuyu2(5z0I2jLUXy-13qf*Eeh(z?)u1??lx8 zN}ExN5beh_Ati68eXJ`8g9Ax=jn@0udqCYk24Zb)lMMUcmE%5kuUw?VjGfV9;Mp%- zMibTS(Ru+IkJQ}1kJ}FzFOG$_^4UOZq;$Jn4OersPTX@)miJg@3~@*PE9%+UcdqZf zKNs07$KCuF6$N6#Quyxv-sw6b>g_x1PJqLl3q)YRnJy^B1>@h$2I2QOVz$Niw3UG0 zZ1saj11X%HKK#Ylu9n1!08PMa2O>Sq2BY}+`00UKt0I-UJ@G-KcBu8|R5>+JW|$Di z?)(XJw-JTD=7BJ#!-a~S!D~-ON!f3G5eGiF22s9j z>NVn`>kN*Z%7B!S%Q=1hE}uux)#YraeTAoC=Ei0{T_QaYede0ZZ|znwF!9NE&qGi* z_UfM4=gTia^jQ)3&8Vlp`7s-rk3AKd96G*yZ(;x1UcllE)8y;lmYQY>S-xUgY3q+l z$hTP9HKD`#HcbQ>0M0C$JaKB12K);yNk(krvtOXlg0tjsb_@3X_EJhT0x|LFQe_xQ zDa~NTCdFXgtH&NJ3mJRScrwGW`TmZ!9+6pCCdrR-avwhm84(5}WlRE@F-PKwuxAcf zfn*l;y_|v8rsb$o6S~*SAKZT?dfW|i7sOzcr#r7(|LXoGMErpsBR!qv@)hDIE;5WT za~V~XJI*bTbO)^Quy4jJ_u00LBx}4acrR+C*km?gmhA2ai%%&fIR$_$B#kqeuNCA# z6*j2q(dYT7+1+u+;?5zU`C!dmN$56>%aa}B#Bq#&sY4ojvqxo{&~((3x=@m%L1!CQ zJa&J(xJ5GbSbB1<%OI1qI&g_Ra`qXuu zG0$rA(=lV#yifR28G)1rHqBnFT+KmLJhhUi0{DZly*P4v^W@3UC_=H;s>-~=G zJt&2De#_Z*9kFlWO%C8BpCA@BJF?8@VN+e`;-TVmc{JQbCRxpP3^8>ebYy$JvD=i^8vTFbx?^)e6zhG0+i6$gaq#MU}<$&K`MV8ihuN3GvA z?`{(Ql{zXK)9>jq-cO_{so%wX|M2bQv{GhdH<9reIGT3RI?amph1U8KhB+AgFHrz* z^&1djJsyk+oT}^tB#c{^&)pD}aP|ze`|TBoA$?%$<9QrWFetOax}Hv9WWMd%3=Cs3 z;cFxqFbW2ov>sGToI}}S>|9C*t-3~2(P+_vehoX(Za@fu>@FI0Fd(#@hqj4_ivAW)K^~ zwtF99pC2}35X3LZ(mr^ylRYdE`{sO9&lg8*SD4T|;LZF6ISI3BfaYroa+YFwrufgu zG*dnH4nS_Zl< z^LnsG3M2F0{C;!i&sZY5`b+k2Y*%hgv) zgC8%#2B1eNb5?r>3uM5dopt}4&Tqti_=e^!2PW@JKji)gYa=MmU9!>&J*Cb>mp!-p z%4UTr)b(Ou#~&)rPM(BPeJj4w+bjA`amC{@w%ZY#-j4rz#dX#F_;%+sRJ++epMaC$ z!oI1S!a&3WeR z2=q9bq&BNDKC2aM8v!k+2hO&zvRmr6iWZc~W7^KwE{aYQi6Ln7b>3|RP>8M<3X+$h zJ6Spl2*?eYcWBC!ZZ4v!6il8nnZ=i~l=YXSaVfBvi! z<6_Rp6b~e+o0u36GG*k4_%0$byk|#|xVt7ncz%Lk z**w{^)FpR-8DNtwB;4_4oK!{IDd+pce7QQ4t5LPiCl8*8yz})$8aEvNr(Se0Lj$(! z$5hpGDSq>o;;-7^@ZOksXFP-6#kg%mJjvV&$OV7ozmi9DiWGi!J(zei#2D1Dqh8jw z^8voS?LYf~per0m%*BL${OE`9L_@yTd-pf?VY9kx+$fp@Z`E=aTI?Pk4|A^Xk*so@ zk8g0ckb1f6C*RheL}Y#(yA|hMDM5K+_^gMY%`eq#ECVIjuh`Y?4-OVm@Z4vo9?6tH zu(Len_y-In`r5FV44gXh(-q`)682TDJg|0`nE0Q0vr?JIe+a(5OwUMUy$Htt{hEQf zeRG07Hk(hQ*@dAYoL3QE2Y33-^-4$+{VJ;&!S9b{a*2<3!%XOxl$j5(kc%Qfd_;J( zcmRXgeSk8y4R|9C&9LE4Yn|5VEc?(W@c{0DYyV}uZwZy5*fiV6M;~w&>`|F(AJ1bA7v7>K2nH8+nh%YZgn8Zf%cQv>wXAs54I6F?+Q#3LznqhKG1L zuHubGGI_1#VYl0YIh9Z z8SJ8+Tcg13w6dh3M)MZU@5iBpxD#f<9~i;?8`o>zA$5T+M2X0s&VDJ&5B+&BEMlAk zow>Q$f)fA|KoR!MBhLftW+9^dDUVowo!O0VUo8#eTh}t#94WP+^KCnEk1%1s99**5 z@~WWkGGvp-X>&2e#8t<{2-SKI8#VjCXZhT+ng!GZI%jYm6p$9K0Y%F3eeb1GK4 zQ+OvfuG2eRkjTc8d=tXJWMzTjs+>8=P3%6t_$ke_&J`9CVrBsUh^e0jy2nK8}$3%!|UruTSR6AOMfn_7dK&X^QzTGz|&R7BmOwDH)t zL5LV9OBW|Y!?rPnyCQFNa+V48yd+(43Qq_Hh~~sDdm|d{6DZw=^597*KEt(KXL-h^ zqb&V}rn}C*Q}D9Y^XlF^s!0Bbt?1YovBB!m3+uF-!vl>PkcJBXSLpMtJ)Fe_0Mpnr z%%3X;j;s2u;a0fYTNvv;RBEifI$<6fw3Tysdz64b^<@pu@AvcXmhIB4)ex_ zv$XecMTlwn0s{sACVT`CPIA|DydyooxsEZwD>MFX^!M?7M?cLJ@39+83JXv86$k+0 zAg*&cRzq9OFIZb%f(3NApw`GL7+wxL^4;0EHEcUg71)*nv$$FUUwMM>(uGFFa@y>+ z$!nR)exnsym3XUb`!MP{{zwA61DRCY-SeR#`6138L7n4bYA!DRcK_lWsbA@e}{NcINT1 zyIG!-9P;n!m(CHdzm@@Nm1s(Pz2Af`{H8eYU1It^eEXUnU$VPf@k=JZxx0n;w}Y9B zTNcbgE)G7b%M*MA$MOdsUM(ugj|IiU4#(bXaVk@Df1RirfchvoD|xoo+Y{rdZu@7c z<(6AONX6+}t5H%?(wDsR1fm1ZY$kX6I&x~xl~-T4iWh0WD&9JwApyCMZ@(Qx`1nv-CTQk$oDj*FoG#sum~AH{E6H3{Q{ z!#fs1G|asfL-=~Nqs4(3!Ve=DVx60hN|6)w-7`xqZhj)(UM`rmt+6S9N#Vo^`Y0=Z^D(#*C7kN&GB%GT})F#W2l?hyUCr0dlBd-IjiwV)$-)>BC03GDP{ruq#1h>L{2?F(_rixaIP?7o3=GF$`|)Mkt#^;W z;&@B6+$jR{f_VZjJGG|%GW1Nr=Xup``r!6}`H!GH5M_f9@PXaB$4>J0C0XRU2IS8_e;6D-B|R$-An{Rdh9ui z*$Uhy)DRGd`>=a+Gg<_a-)@vw6M1Z*TK#5v0Isj^A`i0dAI#AncyBMJ&Jxq=O@XEm zp_u;ms97$PM_M4?wiV6?G}WRp%hfmcd9Dv{cN63@0$P2$AvgEqXAWl9kGHM{74v7U zmqVvq4_En2kC#(ZarX48eI5boGl{0fah5xDa|}L9wA@!- zcc2E8>;dAcsJ3&t{|qI~3rfS~!#=8eFpM~Vn>&#l>s zY}2aFR%@puC8+``paik!FN|&=f9dY-hTbdw5{^mNy;N&7P~&~h(7b&8{SBb=SOt;%mS~#oOH|zR2qJ;;G3SFB5wgKppo887 z;E1HOv=)~Uyo}550m$5Ig#S!S(%uhXFn{uG4v2U$rV2cR26D)n+vbe-Jr)MWcUW_` z!J(nO$=b~lRPwJJ<^Zw@1f}ZAu*w2AuY3VohRoHr5av8wI-K+zoLImOiA*g?%bYMpcd?D)qZ9>>V z)Sh5N*N4Yz5o;eFqkdKIySG4U@pQWR`1q*rb}H8M40K>13+DylEegzGG!B=NTa-)) zpzh$a5YKl#1Hi*tm#=_MnOb?1&C%ixYJ%Sz7-YPQ)wYvv``^xUL%TxeG59=J!3Wh{ zMfzRpK<3}}zG1P^()DCw6CC%1ON}B}7sm_ed$%_U^6>wzL`g+JJBFPoQ>AgXBUSijf=y5WlpYOERwy-E$1|>eb7jG)5U{F7?z-o@rn?XtHEq--fsB;<&ng$!7 zTn8p2`AlvK@6Nx=joKz zHehY}Je-8y+Nkry=lk@a&#rx)qVIJIDq%mZ=d&w!hvIj6RFANT zSnXdF-Kh`=BA^Q-lcSX z$h^5bi157?>u~b&YX1-B+6{3BXf?GtZ2Q-EPFkgp#l^+DzreA$KRvjgyzv`cZnVVi z)qA$AmX()J$xk7@%Aw7@6Sm&C0wTyh{Xv0;kDXyE6B70< z#3m!H>vson^uW!w3l+<8bCDnq0M`e<{oVi{zG+~S^?_Qb-h2qE`}`QFFq!^(oxGXX z=5%Bsin+9D#SvjZPDb{hN3wj&eT ze`^7<=we&LdsF(Jn~0L{jUX6+Ir$2B&VBATW5gn`W=a(Em3KyCe`X)}5d$Z_+JZnb za{chk|5`uQiF24he+tl35);%24hyaCF*Gs-8F9+1qX0b9$7eA;Bab~AFZ46aiT zVvNA4Xq(H?)ydM_PFeX21Q_@@;Hq!WV+4r2c5@G^x`^U0ffudkqOtDX5s%ZiI`!&f zJ`b2vcdJo&4jRCZ1TR(bT#pL_I`v$oF*M{U9}%r8%}XOoHlRJ5G`iNabCy^V$!i!g@ZFc9Qv3 zGLfI_gN}{{jg0rsy57jP4G}+yR-K5)MkfFoakfe`Klo%_m`W}Ku(a4JesUn;Ln|K9 zJV_}i#DJZigqis-=-Ju*+(jiH&E$c>3@=x&)Ve(Bdh+<|2eYi({GEpI>SzTa3X#V; zyVD!U|EGYbmpc!z`j1Q6-fJxNjthCsUnu|-5k?--x{fX^DU7=Ijd)F05x zBN6mg)-%f^(UhT&VMPO2S6x-W$m}1$qq#AnX$3sg+>xP`87d+&)Tn92?8Kt zB4S`H!iY0KLCdQmRE@m(_bQ0f@m!I<)K~O=xL6Kay8hA|Cnx&#YLz#e1Bf)1qHPw9 z79mJPDsy(7=T&V6o8IfUJIMP8!84`WuIpn0#9R&q&7UZ)&dOK#T%cRR{*Vz1&oh1W5I7915zavE=z5oj|YobNk+rQ-iaCm zw;sIkN5hk7u{G$2K7RDq!OJuH^uT%~`e3i9lPYm+u&8IlED`#%FXaU$B3cLx!;W9V z5Dd%>wa=9Tw#3NDOT}sph}^%Jd$#-_Af?6DBmy%aHo^}gH5CpH4iN#_RQsIxRaj(! zxPJs$H*YJ6%+fI8A6P{imw!GABO1`EP`;<7B_|h&^@V)LhNgpo`X`Qq*_w$Xm0YyY zZQBog-4;o345y9P)KthYPTV5<{pZcjp#Sr|FslC<6l^$w0MTEN(9jM+Grn)%IkWKZ zcJuyUU&QeJKl4WT|CGB_ARtLo@MI%OSEbCZR+Agh+s^8g_%cv2G`TBeWRZy+ltzu; z5wUaK-NJ~gnWO1;sG5~pn4j{*$5_emG^%URmt}{BBfb$P3zVoE9oVD_MJKIIl4je* z(yMamPYd`6Ggp+kXe6cg8ykf{wJ|M3@XYbOa*Qo{fuxz1+rr`mo0gcGMTYa0+2lH6 z=|B_K)=uj=5egorVs7X1s2o>jnww~>bnu@YV%^FhuU>ck_JF*f{j*l+x|N@8OdNj3 z$qq@dr)7or`Dj1|ii)dP8LTy9SC#XxAIpZ1iL-cX|L`Ce zs~mZF&Quxkx(T8?t`zw7GZxtPte+%Wdn65|_bJ)^-r~ zl3`yksvM7$NB6N4%_oaDr|1Imhw&Mo{BmV&pGrbDWnbGh5|zA1P6 zM%8X*ZKQtGZkINlCDDbRG_PfS`oYC;SZqOVzO(leO)TTM;*94iMRR^xcczk~2gB*@V=C73bmQX}sO9mxY6)_tPc0N5M2D(Q;yKe)auPQ+->?&!g5i*_F%o z4a-%_hFdKi!!@;MNmYz*%$>)OXEnXJi#eF=u29sP)Xa1+t+om%3;It`_(d9tT=ABQ zIT+?y^6gl8=%a#XEjSNl$pW#FhRS~7zddv6&0$q5n&yS%?$H=4c$r-*t6xG7CiJV_@T;3k%!lnYwcLEEC zP3AXjc(+-9asBg$agWT+Xhk0VE2mBb+w^TD)NIhSTD8_`6oVi*1$bm*BgiXc5RLA) z?iTv5TDoy=6zHccho&=}7J!-f-UlNg4J%3vM zGEUtRm*607rd8^pOGd{Qyezf+WcjDXh;qyhZtAkK==R5ZXls&{;iW}?lG{OX&5leN`jcwXzM04Af(_Xpi^6{dE1<=qb)Hsg4YPsO!I?wp@FqxU~z;}Fs5 zf~|Q`=4ME839sWMpo7k5hQo7*pDEOp;KrTo&ffU8dfA>8?Ze)<$Awg)wNK;4E6=NthkacPNCT z!Yx-fAlEjrYz^`KH=T1!m1k2|rXZHyo^I(&n;H&|>k-e7C{V`4iwh2M+RwlL7Bzdd zA4Tv?h=<(P-X7_f@JBhCAMwW-&fVy4S8_u#-1_dIEmCEea zu+jx+fz90oKHrnCwTzcSjS!mbZ7$;t#bSH2?w|L>Yd zySK~fg03&oopasIiG#^-9&M&=@=_uG)#26R@p41GA8MgY#2Q}k+82fUL>0;HKb7x3vds5ovXXm7I`eF`lK=?TCk{s#Pf)u=W%1ef)y5 z5k`k)^e*p`f;|IBt5v^!s##`SYG8t@;YhT!x6WTb!D*A6yqJOk`%|t8f_Gzq#mYcX zuiEPAh`0Ykm%Yu{+V0^wGJH(gqKaJc0C~B`Sli~lla7FUY`w25GM$rj-}EpN&! z%s$E{8$&;LXhE(<83C`jaJO*e48yP8w5J+;F)}t1xIPh|Z3_F_O1s{3p}vw$pqvF> z_sw4e$h!Lws^Mm$cSkBLO?wNy=wE|J35}2+=|hs+>6#V2zee5as5Ll@REO!ak?;M+ z6MCxtzA5dp4(Ht|cyKC^mo2xOElN985Mz#051xCEq3uvqd&P8MWkjSzRjreMjZZ`A z8xMJw91I4HoQt0qfnL@R-W4e59-HX zK1MtZDyk0*O9j>I4ldV9;H>EB@>VFE8u4r8!)>K25=>d<;dc)-98a zVRcAm4`O(S_uCQ*QB3Nn$fD|&h8Gk<9RLisn=K@@e2Q-zzsxZ8jRWfMGZgi$mLLi{ zy*t-F{RTK8$0HOA&lwyzs%%UXUS0jIc~2Q^kiU}O)z)sasW_`h%hTn#I~`06^>Xek zDZEM;jt)Nxnb~_TSYCaevINKICj#23Z!xuqZ|cU?l`vOP)ryK}ztA<|qG2dX*_cvf z%FJIy3_TBB*3T$Tjo92*s618)b11YCMURXAPNjT~B{ z4%Z%&vS2O=bgFPu66&btrj1oMm3h zIEqc6Iy--b46i!Y`nv~ozFqH06dgW{_RQ{Vn9?vm7~nIy;|_J-EzD0&%xDi~ zH@IeM+3J5Pn^i@|H>sLL^qj^0+h8xLB{?Z~_+Ce9=7f^1tU$cl^G(uTFU>&nl2NI7 zGuE1w7@d>dk2PF=n(yH#*UMJ= zAd{B{RV6RUXq4PxZwVLBQm~V-wwh$aMCxk5N^`VM9)IvDCP^V4MdmPKYyYVA4d@l&JHzfqInUn4kbthkaHa`gS zTXaFtpR|a%2$#aQ1m7g|F#Z%#v8=HxE22U2*=Q682_7UY!($|b+E?zEbSH5k<3)kO z0YUFL7+Bs_g>#RxA3WG!Zjz~TuNX(-H)yGZcPgcv)VLm>t1k=a+lWBBy1UcW_1wrM z#b)O{J{cNaZ!KyLqPFN(-hV4w>DM}Q+};xZDCBeOCuR%pKsqURJSxd+RdE-r0zRk8 zpM`v|QP1FBNseiZa9g>2B}Q zk`n7{(0QHGT$gL+Vp9NnXvMOqY8QZ5BpOw`*LzAFHp%vh$aLT-6&UA2r9OyozV&0QsgAiULJ+tq#15W(qF}Hi0TX`MTg@X~#3R>@%pH5Y9I@OBO%$xEsInR1xh9M#bF!oW+;yV=)GW z7&91MV!;5iEh<92=W4I4i)U(UrpTh8luAOqDmj;*d()ds+XV;DJWtA84y24Xb_mgo zFbO_FawcaJvSr7YgyD)W6%`cI|G8$P^<}6l><=*sekij-#?oVow1B&2sj2HDO?y6x z0Ctmv5F=s@T@le*(P+`tTR^w_ttBJ{8A0)Q!T$hx>-#-DMBC-e94Yjr7W8BYI?7!i z-4?VQ$Um&3^n$cC_8HRa+e^7Fg$m>Q)|%)!!k`SVf7B*j7Zzxe>-!;dfdAV%3xr60 zNCyd&#-#@`=&3$4+}ORq1sUaUaNoTQ>_6-Nf0GMp6W@Z0rGM zf)B}$r0=}tkSkKm2&{a1oW3?fLVo)`pHDiUUOz?nJ`D}q1xFEB*J^a$|DkgwvLX7w z&PB5~Kx_g}Y;seQ{%KRHk60kMf!lk3dX#Vpw3GV8#_KLI`8+RgE$$kWspQL!!TYvs zU-9tY-}Ka*q6gV3G-^2)1n<{b4qO3ZX;^$tLD|Q)E5aq0UCzBqj%D{7Ce?)Ud*t~x z(-&ODmd$lm2jE4hSL-mkrn6e=b$SKqV>7khcfy&Xqjx@SI8N>9QHoV^JL9oiD0K=4 z)rW-vK@V?@Z8kNTYMEM_eQya(!kVnSG~HwvwS!1-)O?pMQ&au>56DFGG55|_SoruH zCd;~X$&ZZ=x9?NdmzkY#{5|fY)L)L!v$Y_Qh_)zH3k7%};<1}A>Nr;@)P9OktCWq- zI=v1->`6!>eod5F#QYX_cF-cB)ovh+SP59Jc$!>@NNqHgti;wyDG1 z1M7oWJVJKvr^27$&cp~ca18|}h3H3BbOB%AvBuwBDzYT>OHC)#%>n-YTFvH5wT{@E zW+Bm;!|ElWQ6m7fy3h`XzmAHEEbycPOe}>kw($^z@N;X-4WO_N2fB1L%2-Rft(Pcb zy<^p>w#FD-vQr9~LXW{w&^o?7--QA)I|_@61aU-&+7L|dn<}+0G5%QtJz#@o#ICEa zJ8d>fVPww2iOy_w>RJ2s>unFzk@`Hfff1MQs*vT6NgQ;1&x=7r<(kQ7z#P@VQ_V>t zg_<}HWMBojwVSu@?(W*`)(@M_6!XDoot6)8Ry-~;K;_&_1O^hCq_Rx4+D}wkS~duV zb(cJkEQXNFzrF^cG$SsK5gCqeJI;xQW~f}fYH(;YIa+dlpUr}jii%4guNd5yLcrqz z*>8FX@aqHpeVJU|tqw_L3k#ae63;O2p_;GyOtZfd(5X};rrJRlVuaC)g|>B@OOatR zmBT3^@1>=rc#`q!@$vEU5}12#h_=(rKfADTx!L({QjTo79JO8s%4GjkjRtdBJ0d(f zIjqI|AbELt`TKW44f|SzdONhN_oib%(G?ZqM8m@1b}~Ic1fJREECCa1NO3Ii_pa%P zr$WDto4n^M&BO!?>^Zp8%v3I0NnUg$JUoow=fbS1a^saGm=41sK_NYar!32gR==Yq zGKT2|6~RhjZ2`1(HfGDsPOyYz3vG_}RmC6dNu3aX_oAN;oS&bsJ3P!!^nNf6tW?Y5 zcCY(XDD~IwtQYQ5ixUCPZR;;-FBKOTF1zWAvg6?y+1*r?~X1fA%YUcOAxe$(h0M;8=DEhrakP&b|Z_rw>GmmEUcJx#46|c3eU`6d30Bs~`A$S=PI$$ChvzRVh@%`>RDBCc1eYo1_a{2SwxQK*LV$~_4edOYzf##&+pw6yUhD=!j zIJLkvzj&-$;_q_CEg_xaF5G9>=Gbi}Rqj}vGXMHn)OqfVnv7`Y6+@pj`aA|JG#VPN zU*fR`zN*sdXr1Y)%DmyZ>0jT4(5ScQa}EfG^=@$Q>6hhyvLJxYl1 z=dmPMzgf?_;MW7ODL!Ka_G4fbq(;1Ai*a4g8##F?Zh^9)rXp{$9i*JUf*CUBL)VG! z+nS3!-~Hg+o-M*$Bg;R{`U5}$(PDZ1wb8slE?;h9hZK*o%CeON3#6?A=o|%<%G`kL~1=jg@LQd>G4%n?{ zhI1Q3l6Ttx2>c4ay1QyJ$B=wIA%{!5NvFlJiK~?vn(=Y7xmnNm^U^E>2NH1* z!E$L2jFY@9q=tW^$RkBeM>B$bB}1RZi%4YmXKW1~B7=fR$m3GSNsr-+L*Bjrt0Qa2 zP@;b!LBdDXb&q9?vqiUTIb8;Y+hj5$rIL-S@BR8d>*54qMs#t~a;`Gsh<2vzXszuU z03jPK=6={@a$k;rej{DIj%zYjdLrN5zr0XiaEE!9vuoR#@if{yXYi+Ttp*jo1&{w% zfuyOqXwsIksmN5`+Upf`fKj-8oa>Xk(p(z}4GXJOuI?OHWrXh;P+pZty}vr%Tdhx- z7)gwek6a9qAmD8{nby>3E3Bf|&)slc^9thC>T;1yIp!;1bd~_k?bEqkZWgJ`bs;P? zC%GXhcyycR*K<4}33h*2^Bym;vqOS|+08B3a*=V_A%b3Sqfhj&V0!*ye>tR>x|^|E z+BKR@oHJW5x0DJQw442E`^?m4tCX>=nHr`ilEs!R2bwCA%2fP#tT=&olF#XH16IAz ze`^7NKNp^8ge0_bHycc;0x35gIOC=&LHua;Dj%1g*$n8*l%cJsD7qf#y`USuXzNZu zvo9HvORPPQ=c8B+KWWh-e>OF>)#*a|Bf%+5S*cY^&&CGT@^|W?MYx&KW~~(}j0V3~ z_G_c0BxgcamQ5giF?sg~C#dD;yG@QwlwW4W@|hJ1xnwF>1sE!jZ=CWAZvX0+-|@>iQWwsimzXtu&v_J;_UwU^`wB^ zORqT+GuJEk>+d3xD}yCbdE9u6nCMKNCJBiF1ALF0`8Wh1>E^z}Freu}hi;REzob4K z(p=DFln+FEGTw)YT`u7?G@SkPWt7|)1%uuKE3yO@{gIy;0iP?BSNN!h9O=qJqjG(+ z#@=X)@C?;G5L^;hp1UUui#~JPiXXh}e&TJtNxpKmbqPO;(BYOwJB zb79Wfrp~BuGM_(gtU{Ik!;Jihd3_z#>rxZ@moKFCVZvX*A@sB=!UD{(-nS1T$^2ujaY>6R+7zcyGgbU=@Oqka7!d* zWT8H6kQ`^nFkW}t3V2OJ zVWHiv3JP2c7y+1tk)LA2qQivHQSeVs2S6Bhv$}-7qKhDoY2Hh%-uZwi5^{audjyad zfwnn}{@X|zQRA#k`nmoMSlzFn(8o)aXR8TvgZRKnGosZXs&}~DDa$AA0}c0)mLoDO z0v0hV<+;P@@+W55^Bp1~eFlF#4cOU+-#PS)u*Dr<;_flcaqzpJK^D|$dH+nQ{>?=R z{TxNG(C8>JFseK8nPhHAE=yCMMf}48}<#(V25q8tppfXYcwZN~c^)8uj?j$00AS{`CKR}$Kz!)FQ>3dPO zZ;G}z4%7(uh1VURWRmaBFP`v2#79OtVYCB@&B1mhJW2^@GMToUY)b!BUxCR1%&0hI_N$976(k61kcMBcs|huO-)ailfYL2 zv~yFd>~fI2VKFi{(LXJ4 z-?+?#Z<~n(zz{eAukA|1Z1{2;`C8Wcw60$9IuKEvw{E)x^e^8+nIF&%(-|+7y@`n&4V)v0R2eA{o%y;%zGAbF}jY)-=zJtKQD&-1e`mWJ3J|P~D-NxUH>2`YuHQ)B74n6pc zuVrk-B@`+QGS$qkb_Y#VEUe_TH1kQ@8k?ZNKojGFwlm_agam6=K3T|+K#~#yb>-hr zxU_0zPQQf%kY!(LE16H&1Ul2z-dgI=KYs#!8*@ufBrzJ^{+VYzG({3kitl18ea7m$ z2`Ilk9P<(%Ni{Sg)b;mUzC~Ykp13}d6kB^T|7zte)VVza0A5f=ycD8vs1mwUrhC_8 z9$m#VzUJ3X_am*@!OQP@%|%=yY`phG8VliZQ+~Hph&YYqZyOFq{y#F4QDBOo7`+NY z7z3Q%Nl@B*wp-8(9Cs-7ael5^p;#ce^&*XJ5>xH!_O26A)KOTxl`HxtLtz`g>;Fxe z?O51^jYaX@o@K^{QhD0}=Xkc(YMFj)@>g0ZjFZCZEijVmS9?p}K7Lp>H#0N)^r?GY z1m>S?+!qQ8aPmQg&_un}Qr>FPk?JxI^`NIDDf_5^}^|wC%T-($Y z1o-~{smA>Oti0-#v9Ylc6#PQ+3@d5z?i)4PFLK-e$cdy*f4%|V#32sgLqDY9;*#d( z`ah79l*awe3uTUJ_xq2SsE-c$W%7s#zn1w`P*e~Y6a;q`^X30m`~$_zl8}^?7MDdt zh)zC80sa3dI@<6L{80SJ8pSXw4IhsUUW(P{@4of=KY}71(5>z^$&;7{h9K5HVeQLtGt{(1jram9sZ zK*v#v-F0YBXH^GV-c@W;b;woSoc?%nEvapoaOcd5XNKjojQQtkp7(vfj>Uh{sn_rk zSNI+__G36>yIP0}k=`$_Ck70UTBjxXSarN4;IQ^VSh(79C>!DGo++hL*u7mSU&sv#hVny&zX_-G-5TT=>oO2aWwwsmd>m8d&7U??;}8e$FZGGEZ%SUiL#Szxx7#~Qp+8^Qu&ac#$9)gKRAqB4xWGj43M!>vdw4BWWM zv9D`UxySZ0)~zVNqo@3M2+vu~9rwZJV;wcT_3=}cQ-DqPkvC>*`lruyX-x>0Hb-fC zgK}XVTW*J6?81|n^S zYaP}Y`mTih(D-~{_>UQbUl433K>Uy0zF4d?hqrDPrd7)>W<8>LTU`XVE%2!RO?g)W zy_=KK$|n&ZifFWhcApaeIOf`Fzq0UR9k>}hnD}`N-z3c~t`Zhw z*Pgc_q5Od3dPvDH%n7;mp1&I_yIb}??JI+*EL{IY48!N7P}l)8c+cH_v80>%lUZ2R z`V(9#v#ajh8&o-;$c)*;Tl%p>1)ANxE7WdFgL{URn)E+;ZnVvL7#}IPFqtfRk6dXZ z-(gW^|IF%#1M-&#O+ zExu}-c9NwY_CNb*v>g&N%T%kHXIIX91`wIEF;WUtNZebVji`SW+fwr#*6wTKBPBW z`_PYGaYzV6!w_t4cJrFL{tx!v`m2pE>KmoB6lIgut7A-}^lGZ@6pSAF^gvGMO{8&pvy9&Q|{~Nt;6(`f=C# zYK4pXF|h-s`;#1*H!iiCb`8siM;j~JeD*&c)a^bTf_L^xoSBdd#kl}0(|B~eTKAF$ z<~rAVCH;&pBbu6|ts-Zr$fNE3IjaY8BDu86LPbZqOvMTBO=(rT>m48i{6mClRdyOW zwt)2p%ON-7);18cyHCJ-4(>izIwL@IT*ryI40rGT>8~R8O0wChMJ81U^}zLinuh-^ z;cI@u3a|8H4&|ZxhZ3L${pFPLAN3o7HWy^IHQjOz3@#!hIceF$j?Ws+3Uf7-Tezs< z648Ifmd?lEf*LWgM&Ub>NgDy{bUar zsoYg&%&qC}RS2c6Nog5%R*ZSiNh>1CtR14bl<;{368~8-^k$xb+)}%8+}duL5v%m2 z*h+-m?lLiGb&eg-n~^Y;$t6OuY!VAiO$JFz(t5-4^Fi-!I=n^Cl=TD@#dc~ot$V`F z3Y(ykEvdi>1jvs*KzNXK{nC`?wBwzP!NVgnDjqa#^p3D(jubx+xc1DLzS8D}#W z%a1UExi3lMYNky7Q}*otAl;nprwC441e{cC-;gTKcbcXSwA5|*RBf8H_%DZBFDK9~ zVNcG*rrDWHFwxQSrIK;_-!Hv?*~bAIr^O5qi>adJFQl@O531put4VgD=U8K>Q~CxK z9KdrD`_PgD-J(jXQU5g6P&Xj$CUg8O*OMVeeBqF8_ER00@0TQuO1xDv@`+vXQ~ukl ze0VnXlOpcjtXs+?vxl0Gjn1(SWf0g(HDkS(znX4Sk7c@gPd6}|de9(Z_a)+JxEK_Y z(YW`u{2;I6Jw?mQCrC)D$7O1bKH-Kw5A8}u)h}(wbYEWXXgMlxb7L7u@0m+_&^}Y{ zYH0*_4HLPrp(;?kezq9W-LMQfy8 zD|cnl&&~Hd%d`sP15#!op(_(8zeyi*CBiIq*|g!gqa*t@0qn-Mf-ArcX`{Fd2&u2U z_ZQD+T0(sJ-mjU&m1yWlvv%~Oouy>?2S?E-(D*@|*I_5~oA!3T^{3Q9&R*gtMrGic zDp>8zP{$#sxO=M$|6Yon3D1*-!dNj`#y=V+qualagG+b35Wsv%St^n1N znj)5Q*C9PEzvAw@tmcux1omoJ?IFcKxcY?-(}U!A$`=~XPotd;tyoMX59YdEpGvZA zZa@i+qx_^?$GS*0kwC&QK=0Q)_T0_lMIM#GOmP{Q4=LUKLa{9pTsbvoKBOd^M}9Nq z#`;>%mdc-h%Z){OoI`goTXg|M9se45DdeXftcO`*q_QBildAJF*_=iWLo4P{{w}aZ zRh{aP$YnbLuC|_D;`eA?*QWFAjm;m`5>eGsolYGoOk@UXB0mZj2a`I+uCC^@5;^px zV}wQ#XoXI<&r)cd!$opNS!De9IC+@p8ZolUC>C-iZV##MU|tzR88i;D6SaJZmmV)H zad9o2jwcZgxuId>x@OAVLz93&UBYqpR2##es5xVI&95F4^i1ZV37Hz+^I)u5NA3s~edvy+q z$f-<%OVXycysJaXOr8+lzg<=nB0-pIcnl9dzr$bHk1{X)pb(R;u%J{gmZ$nJ1YkW1 z(^y=tp$zjMv)|UTu7ZkWH##L*i>~%r1`E|*FX^r<(CX4 zI_2M8bOpUXV|D|&DP=fI^eZiF*&(+`8TNvDivFlW-)gS=NS_$_JWB^{bz1GaVQMSk zv&eYtqFv=;lExgOgL*&~vf_Um29!vM8qFFkm$!k|_7h1G!XFNH4c%Ndv~k z#xY~-{PuWM$T0<3jVLp>AR)rvr^FptE##ye-u((|IWmZH3dy|!1hnF-Va9G@vi4G? zvRR2A$j8ceTp+-3IKp8;=I3{117c+D`0?d!W0tsZ3$GnZ#O}zdZs4c?ATd8CayFN} zIWg$#Rkd+Q#K36<`Vs017@28!8XHj&HTItmn1ofQR=67E$gJ+M`pF?)t6r%Otry5)e>?gCjnh5i< zn+FN(K-teat=8|>>T*I0PyefAz>6ttsYxmLK&|&6hrz%FF!0e#@131Dp{kO61P%`U zIQ}?sLgi$w&(2KsCz0?v$^-(x)V^;vxLnmMy0w1%hE_Qj>Mk%Ya?brB4;19u%l(3< z`7k&>NlgtTeL;X0`@$3w8X+x{m}%n|Tg=4AXMI7&yIEWkR2ZnU$&By@Nso6yZY0b{ z!E;8#u6%QK_1AOh8xP2%`LQFtL_WbInxNENO*{>z@=MYmlS*+7*RuHc%^g3ChuDy2 zGfH*Ie5IJZ1I^LO)1n8+UJ2#u=<$GLc>zJw0@&Afld?&*CdNCU3PS6#N zo|ntw#?Xs2C_JvxL`FnNY*K5++Z(A>UgDV~n@_sh=upnyaKv1=6#fKS!u0wHt{!RR zVNlP3)Ei;t90uUY)!E;*Iv?GozS9>WDO}elX#X#3$MepC_L{-Voy>6p00NdYZX%=B zHcioDdyZ0d4nK_0dn$XoR>vWrT0%QLVynZngISUsAk~ilbn1ID=S}SKJF5-qn%-{K z>UxuaBS1UBvgvn_O~Um374eF$%g?E%;m^zZo~B@RdF2lb)WtYBZy%V>E&Sp28(}Mu z>3Ca>FslCi-q;$038dNm#*GUH%O{hlDF!B;S3cC6w7c;)S&aoZb2P%o@<`$sZXaaG-|NUINqLP5d>3eL{UNhiM1or7>GMP*jGEr1j7Wii=5LHD z_Fu+}(U>sOGI?wR3XPhe^=zHwS>_v%9aa6k{?)XNX|$_p zB)ikSYY``h4VRT1M`%_19Qz#aN~Ed6(St>g!R}&?kVF>`v@owH`dWND{kNgn7|qs| zt!(jjMT-r~5|PQl*E^7O9vyO^?LPtaa2=fNU>ck^UuY z-6!ef^{x5A2|7yXfL_1U<#<>++l>4Ne}yBU=ZhFEA-kC%8!)RWXcX_~=FSyRxTkH( zGkVsw7$=fEcw|`pcJuD&Az&_EyT;NZLxG-T-ST_qjOg|Jzf1j8)+4bN%_p-2iPQ#kX(0ruvI^%zV0n0~PTR@MB zT}@Z%TX94K7sUJdC2!ncd#{tDPt9(JrydZr+(U(G&PWQH74!Z`mvB1vg|H=16dk~I z!pcD!ydI4sejT|V^SY;JvCg;6I6z-Yr031ovzdnNE0=FL*KyKL{kvP_82S4TVGs?h zYzFrG_N|ez{$~|?f&o(Tj>=NYQf6F{2#q=Y(!UMWiZ%943{Slju!(-?dqUa$;)jn-+v`R04sf-J?M*XLT1xe~#^%ziW@n zzT^W&BWV>J9P6eYi9WFr8T~eZS^Z~G!q2B%^sO^kgYcsTT5qv&Z=crAFQ#VahS;{2 zmE>cjcI0WpK4VB|e5t2e()H^$M?ka`0RfMoqw8iCu|48uSvgok#Trdtb459KvHK0v zJ8ptD728<2mKq7X2x%63^o`0#KsbBMUzP|7^pgc`Dsd2qAah?&uWw?yCA&818)02O z4$71c5JKHrGS3Ss1=vgg#yPCwKf?oC5*HB^dP+%Z&erM>{|3pc5`VqXcEnnSZ<8@d zf0F)Chf*K-@xvMSYE zO5TBW`vm`B5}po5VL8TU zH}!h!tfQp!D9jl}gkHZH{pkL4O{V3#p(US?Gd*^x0cGu<-S*O}bCO`UqI}LDTEF<+ zZQL$Qs-*P02jR4OnG_2x`jg0V_1Tau2BF4v?M5dffa!moBmEWT)u=}SY+?6p0~eiE z>Tdxw-tY}wRsc8;gGg!eEI>d_53H(N~cp1-q5XI#+*v1Xo zsOy+dpL!o@3k>4vdCmG>7+5EP>@=6$-9W-)aTpLTM!G=T+ zD*YsYjXuMMbiPTUqt$*!lfG7^3c^!rbTZkCk)J;_->D;M-_dRSEO3V1MDXd10K1V8 zThmEtcNbMTb*xwS+E(kky@W+)+&EOdcJCJcIFO;EUnXELor~bCk6v_kYjq$#k${{cBYIj;Y97qQD&hpqiL>&?uUnx!tWG& zDjM7Tk2exjtX3(EYZb@b3pSrgMOpH*m+U}Ptk!Zu+-Pp)iGeKNt7<=cB9(& z!-3ec8`6VOev~QUyfS>J@86?&i(cN3#aY~{8O2eF@V~5W{6k#7?YKj`lzcw^nY>U< zRrdaCVI7Wv=RO6s!|>>^WY_&I#TB`IS$ddTgH<4RB5r!E<=x`WK3C7^QcK$`T|wW@ ztlH|xs^~IbqyBD-Rob2u-<;g_P^L5EhRbM=CK}O~+<_=YY2dM2JUtF_N-Dj*_1Hb` z$=GgaK{GEsO_7W#SWoi_6ih4ue|`SULZk@!`@=#kaDyLbfB0i*btW}h^Gziuh#qnD z@bE?1dR@#X0xk3|g9lwaopD+4PL@$Q?>*~v)^MBoe$JY(loKL)4`p5NTaR@ZV-ehc z$mq9x9XVi}w%MN_QAaLd<8}NXvm#uSL0^g!y?m~3bR4# z)|tx92AoX2N;wsuYH0kQDdgt9LZsaDwQeeK1`&nyVQeK={~tqJV>Q{_yR#RgA)ke) z?g2CaG<9tB9w9EY?)Bm}q1qIBwR!CV5{7Skdn8wW-3s`4w=^Sb0;l6SC#S2@psAf> zXE`J2ubtT(#>t#@i`IP=VHc2!W@L1zd$M3F%DAk=6=(+fw-Pp)|Cv3=q3DvypGEj5 z-B$pwuY}aa4!NUv9Q9(0K=%;f-2*4xrWX^r__Zt%(U?_aXc747E~0_EXd>{P91h%> zZl%U*Z9mNPjdr1yaaIF@;u}t6e*$WWk9#TDA6RQ;zyHU>&iYc_N@AR^@yV?Bn?=xr z9oJzW?z;!2$N3NP+*U^?`lfz$(Ajx{LNAff_ocdaWjKZ&B}&CuA8`#`jdy=sms=i1 zsV$URzDXvh<-~qhN)+u3QH)xlsKvaz@9n^n3`=IkzLo#%oyk_0PWK3o)%&^sPbL$O z+1R>G3}I8%xT}{o?DB8bIw}l74R#smwf`{jc&C4srhibosTKPdX z`n(7mMTf6;vCx|`dU9v*4M=nZ%P!K5XJMNNhP1tweMCmVdW+`X{R{9XtRL8P{tkKC zqbRRt#$231&`R^6?=v_n0Z$|t*8hPuGR_m{hHzyzopGt6jz9{A63%*P76npYa+>JMHKOMM;CiIPxC(Flb{|Oxf>??^PU_{MZ};r zrPmn&dkkXq#t@8&#k?*}zsE(bJTJHmu{d$uY)bsO@15DX-^S!{`Em(& zZpBiK?)}Y5$8*|a1~dE=VyNvie;ZMlWILP30zkXxT7_+^z=P&~n=HtQ$VSxv4^qy= ziJ}K9mtDU0!r6w^*i)8I)a6Hi36AG>K@xH88g`~T`!)&Rg`8OSGB7&&*TJ~gmh7eTXvuXQu92rna*q=w5aM9kfN2*w`8gqoK zeIr+F@-y-hC8viAA*p-e>scIj3(+`}@bdq_whvmMTf@<#t&U-lIUSZNH zn53m`!g|a2_E?GjvbxultvtKSX)o0-dFsyVm^j@{RVY1eQEf}d@C_l$^*EHb@vMf8 zdxfzUt?nkvVyZv=VrBrIR&*~k`h9|XPoPE#I`3a!?IFhNfrbfe0#JYO+?p%6OK#9M zacwswqjrQ~vL~tD=1MFLJcd5U$$Z?=6DC(6!xy0U#<P7RK9ypbLY?U)u1L7ualeBlcj3!S)nuR)T5PKU9t7#1L4E0(Q(%*MC zFHCz@IV?V?wp57qt17Jt0K3F`YP5N`)?xGbEC7Q%2}%B8iqJ=xismF$ALpSMENEnv znjqQAEZemf^{~7{v`RlY-_-)Q`u&$zM^dKR?-vRrlibarmz8UCJeB*HU33DT*x*=8 zPCHQZ?aktCbCWpk{Nwo(YVE0N7z(m_bKMb1k3_=4 zg9@bh?kA%nT5tNwsLi1o@8vnFE2)6!hhC&*)`$DzRFyqpUz~So&BKpJQMYzn+xZKK zWXnV(upz}KkQw%AU$}EMIlW*VzjB8D*hKQ~14XV!m$Ryzeo_NGVYdH6qwuXAC*IIN zSg8s0CG+r9rbEG1j-T~vjM2?ar2M0Kk3$BNB|ddUNo73TwDR1inBUxQxx#Mcdf&OW zB9pk8-M9X@d8uDcMp>Mr#-)1xxC0`t>c(Xnaf&3`O!dC%I*jDr?|fc~U|tXt-ho%~ z)p@HwFgtQ99n9y!n3Kt;44pU%PtcBgs0q!^U3JYdKPsz?+Bv39^>^d4Z?ZJ+{CPa+ zLVh_7PiTs_NC2+}7xzvUlEG{#_2?;2=NYe2WZb~l7SmG7!|aqFMS4%C8lwC-Dy=R_ z6Q>}1-xNZdrp_)NKPdK^H)UegkJ$fAY}!ovJO;RsXN8CxmOhUY=m4(+XXoPZG;eT*+NJu-IobUm@)`%$Kq;Z>0mJ zeG+wj9rR~wk3##|v$H+;yMj>ZCWW|NXcO9>B^_bL&-O}m7^8!{d6*qr+I9-!=ATMu zekw`p^>-hlih&p3P3_c3zAfCa@YM6XK@9IZ1wOF;lZu3*IWw%L48)(A&}7#(&j!k1ddX zbdU8l4#(C3Q}uowW1q2ZU+Qi71C}71bv-2@;POE!y)~dF+{>M%ogspL+}Jt(yRg_a zj7?j@CI<}G-QbV=sUh~{0qrvj-j`MJPel1PzCD3ptrBEn2T3aAT>*0PsskFG6=pbH zhtFLxZYh})qlH0kJhRZOQeT-gBNt~F^r_D!PKGDPZE+;Iv6$lH0HUc#^>FkhV@Vhx#e_5c{RT2~TckG0vx{noR@ZjH@qBx-@1@K~Yz3p-QvO1flQ}(QzT$WdvQyMGJq3Ql zT+S{fV=W|bs)2?=IsOfal9ooN=OYJ;tw6x=7zAP5Kg(#kq2ZE9Zk)Rw4X1V)_o)`u3lKDqElNZ}zF8XIR5W^&nl(B0YPGc`2A=jB>esqAaDHBvJk8?%xO?|mS2;OI(+=d;nmOf z^jPVk4Z(6E#kF1zVAOM%jbf3Mb~12HmR;y%>|?XlI-eUY3fK4yd6szc<#$*THPM%& zup|StDS<#M@345YOugF+)WP*k0P1E;p zTS$0fr9+|aPmnzB1L-^PQK^nM_scUUBf)YE2saAE2=`-~rYx%@+5fQx>@HbovY0g@ zk8s#~-ZbUuJ3ojT%=E(7W)-2dF^p|l&{dFsWM-lHih}iX*Tm2U$ClK|ye-$$mLq-l zf&T=L2qqd^tdQ!SLQ~uiv?Ga*S1qf9S-^5>q5+C~5gGg%l7>85EhY^u7D_(V!R+7E{iq5% z)1Z-l&)3hyqeUD~EK+EC1Z!3ko?CdpKlQ1Xn$w*j%6l#@Z|8VzldZlAY(nYAW$KCT%0`R~bXi>TjJ@tp1*6F|zVN$v7 zDKC%ghM&iCl9)G`LLD^|&n zY>X$8Yc{Ay@l*&1Lw>v;S_}I=q>~@R5WenyFqaxK(Dsp#^0oDA;q%1(XTjJXS4vZ8 z@o!Z=M7(m@nVng>vh=k zvm>ZpevlKqAnbNb(k@D$qvbI(kK!R~v0sccTDuuW=I%2WG1W2Vyq?OiszX(g;eX<) z@Ow74R!Sats2y}DS~8L4!I0SzY4EDH(aL6y#~>3BHP4p;>29^Wy=LC(dTPpHEOD2f zjWAw?yuO&N@6|qAa=B=1RF2oWdl<=iAR)^L|8#R>A500ZdUE2dDUvH}skzLysucYs zW2k%(w?vUht*dE9`1dxY#U9&?c~d$(WsZ4aE;Y2$AIupv0O@p#)v_AQ6?2O?vI2!( z6@5;fpfV))gus*Y8=SU)5dReMdMqp~YR>OJZNKElqaeI~Efh0^fTi~fx9)NJitksF zfpmAh&T`wsx1TO0raIbP{wxHxIgCa@CS=i9KD=$Svi6(OqZDXoS>~QDUOh6>V}g3i z7?B9OxXRWdv#e-ET3U?1`9jM{RLa}#@F@Chkd_4*TjdE9oNDf~l6h_N;MmQ`q}Hxs z_T=|2%Z~?kw`31dcY$Spu0s_v;8DF?d#_9nLRk#RD9{p9H(T42CIwdgc8G#8By{7$ zPJAkh?OqMlgR9+t=5E7?`Eo-|ohRGg281T#wN`b`-kPpU?RmPgLX_YCT6nYb?PDuU zH^VsGs}wG)h?*8fMln=i(|gYkX*AtWw|nV0YkTj$ZG{=4oX5Fo$*iZk-3&WjvU+Mf z=Ntlza>JT}eza8OijLiOfxL#^tyB9VxuzMor4P=wQ%@^`@S8U(fB@-9blbZ6XTPEg zA>88~BoHiU=$c`J;)M;1?uf193XqGK9)>nJLe%n7B9`RXwa{vXHu84e&i3YOY=lch zGjUbHIJz03T6Aoe`^@=Waw5%fskhrd%~Gg~0kEbrk)Xrfhqy82d=#ynpiqgaUuWA7 zqgkIQTMFh;|3GVr&s%;wA=o;chA_o4!ADN zb^5(s26A*SsZy0%2YPL%ja<0{70w6hMH|*GxzLIk(@VX&W-rdMY%U;nyX*<8hg=ei zd}(5RWd&5c-?6a1 zed!QZv0zibH8Bj&)@_7x!O|jf=|+iVnHh3HicD*D{PtF0UvC z4@B5JqiRwW3#>|O{D>_a`il-$TeR{M<~*NQN`S3)8g3j~E2;GlbEC)StNdM?g+U)% z`2s7iFV<;N{4??%-#k%q>^a<R<`RT0>0ub?)jlHA~d$3pkTe7SJf#XM|WZJE5s5Lulzb8XLL zgK?ZReQ_b%-KMX>QUmAjw>|p7OD3T;=37os?Rg${am+skcgco-+_$SlCuiD?u z!qTY8I{@Kyu|aoHglR1TIiWG8O!nddqLk40lhK=z4#Qh@0v!jog^gBod?&RO0o^Hy ze?5^2@L3WM!q3AwBI|dG0*iWw2S!XUqKqM>$>8{^hx4LBCnt67&7pb4Zx$@1f*Zv97_=*ko*y>HC4IcL?nxPr z##dNMsNGTlnh@>caz2(1a93qrQ-rsa@CjL~>pS$Oh|Iq?Vnho(6N=t<6`sgN7tcwJ zheF%5eD{{@u@@D3vTYu7MsGjOp|lj`TWg*X{>;f&jY(0SS439!Ci2h>v$e8%UeG?J z+0Ql*kYZn?QDj0Hn+fX)ni>Qp9-9{&j5QyO9ur-zi*|Yd(a0XvJ38-$^D07M;SvBC z7ToM#Oz%^62@=jMEm6-!zz!=L=3T8R^lF(S*6{4*l*Y8{v8$ZmwfHx;VpdA zE96-`G2aKjzpc9l(w+t6Kf6u(G37wk{odwHbHwBIrSp}`e!N05+NLXjaVez%i^!3t zW!lkE!mAAWYC#cdt=25ES9&{nJSI90bQJT3>L~pfo~OgrHG-1+p(2e2SI0%Gxx56! z!|zhLj63F>2@VUq&@(9>#L7HCBjPVB;X#jXZPz&q&yX!&phGoxM)w5YRQox9_|s+4 zO22G`N*P&|a-(+MJix;0hQn*kbxlkb)5Oh8P1_z(dsN=29E>&yHS=vQ3JV>^K_=rc z)p|y^o5MaAjY$R5Ek3(W87xreytvjxqtohY&7iv(6ps4(23J(3uQ+Q!x}4=+gUqW} zXLGy!tz=2&(QaI0n5Z!dHE=WL4g`}-B`Msy05hz>*B+S++3SHxfa+4pZq3VS3N zKqDd`Sq${V_`{YE%3=;>(c^%O3+C$BRt*!K(qag#K^^KZP$uacmiHcEe;HOU==Z+K zsZnNBu0?5J#jkhH!#JlI#px#T zj@IJUWYfAOKYKLqX5xex7S@NvN0c7f%g|lQE5LT}THWOcOq@1>DS&@IxBTUj?@UDF zhm&&@L%73r;{eG}{9l+!bc|vcCcNrfw<;LB#^Sp`|VG zhQi{;1L%e`e3As0{g0hkwTnQA7grjT>MO_ZPa9e8hg+D}k@1l)_=Hb8l3lZd(E+_;~ zdrZ*M_1)1@Uh4Sd5=SAm_V7!uzS#t1_;_KwjeQM7SJJtHGp5gUQSqa>2za~a#C{-= zPnJ&m=c!NL=NUH6Z`+43>bCUAtH-csVa8+`8w*RB`9V~U0%;QNsLYJP=&`vT4KcF6$CjBWGrt6up75(R)pGL>8ypv%?mDQL0Ca&? zk5@V_D_c9*lr${c?yeO>f9#=m-6*6bvoM?&kc`zgZh;i6_Eij+K#hA7wk+Do67D9$GsE4JQswuFV~U?MpYGi7SEf6XBXN zk(LG!D}c1!oE`XNjS0HKq5!L%+VUc{C1+BSy5aZM!V4=_l$a+C<}8{>bMcn1l$C{d zePDj(LJ>u0v!GNlG=UQKj{Yq(;jL=)A>_Vuz2;bGo>e4HCC4sJpzd6d4Q;FE+sn%T zqqdHwR!_-!uh4Ierq-YFeD8+Z@ayzc{@UC5{W1nVL?q)80D%x(G9ac%oX|8(n165X zR&$Zu9Wo7ZU~HTSbLMzSG zN%dtN;cIo$$D+vh1^}G>_R^m#&2zag;Dk1-2=S)m5PEmQU@4fDKYMojMUGw$U^G#Q zFRyGbCHsXjD?P_~P_yK3*XneWwl}(2hr_O{fNM+Q}+*CHtL0WbY zZ2)|HZ0ie~u$G}zu#)q{lwW@_c#SOS{h%FPs5E_;LxGy4;EE{ogl|5xCgP`Yz(@EJ zLdGkXruHtq-Cs`jZ|^auOEjCglE#OhPpALK79c5k-tCv~3*#N0J2uWE?LF0*Bra~D zh7PSDZWu22lTA`CMuO~1fx0Syw=P#bka~afTm~70{FujlV%+4qf4~i(6rp>)EI3^u<=_^QmC-Y3C}12 z+=|)H)%r#5Bq|*o9u`&CU_dop=unhw5@gs?_GJDTw$eaQSYX_1@v#|pT(jlj%r%D* z1etze<|`IQYduG4{4*giqW838W7CR;Rwi#3f^34rbfHhodMx`6Yz`k*CZWdVOV?k& zG4o1f&R%^OMV^ikV|DpgfZwZSVMmyE^?mjIecGi#@^Mv4o~f~wJH$Idh&H>r<F>U}gZ5P%ZhK$%51;9)-ApO-vtf@x(Q{_5@2^_E;;s(}S-B z!aT&FBf_8bX=qmBY7`{lOsfwKp|ic?M-il#I+?vwDZK~EM$oNdYrmYP0`Qowsfrzj zbO`^-_4Nh*V^JRhN;&(4lJQOf+Av;u3T8KjW3Qe|aW^Y_>%0c>!7ETf){(yrWO*<5 zJ?$Z_j!F20*m4twkcn}5u~*J3b$3?}F~T7Vqm~R~>Z7-StDc~6%EXRzX@@W+2s~}c zx?_>Q+K2ufm%O-$BOaLEVVfR9K?4Ixx`rk{K;&^?e@simk4J=UJKM3aG zPxQf`o&tf@PA)w5xh&dNOgDAR-Y{MKF+d_%w)&uY$c~}Vdh7y0msD7zI;!+XrDmheBO z6fXrh>xw-#JUAg3S{SVfI(qvFuHT`a#ml&u@rt8)vq(5_c1g}L4vwd=LEq^-R2*&V zF||g;L%*?N8*o}e0l9n|r{tWI0w>Qsb^chsp5ZT*&0uqtOYXRfhc^VYA1Y4hXckS2 z8CrWdwbN#WdJ{YLT|3MSo?|GbD@##BmjN$q5;;ELME^QG#WRvBa%$cJ2Y|y=W7R{- z^v?ak%%x#sv`VmJt!w}rEduVLjQEZ>uj3_x9h z*o)q(uPRDm>TQH(iMF}EaL9Q_+Or~DGVAPWjQ&;x{Sn*Br4etY*n?HhSd!f8{D4qWY5Y8&&xI$r1P$uOB=S`KtmaqA(2UZr0~v4teK|F>wCIe%EY|m71wf+htDyz_rhBqTOno|##_r@uu zE^>mcgLB(E1enQY zxkX`ypVWR3wfxC_gDq3UY^U=HxHg@mal4{lq6$-Fa8gI^MIv>&ND|aLV=;3pi%!+i zql^8B{nq8w`4Aqfv42t0`n6IOA3y=tO@pXrlchsST08_n|Ej)QF;$gBi zk@mG`=8kbU_Kx&QkvxJ)%=9i4Pa%QliS=A3u)yXy=;m`+V9jUDq}7cb8w#JUEWubH z(p6ZZ9$$@+KkW8lKK+$+R7wtJrgW7WrcRO7vJSO*lf>Qi&=Cop#$Xj!r|7j-SQ>1R zrG1I6TCLz=uU~~vn7?=HVfaG%J4XdL*Ir7-jS34Gr*!om-g~KnxcTG4X6q02H;=e$ zL3XAl9((0`np&}h3v&n8bUAkaU%Ds~wT@NM>5_jHshPwm#ne6Y!J*$(Rk=x!~jZib!)~rtnQ3pIHv#^D`V~)C$x%9AL_H3h->CN^3<0YA&H6abH$GRQCJc%@xqIYSuKAsX~7&XVm0~3u(UfQ zvB$={6)=P;?d^S4G$bMu)PSM*Po{t7jBkXvPse1P%uk>95kzPBc{RK_uJX*{t656B z^ngpXM~y5Y?H`+utYj3>p4z+fD@p0@)|61e#ydDIjzv4Zl%qUI#1EXr$%Q)zc$}Ks z9p^dS+gn(mfopYk^o%Q8Kp)S}QLmokpm>bkMt1Rem_t}ew2{o7zQ}FzgXml<@nF-p zbAM=el!q8h-tPRK@K_!fO<0k3@P+iWFEz;xC%&zobnWrMVEt|2THUi_42BUw4PdYO zEUZhzF4Yh1kEs8h|H6Ou{ofHw>*a&T|Biv!_^(8SX z*8h$l*#4i}NPT=n`R|MW{}xzS|DU79vu7Wkm1}TH+V*~XBejbhZcBNlXihoJfX8D0 zl05Fw!#jCy&duiqvdPbY`$l|rz~?00-;`fHvFVTRzZHh&Tg|fmH}%!Ak0@2EP2toOBlIp#7_6GGf{88y?l8?fzt|+f0Cl^i%W_FlCaD3N_kc! z_VSZVxCWr8bqx{KG#EySu@aN(>EDG<*;W@+dbAipNOV<0LNEW7T zKAvH-_2(i^ggdw*%>zs^+rA3Q`rD|0!OiOikqlZ?$rPHEpgJmxh}vk<$@vflfzF3M zONtsW8l#WVzSyeJ13GJZTAizowq)2QQ3v-sT<<)bm7mEhmCWkq{YQv%O8nQUPx|nF z%41v~n%#WxMOuM)h3iP20?*_WO2=1lmEz&#NG5*Hms=0Env@%y+R9%st;D1Bc#GU>?D5 zN#oi9JtseUIrrcCeN&f=#XPJ`Ij#J+)cFstVLF+-t)hmRTcFdPdt}hN;vhR)cddTg z`A7Z3KSpk_gOq&qCfUNZO-x=8pcxIJ+_Y-S887x~v9vmm?ii!3a#!g&j|F$*b4H^y z&Hq}h3%u}ZD5ne9yZtos!?g(P-zV84)|n&TQ-gOu9K6PeA;9-}vMrGmQmy{Y)3)sX zs1o)kRClu{8FcOp@VR;Flp^eg@h`B|PIoV~he`6EJ(f)P9I)A43HtjSH?-;l_Y3hC zzn%|S?2YggBQ%C-RGrQXd=JuT3P|s9OC}2PlUMV$t!yLi?WK>NVFvpJh$ncR`U3yC z24(b9VtR2AeO>Kujj!It!rx0xzW^V`>ysMfrtT5TWBL%%b$9;SP+-hBYiagt4)9H6 z|4Ebi1=qp+X2r$e+c)2pcG*@-I5;?9e@pgg!J@bf#qfS<8$h4hy}fTFd|6?w6VwdW zu@b2(=?K2xvZWh`aIMEENq7&^{^YBq^wQSp2{UV{b6u(b4jO0R{K?1gi!fp=%eAQD zM8>tqef~z8tdQ@kk%+zS;v~N7Zj@hv{1UbN#G|vg>Op!EZuVL;bbfoGdh(A?BaFMg zZ8kZ=w+jqo6BFiN^XPiwIR8`;;gcH*xqV%4bT2i49NR?u&g{8{6LLPzoFUFQ5+Ol61LgVZ75`|o5u zqE=alqJwtTu_@y+b{h84COPW5N-l_Kk+vu(IEWIyC6t0edZO zRy-Xd&Lh{TJQPw}C1bAWV2sPEEo@8jzO-yKT0`bG`u_DYg>>j@3(o$2!`_O>?eFUG zeIw7K^_>yoY?;$(ls>3R)FT)9=gbLM#wyeLPNsDtnjvhw&BsHqe6HuYdfQ(QWWi&9 zo&q|TN|Jy4YxxNy$sY7rk%h^cQ}C|3-U3|M$q#SU78aOpMah4L5u7;*x5|=V$-t!EJUWV_oSgU+@96CBlY7}BGl)uCTGZO z{svgPrdV;%JhTomCeyb##3#GT_8-N>{$7^s@{D?WRNOx(H*0TJ0K4BK!%g(1LoxJM zplCciSiF?JM%lsI#Gw`rFa?7?%ygz6OO~*O)zXA-cs5Bg{Nog^KMLe34aVGzpB@zS zaHrr!IlN@8mQ1+$@T8dvG39Jip2Fu4r2j}VUM8>m(^vMS(bP-%Nw%B*w5^}$kKfMQ zxxVkt1gk%#p=c9zp**=eFz&MXxj|u){k2LRUYkq8&0S6DRln#x*?J!Wqnlas2kV(s zXMAwS$h8Z=8iUyN*vIHxF5USXBAMIGrHqqDahx|AQj0amtVMZH!!eQho1S>>IUVp( zs|zx9^-!jgKE(I*!xWQny&;)SmdqVVegamS~pI&6i4ZXCZ9niJoihanZfi$JO2}jb8Ns(A8NwISXRq?V%aDF2DnntVRs_6}|0aQU9;07zbVR#SnD})djBR+xH>7;;5T29#s zoGzX#^>0EMe|&0$(*efiPRq5{SK7Q%K$Kobt(TG~nm$q8D7tx#==1uC3m!$&gSB>~ z^8DBTQ{8ueHMK?SVnIbjM2UbR97RNFQIXy}A{_*Q5UD{CkP-sYTY^UvQHr2|bZJ5o zLg*cWdH^XQ3WVOIghL>N7D^J}?wot?AMnN-_m1~|${u^o?6vlqbFTTVZ_cqc9Iy^2 z{N2N`1#>#@XVyMr5{*#KhDeo^H}@sJ8uSRA(k%pVf#lk-kHqBE35al>w7|RL>(rAv zF?o0MBuXr%f;H0rq5S!0XY|#GmWr>l#2CO@drVArm#e4Z)aLbFH7iN~NI-56*7-gr z#7IdFQF$|iw=oXgZe(0~BXUbzom%?_T;e8B8aB&M)i)K~6zN0t-N>EM%AGM%V1dw^ zDhn17dS;4JXUmI|O_YYh8gkh&1;Iq=9!wYQv(6Xnducr} z58`JJ!i4PPcda2In>9B02ND1vWvN#V=~-OKe$$rsFwwj}Vz$%Hk#^>J&>x#Nb#emm zu}`aT9?(AdggkEy{P98G-SsMTY=?ASCkAmPz`dNb(0du;?MjJy(R4-e7nzu8`r5L!K#6wUK=Q4Njud(-W9opVHC1={(I?ULC0+`eiD3$$zfnfhw!M+`Tfdx} z-~KF6(xeWEqjz-^l3kN}dMR0o@Q%1*WQlaRDh*Hl-+Au&z zTwzjYovxxc6{YWAy`ZfR>Ni($hUwTOW{>I6TJ>`M3}zVcYHGJ1)vC$5F6dyRR9$#CiG4mx2V~@TT6rtwH zFb*2|y7RJkx^LU(Ho^krj?3}uNMvQ9+zt$j(jZ13MhBzb(Rt=lOL7>W|4DPytqX6L zI|pt?{1`AU+2+%s`P(M|2Lw_kZ8cleQwQ$tWa?U_hc1J)2IxjESD$D$8zx&<(OXCX zu`F(m557Ap?0!3KRJi)}&ULEtl|(rnRq&bcx@6BsoXHaei*WK3-uU4D}x~X=Q{P} ziLq5HH5B~t!Ze)M@lFvh_H~T6WX{}2!yqVuJ!a}BX6*t`?y=*CQ1$RkQ9z+P7k1du zC9S+x+|aC(zBiP24lmuyWxSHic)9yBZdmulP8xTMrB3_j;7xzR8u;ta?RECcxZr&4 zpMAeZe~wrSooch%9X7YgQa>`u{H0)BebZ`up^UESHd$nJRiSg^n&XzbXK$R}zTMU6 z+%M7JFe6M;m0y{Y2Td?r9>+auJ#`zvTq-N^Ayl_@GQT3+re-BhyZA4ku!(Da*-n&n zM~M$^lvMs@k2eVX71@#t-k`7``LKx|b;lzvHo=AurbYfnz3r&omNKK@z96sDr98kF1yl0N2~a zFM(rtJn0|T2jCc7Ni2LXAFe*MdyktNbcA?58SM?*d~0ytIiu${n!5d164I!Z1sO*a zKyKh1)>2e|ja2p9EA6gM!sM-Xso$dUeq?kQD3@7YDlZl~H~%VGXSTe|=mJh%BgmyA z3EGhrX2;Cc*N06YW;)k=u)C{Q6OHU^ks3O>XA}HC9ZAxWgh-7yTY%CSMRx*Rw%x?h z_O=Bp!{Gp2DuHceW8Cc+J`TO77a!+vw(&52tH{KsSMknDPx=lX+BcQMmsob5=XK9@ zub-rMg7=L>eu?je7{A;-cm_wgZnWzMtuR|x3AFC8%c7{3(*5L5&ls$T*NK(w4ZF?y z%1wK2&26dSPs+0FIH0eCKxj>OTw8_@+Uf4;ji!A{+*#fCQwtm_UR1=7!ySrCo)dfl z;PsSRxHxjQJOIJgs#H?~y1unvG#k)Ph6FndT#|jQtKqjz1=S!>{A-z~xEI zU(iOU7AtiZ7eZIp0s=aMH=rCO{9~Lpg5R}P_Ncm2W4ke}12!K)^uB){2*l}s9V7IQ zLwZR@uza5Ue9h;hgSY4DKnDDIP$JabV&8$MP& zX(R;EtO@C^NDx;cg&LUI(O!?#KClMkLhei3d5$*uZ{dyuj^SEU61JhO+#6CSr5v~t z67^23rIeCigxkEm`^QWws=vhlX;d# zN-E}GO4Sg*sD1x4_U`u;r9`NOgNBa@awIaIN#whD-XErttcB`q_0`kB^woR2$}B5X z0b3Y&^+kx-Fj>RnSdMnF5jw}NXE`Lwqp=UD;0cZoLq%jK9dZu;%BJ9CDiCE}#532g z!%KIoq(&Os#HFuO)hn94EoI01_A(s;{#%bmHjU2>U0htuC|>UFmk;UnfsS_+og#sM z@V19nNJt+!1bh4gHf?+~Tq|eO_{L1%(9i5}E{p4evRxZeGckj`!B_ii-b=m2oE{Od zDc59Avoynd_<`x+S{C7uT0G~Ite0=9%6eaR16B8XSki{W7@Eb)bH=^~axuEy)o~K|#U5 z>ogkC)Rc-B<8VZ5KlE*ro5H8&=lP}a<}j0>p`j1QW>GJJTwr>a1 zXOaBsVj%teUApf%*N=Yy2-fTqWnXe8GURR2eLDMmlm*;*g&0O+n1c7`{CB z#cvK{ztd#_vnE5W8Eue=2+&OtT%zqN$fkcFTO2865mKEO~h=x*HH zuW)j5oXz@9mnF=eJOq8)`i~r7dLSX8AbmFae;~h#1gag!%sf zwVQ3FVF%=q?ik+x)9U{prZ~*;x?uXv#mEm2G-+jMeJd;LK8%76H{EbgYm?7lSl|p zxZHJFjrC$@N95Wr8EYM=h;qr%YLk@AOMRfS__2x2hVHNZh24v-B+1Fjjyu3={W?x# zeGk^oJy4skVk~Ket)vx&R;(u_C1EgBZ&zWJXYU1pj<~jNTSb;fU-DlSOy1jPkWlPd zak1d($wt6lwf6u1@OQQ~!~ut_7_&>y7@)iBU&1OsP*>Z%-8XWB^Y36`kkj&GOyGq$ zJ;)^X%k0`trJ42!9);S#J^v@I+y%$sTQ(C+FDl#nFDetUUr?*F89Bkr0?or%;n%mqb0Y5L znqNKp5G*4t19>%KpX@$*t3%9Z`cL5VKt!S`f=%j)yEur|1I?u^n1I3f6Vcy+DrIM) zy|rU$o`rhhV9m4N?0R~W9B1&$S;bt={B|||)2Ah zV=qaX8v{)V+E%;|)O;#9FH{zE>k30a^1s$+F)P;`8Zue!!TJxKHxt+=9MT@edigvYcHXUV;c~FJqrrQju3Mh1b4a&P3_<@5#+A< z@2?X+`A) z#62PL96kImT2M^i#XKnNxJ08d{bKBSF@3&aVU1zvyu4)y+QaA01rW^O@#Dw6xZ-w` z3a{>NlJeOYQu8J!U3y@06?4FUnzI)ul2U1ove^$-7NVRHj zO$K)*VmZ@ZT80H_@!7&Ybt5>ybrA%v8P`LH*i+Z#<$J-`PrSOv z2}zUZWA1;ubP0e{hj%OzxtvM7M-R_~K#9G*qYmoiDgS?EzEH2YxXeSy*L3{WEG$3Rv^qtYB=R2N)(x-^$}0862!HU-;Q!jw>4&g z%NdbpG1ZY1>d{wVXko-y9~|=@YYN%m?k`eSQkqv$`Mm76b;Ny zeK;8d!a@nFea0C35?d>8Hhp`d)~oF;#B-?Eu5PtygQUE$jM=jra;Wv+eNHZ(wD450 zT>AFyTSs>AOn`(;yGA|P7$88h#HCW5wsZ;g3rKOe0MqV$@+Zg6_h zwTlbLMz)LO^yJ`2$wo-uj!S^)OdAJ~$xtW+x+F)P2SFVqBk{)o(@~69SvO#?l2aY~ zE{aFYC3==NJ@1IWMCo=;P+uWK9c>8S6G#>{o>9iv-6OK+`pwMJZo9O-1y#BcsoGAFT@Q)N zqtAcq3S|LWRIfoCbnEb=?}W_#NE}%#HuC1zntNZdz7B;UZGiK(nM6dd#!^YGE6F;; z@*UO`8{(?5ooN&b#XD*)WCby|=?6WPr1L^TSbL@!YTqv)fYAem#kD^HocuOa#ka2@ zj84!NKB=khtY>W8-Ii*;YHC&`>v#H$5&I*g>|iRIYT|p(v^Z6}aXy7tzR8>m zcIayBuxGL0ubZ*(v6=t{(gDztGw$D?R+N)%aJ8LAAb6BcNv%G4>E&`^~RGaxr9YK&QWRFQCu$O`e zVY5Lk`IBm_Ouu%sPoIs(#ZjAyw%-|3`&-T1alm)=twKm_YHU}&X9;B&mj3Oes!&N;8OOCG)rqcboQ{?hH zV$m2328&(tfO`lq>|6+Ze0+t4g)fisTdU5^=)UU-3R)Q-86PN+#yV63c<8Q6n_E`I z)}mWx`IFt}$)7Pvbk&7yEJZ{!bhD98z+VeTvk-A4`%Etj-xlSd70vBQ^}yvvkX*ik zr_n;EqajWnPbL}yrc{1MEx7iW_m9&T;^I89H;pL&mfN>lE*|&p(F4V&wK>8SG|8#^a_Bs;((>3ScZN8!IgZriToc{U`);tpx diff --git a/docs/images/catalog-dark.png b/docs/images/catalog-dark.png index 3d8e4c89d8743515ff39c95935ba38fb737781f6..9d24e225aabbd90a724985f02e052e8057f2da9f 100644 GIT binary patch literal 25706 zcmcG#XIzun+PCd&Mp0?wMg>Jcw*o3s6zL>!U=#rXl_oVZG?5Z|fDkfDo3Vh>LPrGY z5PDA{K|ngvLQ6vEH3@+vgoNaY`@Qe?^Yi8TlJyH*xym}%b(a5mtbZO`7>k|0c>36} zV`3%`|9E=r*zr%tjtRw{{FmS_(`vuO9Xs~hF_S-jw+_$QCJ5&^c@)sd8X9w)Vpi<0 z|Bgic=koQNCr|&WRw^aJB%*d^2TJgXpI^WEPk_ziKdj5GH_%1DKeTLO-C2QYr`*am z{rAI#xV69`NEq|cOt5D>oL`c5_H^8@f=gZY-1zSh^eXKl%xBsiSV7MIlfBmzxYl7py26vUGGn<&zMBXd+-qKrY zKBcG&+U3BRji+&jSAJeIZ}P;6`Xp6*dk2T7U`J>BCwbe_X!y}iE+Hmn=7LoZGL16v z{^ox#9?s0>d>_l&?C&|8b?^1@qi%0fa?u0J?n*(TKW{EP-Enfv#lfL2c=N$;a@P%< z!bjbe{{8bI7VN95&=r^_y_tV-U2($ZVgC`na9r?gUaf-rl7-H{H$N@Wyd>_?X414E zIQ-aWmgwy3!qK+N_Tj%Lk0`ZcR3}?}ej2JKH1D=!!17*MZ**!F+MtoMIgwl+f9K~^ zew~U~_3eD}9Tj<1DLBKX%IDx|~5PaTlenWVd7>_X@miOutvkgkT_gmVmKUk!tr>1%~?L<^>e^-HwsJ@h|ta>Pa7yTI% zBw4jg={}0nlnRrzbFj)I&gY(08@c4@U{^b{P#3^>ZcLgrw_s*j+h&<4SjJVzZQcOB zqA70!3w*aA;fyZ8z-+vK_(H=QfTDkwSo39WVzb&NtU>n{YMGXqCOmW4xd+xX*+?MPJv@gwj8ct+cZFsGpQf-5A?Cv&(g|htpUkfaA z%I#u>GIW!yoB!lce~#Yj>(Q05e1WG4

    {9z!5LYc`mP*~*BgqQkZo<(0G5E^*WD!(+M~lINME zaaPX$#ND_gaKmBeicvz&(Zs1jAcDSLg(a8prbDR7ya2MKnk=Im3F#8k-bgn-1V)lK zA8Tv1ME9(Zk#cU67cQLZ8xEz8@u1vDkKMYTZ_UUvb~?wD2{*bkn!EKZ)jK22U@x}q zHsVKJV4;lS`?>9wvZ_MmNA+9#CL=RFKWY!MLfQ&H^M=YMtm;(}y%85}acd%R=c-gB zRKW8#-)-Js7*~&8oW3zGuD|iOz1pR0S<+K!k{0J+HvzLFhQN`5zwD=ovxQ9K7QczAd#InEcX`8yp|o`C zSa}=UNd}`SA$!6m&Yk&f$T#K+RAB9_RzUA?nQ30^2X$V!F@8&e@QwK|P)D6gM0#6< zE2(|NQJ3pyq8BSqMYxXQlZDWWPV=Yb>f0# zlZup0lP5BYKI`Sqf7tTtw@!rP9K}Sh$#b8N*n3sUn&GF4F-JT7P_)mEI*ptVd?l)5 z|I3%mNA%S7O$q_mB-3TM^lXy*=n`vXkF8H_MPSja>o+3$U`N0+p)%KzhgZedX?1k&ReHK<;XQQZ^$CKqiqY(VB zM-sq}74KD*@a6byh*_RL{OAb8nT#L+(0x|do%430b1b}}f6A(pRGdgn zD;5@x|MvrXMH}hw;D)-brr}rUm;GRk#)7(QDP~{jjWo*@|sz`J+9nbeFfN0dF zl|-rT%jSL+Pzndh;E#YEDbFUF8&n*VEp~2~ZD-7+fy3GBL$b>IZ4qeYDOop0P3Ae0 zf#3G`8UAZs8&FhN%VMp%oD+~X`sNk6>l2aw5d2UM3YY?M?~RZe>1HTtXM^OeW>pBA zFqG?Rab5|yHQR(Gj03*ZE<*%{?(p<5ridW#9<6=EkioRSywBz9M@*auj&_5`#F6+V zOK*f>)sdwoh1&AP2y&i{8VPlHd<)VcIom66Ec^ii!45DjiOJO{*yf`>eQoSF-vuW` zBj40bAb=Vw0;kyIa3$Z&qdf&T`pdbLr~@ldzI(eyW_8iw0u6dN`jT;|n|Tg@UQg*- zS=}|8tfGPVMBS{LCZ;|q6H7}UeYINGKfjW+xQ?db8ih@wCM!&p%wK}29k0eMNn-(* zD_iI1p9hM7W^;t)^EjxAYJyLau49fT??_@8K&Qo|c;(iWZ_eI8Z95pW742-+#WUY{ zYov_z6g*Dzp+Yd9nj4FJ7S2tkB;G>F3cU69_ScUCE0KXw@(~?l0nJ`uJSJ`f8|A(C-R)5&W$=ic zK1^ebaA1Z}7cjs-GP|(MY1*o^3ZXUMc`mG5h>+beV#zPt;Mf^pRQg^|#)wpV>u+T> z1tg&KUet5`$w8Q{U0AD@W)D)G$K)}FPuY0mVhb6Yy0UBo%*?=XYEHv$0&YvZjKgl+ zA_ih z`f@eN`_8^dwVdcbn|vZ*w;SfS_^jrBOe=06x@N!lcoK3TpqZ>Zn3>-ZIPaKLDUaa^ zt)cYR;(L(}>p-<5k^)4hGHkhC5Mi2J@Xny>bwo?ZM8gNYc-w8H-Itp(DYrS-%IMeZ z>+4O5ftt*}iuQ=Od(b4OeB>08wUE&CiGZ+x=Z%D_6Jj>JTnzBWoo7?g*R_oL^c| zfu>m^&}2|^oR2v0IY}E!X%HJu+C^2M+V|jf~I)Q~(*}N1*1o?2X%}0B<)^lGCdlv-I?BwLC!lnc{lHd5XJ;#O_w; zt0l37sbZ@>WeF!4%?F9z&6BiW6o=6T0MViZ6o0^(&|PrwvE6pcFGO)xeL~(|Tac2v z)cM!H=eMKx;8?It$2{Nzh|Dp2K=FOrQL*?nI)=mjVdK{o>gMaWdH|@1muKyRt6L7q zdG1y?IF<$Vk@#}E0JAKl<}3Q;(}DZ>W_N6W$+#jUlr;8=yrxm@FX<@nG1k|ZpAkt# zdd8x(iYt7XVMiER7Z-2L<>4-gMg(G@qdG_d+kXT|Tn_su96ZU0Ruwwq3F zJM-_!cR#*#54iF=Hl&o3hfZP_Dn=ZDPTDwlqJb_27TtVvkhZu2cESVcX^*Y)REHBt zd@q=InxLxtHD;r3IAmjCx$tSdrTI=+(951veFudpV>R(uFY|?Uvt(kCZ`7T`(ur?N zR^yfyLk3j{Zb8+I6CxT7A5bzkO_D(#_E{dPVZznhDZiyB$lB7!kmBz}EOftw@Q)Hs zC_*BwX39#$or61%3qhiTq0Y{m%HXmb4W*;^ko_YTLPG2`&QCu4GOHTOm=7|e$WpI^ z_64a=<|4UuZ*7qHbU}yvK`0M8{kvpO($Pc$$x+@vwvM9O5or5`2jgOnhzp zM&b;;H-Wgv46s|{lN7GxUbCEiZX8^&0#0L8#TB=o&c!RBwG{!{@gi|)&7Lo3PpcqV zD2E(_t2+}aI;UrR@f~arNkxM;_Xu!7(8T9-f}$KSGUVu&T<0Zf$Ig%a%C>eXsbpN? zQpx*vFCNcL;sPLgUyybbQZYxh7Yc8t-B`CFNS3qyH`q<8LHZzBKXj_lpcAiV8{&AJfjFZuLB4{~O#G9nbt>2(>CK0zx;j_rS(Lyv`*kHF_*XW#4} z4ICyNt}cpM4Bz?CIC0zQf3*OWs6WlGp0ogcrc`PtSx#*S1kOXxOC<@>GeMe!{SEVN zmopD|UC>axR59L0bSReWr^gG2;nhf4McT9X+`=s=8WBF0@ptFN4Ii{p%{0pU<0<>O zSo;X#kkRDEWL4PB*svI6vjK|I3%Mj8ox55+9}RY}p8tAnxoB}!1&nEE z)v;f{h(v_b_dIRlTMy%~ca!F?pAhIBitjlX^wAG&GcY8OFz9ykV(~S0=e5c9RcWy@ zO>h^k^xC-xQGg10UpS4Lmf~{zB|b&FrQ)q>x$bJlu$ARbd6juwb>QMZUrKUqFRCWU zqxW)G%#f?7c2&-gq;xwQF=s0Wy^xb*5>cN%`wkrr8Au+~}jE}70oZEiPY7M1#GFDdTFfT41dzk86Y?*t< zd$AEu>Ih-?j~XZMByVjEHj->!v_opp1-JMCV^gUk#qmsLzzM~b$F*EtjUxcEip8Bj}IqqNR>O$nYCc~rTv*Z{r=)HJDc zS!jXnE_sczw}cg?l7i3V^-)Q>f&IPb&*zVriyRs*wCvS3zN)IebCj*2c(!ET=P8>_ z%?RJLOn=`o(mqA-5_s!(|tz+nJPt41mubF{~ zD6AY>T8K$qEWe}CTTKYN8vD=B8b2_R&>BAeYN;o5&i`H25O;yG^aeY;o;Ap(<{Ttb zRZb!GBlnc`gCu%OTaVadS8ATS=ZnQ&s5I3@q)( z?ov;BTDt#{xox!rz#BLF%j~yPOLO{qB5{T<3F^N6%l{}{LTsZDe8gG~luU#TM{LP} z$Lm)!K9Sg+Bk@$%$?}56B}6;7e}ES{xv(&+In+%mYVzEPKCH-&x1F-cYhgYUh7HcB zaOcMLL+lDK%HXgoD`b15dwr`rOsol|Y9APJpHeVIDLh^Fs;E^h+da8%%-VB^`tRt9 zSHAXP>Ug5nb-(iA`DqgWAtKsvC0I9L(7AClQJ6xl_Nm<7$i3UP$~^woKSv+&$Yv&o zP+~_6gl)LJjZdtFadxa&zy)%GJO_bvv3GO{qE((EppS-jauHZwybxgTjwg9Jw}<(}LFJtD zXI`{_kXc2^1|>L(QN3>O``wR3^?7vHhGZyO$InSs$=YF`TTFgmf@5Bch$U$U?%^6) z7fd@l!S<=X*R1~%MzRug>N*(-gQe2d22@?him~dm&~JnNg`Rd8L1k#~rp>{~TBU4k z%D=oMp-P&Kv3gXgb_(ozDB{1>0Sg3M*q13S|G2zB@8$u;^l`lqIcAt;%IbK`k z!@iD|z$gH@@6YP|=CI*Qw=3%I?j0CUpQ)E4{ZJi5e=vy00$i!RUXX)j-3aW(pBiu8 zyV<@GwX(bkT^3^C#A=$kj(XpOgsf#lxOJ={@P?>j+*D_=aHMU>`x*Y8O(B#tfAmg1 zpSK`9&}A1KOx~{NfQ~kZ;}%;k9pREjt-|ty=oJ+}Gck*>d^CZFjMyu#ip&1xHZ}Fg z{(7aefo|@*+&M1A-Bl~Wl5NQSsA!&~ zTdzU1=RW~mW~!XEIZMP*r!b}6edY}+YKtRdFu&eYkXX~!8qr_$582Si5oZ5TqHKpP zt6lOZcQC380Yy1wRQnDCBB5i%e3hc)H@;8nJg;eHDDR7=E(;6mr%Y7CLMO2`?%UOA zlZ|~EeZ%6Hb_fJBDV5>SFuRZtJ9ks4tX;0!*_$_&+hLJ*D&hO_6!oX(RZcafKC6^= ziUA)#-7<8Yf|z1z2^9a zHJ#`t^QL#=V0Y4w6gfE!uGa5ZZ%^`@4nHJUsc5O5tnWpMUF!DJjvryyabg+bB9}g! zgqM>~Nq{A1`)Mm=N59M8;@%4xOYeahwKvpw#N@@qn`BajJmnOazlUGdmF(wni6&Pi z2Wxg2na0_|&9rASB76LI!FD_brP{qcwurNXXow0dJkE-ZUt^xwQ+LmOGMTyPa?_Qv zb7yrI9BjH2sTSkh>u2;yAckv|)wz0lt~#C$9=VmOtzeUqV(oX+8Q8Fj9nD^zvl3EH zVKQ&S;HU=w@#9y{zVDSApw=3hSwjvj^`_OS1By#1ug2SEJyy{1Du(rH;L`-w-{m!qDS|Z>gX>BJk-f#7^T{ z&RB9X$R%hd&!?$MI*cSF!6zpnqt*eCeDu~}pj&Uz_y)bM&P1}8&--jPtaH?uo9he$ za_}gel>y>H#Ef9;av(^|1!KZ1AI z%Y}y}Pc@3o+hG%>YyvvX8ExayY6wIQr`|=q99{`hRLhWBke zfJifsJ@iqGxih+5^KN{0iS~GDbm)0%N$+UnfUbBUMo5r^Lx(Qg+Y zD3Ea&D=m1(yBII)TWn?}i0~tVhtPvK92{KHHSEyYS58u4GY9M54;R1=2KyI*Id3m?pl_E2X__?oW|OS2BO_XZiGawE3WSherzvJZGR8$9dy*q z5Hv$99m>5-%2D;Bjo16MzdX&nmV3crJXc=S2zzZVpR#3g!Y&R(){dCtPzKNo$8$j( z`$>Hg`(V8gjehKyeY*CllLcx<7TwtR*d}drOL;!nl2Qdxfi1}-@P5pF4D5Win zWLjT&_i0wxQ?Xv+k%DEIE%ma#W`Mh?8F;r-B3(Y5W?bD`)==eU4R=-rOKjpS<)8v$|TK?ePJBY)RG?)Ii_o3Dc61CQp2TF6L<+m8c zU0XQ2J0bC@Jhh|7R_UvTWvzZsNPANRc8#D8*N<8sNuzc5Sl>`mBhpjf+>F%7<&gL7w z#y}H>#B!H(Nu_~F0=8YazU= z*c}DyXhz!oYk(t!$MW7cP2q@SPz8tGGjsq4!3-(RRR40x|eO#vSAU5{YtN z(<`qfE9Z0mK=AJ=rl=B)Xv59uT@@d?am37QUmoIH0omuE`r_smf303#r1><<-ayjN z#$#m!K^r2Uiste$Sq9$hC=6`ZW^y2U2D>GrKbtgo3tOo^{qKu5dR8O=Y005Wvu zKjJyIK<+$-$7;OFRxd|PS|a@@KtYauMO75cU!cEgV-2AqM>=SlQNIm+1cv?CT(p!p zq<@~z5+=cd*fS-T(s`e}o=b9$V!g$mvE~hnBN&#*A!qhMC8sm3iVVHs0`&E~AE=*7 zzF)Ie)Tp*k3m_e>Dbl~uybGvcW_WT&^UEF2v%!j~V)9eVl|FOHN_TQn41ehO5OPk9 ztf*ie^w-SB%Iw_|5@C9-&PxXe_szSPiuS!k@_AM%LW0iu*AzSHZBqw|DTApnC`Nwy>aPew?OFnS_J6OhSjWP z2hUxpF5qz2<0|-!dm}&Um5t8p za!5Pfb-z)9eRmWsK5?>Qd=Np5(HcV~C6{*sm`53F7il!EQ1l5Pj) ztKGUuWH$!GSp2ci0ep=INMd^fVtJgiJrJ*@bP9AD+>N7?L6sP{Cx`>B!B-zoxlQj2 zG)5y>ENQVzs#3gaZBiH?U-D#r>MB#z{B)=K64z~x8e}DALvA?Q5`hY()7#NIdj}lTh z2a(s7#ItCxEtf@(DyT=%BgrtC@faHg6W4`2C;Z~0onola1%O`YZnq{J9v87kKX;S)_HWu067C4BqOF^N-se zJ#nJ!R6AfAIpTGP-In#}>9hKx?CzvsD+kZlkxo+NSss3p<&J5X3IH6gP{i( zc~4%Pm+J5thqM|&%fdH^limNkO^}jhkE>z+4C(XKGyndZ9o)D79}-xvW6~rXcs2iN zcrp?m&`;LD`#aFdOERFnl7b;U+_c)LqQ>FLNYsf$4~ugZTI6_#7;<(A7iy8zlAhbvYq5jS3472Pn_?qI}9MbmoM z8f@(*-^th|TS&cp8C*6o+WH!wbFKA*XzpE|C~)$CpvOLCEVlu!-QJ-_Z>82|18IK* z7RGMKeCp}n3;n}MUjBONwAGD{OrjD+b2vt@1%#Q8p|-lP@0T%AUWU`&71ARDQ@th{al z-$e~|B``&27TwhR4Yqxf{Ak~ZVeUuCQ`;lee)8<7#0;B=$=+hjirv85f0U*nFWD%f z9{GT(dc(rDgRPi!N55J%n4nYwCnpy_PK}ow1X}fjYY{^wc+h6QF?c?FJEtE@n1gQs z2<)h!(S5pQN5}*gf4pv}QBJ`4q`0u!>S%)IOr+(;rhd=lU=*3i!K}gL=a*nfmXjIR z@;Cfo?9Fz>tQMby<$6rYdqF18pV$BWx>S9!#Y5>P-e1aR-{bbjBO}{i-pOQ?L@J8s zKD9K1dT(rVnGsUF{m)o;<%&BV9(}9K%l&KhLNOiy0u{bRWUGgRnox^qp};GSt-maV8G#0uZixma?&~I7hU%U-Gbgymf-!F ztAbV#6YD~*@QpGGi37;B1jCj_6Vb2dF8gLbis4zF3t!PiyKC22{xg!^~}T7F(oE=gp6-yKO!7d>{2W_?B{tTGCiKeQs! zm!5a;rb(LnnEw9dS_9}jwvzuhN`*N8`xF>Jc7 zta3w5g@Ci0X_MsQQkwBa^aRYErDvQ2?$3PJB}P zSw?%%b6$C8UGl^)l`j=IgOaDXy?R`H;{JgKAD|bc3Jss!tQ*!@f@Ho6B*IxVI zjz5XSufd&PpEG|bD4Us?ZTeI|=P@_`KT@#3e=y??rZS{NPVC--;K0Au-q@(@`0Y5O zkA(UXpV+%hJqFjOG!w4>7Y6)mUH8QQWA^cXkV(_z5=Z{m=Z;=bJ17{;zo&&Xn|4tU z=m8P-=ArkE{6Zt_z3bebxy<#B&DgF^2hR^j2-ISlZ}QEvr~AgQiK_f`tOfV|LOpCe z?ngv;AH3Cag~gHW5&Y+t5U>{%B7N?W^!UL<;^|u`rLO-uM`HBPb8PQ598X+r8dgxw z>bX0W^s>}Eyu7g?RI9PcGE?1i-|8+*1y%ClOMTT~?OIC-S75iV^<(3bI<}h40_%uX zeN|IlL+I;=hcaLEGBC!0EBK zwN8?L)V0o!oileco&;r1=B4As|J?9Z4&K(K-1zU4-D;{I`8)oBWV-@3hTvXt!CoHn3*tIQAx;`+_52UhH~e34A= zixD{wMy$=}z4Fz|B`)g&(o6**rMxbAO%2}3Wl~~ZJ%aDnON|cSX}OR&MlUdVui9iNJ**Pt8Gu+|RZDTH3a*j|2T(vk`bCN%^dt%UICfj>LYGs_G5VyU3|w z$FZox>pX@)*LtzFU0qc@KXx|G5Dc>#7>lS%BBtzYUDAj^ zhTLaN6U0j|0?OchUCtHATbPZ-6)0FEY^1PQHY93iy6lvr(jZfPuNSh^$<7OpQ0PnU zJx>hYQrW2*5>Q?hQC#uFiZ!t%aT*A_xh9VW8f413xg~-35)> zs!&GE-m2QODZO$`SaTo49d(L)N<=J3DFz()Z`6XaxBUvoY*X|^?A7bA2eTX9pl!mR)1Q$j8rkZUeN-Eu%T1aA6^S%p>Pt2c)2Pv=c# zzNKK@crnmsV$^u4uJLwL=GVYoFdqo%#7SWmN`+V?=rY56NM()7o?d(1Oy`sL z@B5nOF#mO^kv~Po;8vCVj!z98pF#N+b?|fNY{UwC_Cp$>&`<`nQELcYSmJGe)Sx?4 z*Oq3Eemi>tS^3hmgr1pfPK`MdNcSinON?w7Y_@d*bQ*ulzRf`N&jdm9+@bv`%h>BE zXr}I*5@!Ft$Coesw#`{o^p-#Z8}l0IA7D+xnM>Lb0Z&;CJMpJ}JR)Pqbyex74ejpS zJuy7#Nlz&sckK^(xqfiN^^?Z4tkfc_bCa1_A<<9$OyCj8Whxg2i*Js{b#^keBo8Iq40r|$#gu98{-@FNN2Ttcf4Ht2-b{1ZGvDnu+*Q9ZBi6hyy{!MAl)k}S zEKO=H4^INs^ttYgttcB3PI_PcHPO9yIuNz?Iem4Chsn#D6;pm-?qQ?X-hmEP&+ORn zJ1-X%?r(3;_HU=ug&l4-=BqT%hpin)bSzK1<$>Uv%L1B<0s|e+@S`XxBd(P1EeBS~@2pW`BM&IVgM6JK{@$H>_or31pHBzOhK5%a3)W9pCle1nV+)G3EUR8@n;niy&u2lE~=OC&yV zwrSkGc}OH;?kA~4EcsaK?=J}9n(IA>Px=(d3nWZuubn$Vi@_u9p9(Pk!~XRTh+*CR zHo4BuOFi#e`JD+B6%{N3EQxlB(#FDAC!ZF|Xq(Miba&M}5b$k{4Gr;}OCbO4?+N1f zJp%#)Nbo6ro%;okPg;8Kj(OCFWn%i?f9<+C+Pb&Ok`Sfn1^0Z{pU(_pbL*z^it~I0 z%x;#6*Pe;D-p0VWezUP10-SB_X=25?m+Vbhzc9FBW;U?51yKyx+JVZp@G(DqNC3Mg zMuL0bg$Y?Bu;RW3j{d?CvmBq8r=yfn@X0bIKia74zy~7Jv=&i&& zOfAEMV^O_v;d>)%xd)GA?28%9PUr_(n;Utwp|$ltm?=)A0ctwh%iKP>ADvk`Pdozx zaORC9^qvo9JV72U^^3{fEqbp{ZPRvI-`tG!?Hlv9QiK?IWo2c_=QUsM_!F9WTcy5Q z$ADY`_bM;HURHg}TtP<8sNr+n&D3^cUfvL)roMh_Su)9B`-6~n3vboYpCrI$OLF$d z0(3i0g7)H20NO$qS!44Q$no_DCeAlvyH>%f|`&Aq1|ty>sSn*XE3%2W%z%Eggw$Y=p*k z-aFVsH&ixMRJ1oVEDI#yXVMN04NdSR{?YUG=F*E7A@UIiS5Ue;=g3XUc2(j_Jq1dS zpX?$Kye--fd0gpgJIKIiH`yaJv{6_b?&~@nq0&|!IglVy}3OS_s~v0Y9V^A%?X!s!$mUhsdNv25sz(WVT8K^{Mm18 z3W~TaB~ZhjjPaH1oeSH!I&OnpkBq`?kB$i4ORM7OC6O`;2J^}Z zm^%`Wgdeob9$dRA*qlICxd2o+apHtg`HL5&YA*bAZA@_p;ReW1|0Lf(AO|@FZ(0s& zRrgB2>?#Uk#n50srP5vypf5jFR2;EN64An^q4`<9kq zn6dG*Vgdkg1;E8`pyb*2>|Ef^_dVDGh`2a_Ew#q;4|q!^?3-71JX*KZHx_a$|x8D!2;J2=IT{(SlxFjEwLoELxxuo8O3G+ehcevoDNcTmxo z10R<_2DU3j4f9J9V3mDkKUPLH(l;<1%V*Rf^1}K}3j7TufkT4rfCLTYfTBAJ3g^;H z9G~sxTLcNID4$1VwU{B=+CJ=5lvkRp9NLbwYv%7*e$nu0s+ACWC#Fq4dG(cyf7?nf zpRr0yI8c=Tj7GaLHdi6hEMs!g22C3(eJpCRi&|BF(SaSYI%mo6n^`0BzZ#K~=r#Y* z`}ObN6dw?o$yl-X4PCVGrFJny8I~{T`2#~u53GRPzG(VCYSBMxjT3$!Kw{bG);+1% z&IoB;I~!a0>Yn^bsN91)t`p4=Y5owAAy{dbMgIcr?G|ixXb`S_ofJwm6(oLmXwQMXOW- zT$-Xhq$gur?*zTBxr{6IUU~(7^6(*;pFa{Gito0Mogr*^ofi#_UU#?aYo@`2NqBf#dWDuC$kQU>%F(ooq}-tB)ixwTHO+1aAt?N%Z)Y!5&%{hKET{ zUwx#9_5E|D$Xn)oGwsT6S>JWBx}a~BG@J+tvzn#`qW^7_uzK%k3D(Z#;XrZyH6ujW zyN;bzpd&A^dy))4W9k%sbXQZLR6FuH%~<|cYKrB zo*hDnzeUHnU?CFW8B}6~llIGo` zZf@T6-#z0G8w%Bq0}Uye=lo`CmC!lO&?a8O$YNX%g86wIIh~bdq<>J6?381C+Zolf z+Fw=WX70Y5!X3pjO3A|s;@rUbAKi>hSJB^!^R*G=toLnp9qZGDZ}17Vk2l&FTF>*`Dpdyrdp9d@sL43n0sO6vI<_l500xy%Mzl(fF*A2>Mo~orK$W8Gnps$K#Rk z0hzpXb4oo~Oye$4{#MabXLt;`xxA`6N7G}lD@j#Q_4;mXI}>JQqw!1%uehoDgJIAX zq;ofx`gfJKf!7=1M6tfhgXqPm_AMeOr7B#o(%FVjzc*`7Ro^lhGFy{!Kb@r58Yk<= zLOl!Xeekr{^z; zhvlOk30ZF6&VzDs^K$HaFlsAoU8msIJd|(rGhYMx^}{6$CVtGE17dJV>9=Ge<= zyg9#<3nkfEIXg$Skh3Lr{$*ujGe3B59eq1z7=bq3qBBMmU|(A7L9;1YN`sDJUnSyJS00f_^MVuF zG#>xj*r-WvJ;xuFZw+fM_T2eJAD5xTg{^4v<#RBz z71n|0zg%VMeCMbQn!On=d?TB5^yC@#=p!rNKIb|6O`7Xaz%yBNU+LMq?=nJNnN3ZN zQfd+qtfZ~66vCR@6rP7qPj8y2oH&yp7d7Aq#n6r$6Cy^?O8E!LAEE8=0gv#Km>MS_ zvy~gjMqV|J8AP5L)v}9~_?Wv-^dWNar;`1-$$pbWm zH$rPxgKfO6#_odlq=0RtHn$e1h@ntBJ|~{9cO=$rbzE!?=Vj*TIJbO89?S@BlmHl@ z!du(QgO-hyd=ihEv`&A4{<&8QRw&F>UYQ_5D!tsJ88=GG?8 z@y2o$&Gh7W?B$F;QTVc!W^?S8aZ~oI&d#y1UcS#$+a}_o3=sHm*S$8*y=#wQ@W1QT zS1K)JUFKY-b?GlgUeN1BX7m^#lWN3zssy{hnXDgKFAHex$r|hQt-gjdLy^v$SG(Y# z#I7K5qYtvOdBKhcDe&c2f{i%)boGZi%=N2+Z7aXr3uTPTle*u3F0pWd9s66oMQka7 z294UC^BAy;-6j4^n}^;vG~ar62^ao+qi%(>t=ZrldBCtm`2&|f2?=UU9C*(xT~*md z{=}*7Un1@_h~aWxz3OD;vccUD5t&?^hCzyyEG*>Ubj7ogqaAr(z>jXrfe)7rQ1Myi zQOTz?if^S{)D!*#XqS95)$>jqeXQ;p*zW(dccoEHW!ai;ZE=8LH4Z34+EPS|AqXi23@_n*g~{JSgnp0m$Bdw=`e-~PtU(=C%wrtqk)=Cw{z`4lrHqG4(R z=pWFIS1rHc5g}e3>!jC^=6X#MJC8vWYcF^j+5M$_brIQnpZa(#xNIB)TzoM0H^@##BNATIFm0*$h z`=eH`(j|%!twJJw3Oi=O>vpHSKa)k8{kca~bx}C#6OdCqjqu*dSYVz)_7u+6 zY6@6DMI$YWA1z;7%VYrQ5^O@&CDn@U^B@Z4qZ#{=7p>L$NTyfsHJh!;8c+;_7APxJJ zXse+(;PNad8PRc5I5siBRCSw7HOgxLW{rgbaTtG`zNQj#(a1=)Hf^7p1}D)+iy~C+ z02wxbdSEn}+7{hVeB)V*m6|zitb}}sCWqo$rd|bYbI8L$uxtEY=Vnu2JA^J7UtO06 zH%?IgaEQ>jHG#Mm!&HyLvXx(wMcGVS)>L}m7UOu!R$-Cr1lmm}eayladC;e&VQ8MA zon7hGT=ylDL1jxdg{-D&DKmQvW8Wht&yUAjchWHV#~@zRdvl9(Jx!xb-sy%(6=-na zHRy5aTBhib27s5htj)u@jZ3V!;mS^sCA)~kf28FxQh=Sf%MXegs|QG$6JqJ>*1!f+ zd)t%d{^8fov`EL30TgBZhvIa}A`!da5H+0?+HMOriTn`^gdsRGT@UY}afh5^Uo#(F z02Xgzy(2gqmX_4^P=60)+;TW6(D-E!2n51kditS1Sqkv{BC814S0TNBHrbL-!7nII zD1qj~NIB8fczwHb7i@ zd-c9%{=+Y(+K*ZJ(#YnN7j0_WjF3L1-1ift1pPBr_j;v?{%>a zYWMo3k2K)CbZtC{yAFVUdb6X}S09V9(h5Fp$?)qsdzq9#6%;5Nw9e134t>NB#@qNh z5N++1YFF}nKv}!bbW+syXmDW|3rx#DH^x^2n&@D_)2Em0-aVailX2ydl4T$p;+Zs) z!@qtuQm}fsNHt{wXk?O|;N96LcX>H-@t7Lo)zjX&>jt^H4u+_lxKEU2>LCL8RaCp! zDBnOy0;yktEacS`MU)>kdV;ayzIl*lUKzg}Tbkt+)$pMHXX}|~MYRe5s*zmn99p|N zlWy)mSN=9E9ups$S9MaB3^7HqLseBLBxyukl+taZUcEX zfSXDTJi{Ck^b5>6M>FUSzA0?x2ep;r6-;D&Z#``sIeZV3>Xo?QT$*&;HCj5QeIGNJks5M$ zc2IE#gQg>`Yp+#an&{8wqBs{t8@~tYWnU}>3-$_0YtIo0%}b95=s@baQ~ID0$$evU z5d<^MktUMa&uJJ-!Z*uT9$3qK`DlZ0g4xl^=nfXzIHF4{Kd(enQpt#vR$YVk`^72dfBrKQ2O;~j;XpDXfL`s1N5OtvGx#X=j6=zw4k!X}3hhluW!>Yo8EJ#35Zegb2S0N%C3 z5C|8$K!>l|Qsd$fv$d@o2z89d;nOgbzK%jIy1e4&1;-GyvRhl~A2yRZqu478Xznu3 zDCkO!Zi9b}u4ltKAx9E7Qa+^r!b;E|7ee|p1@+!|nd{u4>-p2gn82CDcJ1$?H=8jl zW;3le_SWx4{YIZDr&;>voG>(Q%nCGSjxZQEyjvVGa&fBwM%ZA|Vs+#x9>1dH^&NWt zZVf4HDR9k=S?iT)#9WPvj7-)NB()5JzBQY$+D`1u!#J0%=Y{13+lA)cDb|Kc(}sh^ z1L|vOz7$9M?azxPMEsJM~liQVhtyOQG4g+tUG(^OB6(t8;-2ybtM9P$$KY|khA~q*Nk90g?1clxrbyW*HfSl40I&u|W zPvMac|E;^b=lZ?e@V94xu6M~zyM|NHkNIB>&dbNr_98SRRZu0)W%`TSf&L6(Wct3g z{s7?Yz+bd6qbi!TV*6J=G~Bb_R<8?pGM#_wEfbYDi}PZdlOJEIJ89*e(U-blBH{x} z{&AN4r;&%GZ|i2hqodRC8FWAS6v5;Ic)y=kKKS6>k%WZ%r@TE>O;h?DYm?1S7{A)JJ0BKI>d zgXN^CrdZG5!?V-XsiLAlmA_TOwNG}mJ#qGTHz~+Kz3?Im9|@yb2R4d3L(rq=@-D=o z*3|HH&%$x)OlM%&pQESejKUYQXu-}}4q)inmpwTtl#E{EK^NET=R*K#&&}oH5TnF- zp<^_!O@LG0vbE}Re-;7@`1r>CWh}3*EGu)9^QxqFcsVyD#_wLcVw5g?+_(TsFC@(U*GPe;@TV*IS}f`yr2&cgh6_|@!K1h|I5?J@2ihuo!u>KO@&Nv?$mef{!g|1q`} zxwh$lAj-u4-nM^ih~#TC@S-DjyN$bFH>l9Dl7n#l#4ZgGeES{j)GvP{`utiYrMh?b zO7=WJJn;g^L;s`z*`3X*qSlbS{z(CHjA-vaxL6u)9o)0jRrL>c4zbvKr)s!uQJ7GC zPQh^0jNGPQOu&Ahwg$VZf^!Ap&5X;|K^tkR2mU+vh_^1kqowv&tvkqoLDgo~9;lJ9$-F^GjGLyT;wQF4eHVxBVdf2qtl7R>eoCz8EKuSPb=s}Sp zy|)C2bO^nK7D%4l_y5Da*1ewReV_C3UF%&dD_N6$UD?;3nLRUmelv;Id!ia(qvuaUob97)McLGj#EQ}wx_A9`or@!g}z;JpK{p#jC<{@URB zawByOjccQ^pZ@Z8jr+sa)}nH!v6JW3c4yNYpPWDj20oZ3QvpErca zZ`YW}Z+}qz@5y(hThcn@p^z#k+YgU};(L3Msj2_E?H%*V-PO?`{r0Ie*VI%gqfMN4 z^1GYl3$d)+MGwwx{ADxTe@)4rxWxSxLU-?8B9Omqtk)^q-SE3ahZ9#CU2-k?;svIn z2jjK zTG68*sOndU19K1ZQb2|sdTts>K(mM=$*bGP_O*7|vlXR+@Wj7Xi2(dnRO(?`+&V+r)PCx6-0-6aCvuGwBeIKz7L!Caw6Z{hFj-l#5Zk zl!N`SdkIhmigz-(CpsJL+eJJ_f#<^^rP~da+An5dO6{Ps@W4mW**cRBc~VCgi-v$M zn*%#VkAd-$F7)*@%eIBylx04h7@1>DE?g6j&*9f;sS4TVf zZQrEX2(_ta2YquUp|e#sxJ4CcyVEOy2{i%!h?)6y$Vn&-U*h!H7NWuyXc+MXoN zi!-X7i2XBHx|_w?>@XCoaNN1pfe0O&(qDJtE~3BK2Y*?pdHf7=vHE@Jk?%Qa+o=K= zX-+98ZROYLP*|14p4Hx1NK)umy9|0N#9vq1E06VqEcY-^szM$y1c%cDC}E92HC>Uj zh2=s#D}MCIdF-HCq|Ax0koG>v6yz$u5rR{&)OuCb?CnNY>SU0MbsRlNgKILpvQaOQ zqXjg^VjWkV`sC9e6jr0(GE&j~vEKCqMX*v6cvkpmlw|rtPJ{b^d-qs+=%u7=X$4@% zH+&{kfcyfLGGeJNU(0=51BGlgxgrJ_@6KQSOabExwWiN*Zf{s}F?<>8WPXT0IJjL% zLN&j&i_;S?4L^x}zTV({80%9u>117ZiWaFp=?__OQuh5SCSMBkoIrUw3X^#Qpiv6P%mrrv{HPu z_SBj$jb>2PcwICrB^l%lb>#`YICh$y6o7UKvA&eix}y~ny(66$pLTy){|SAHK&2~B z;Z?PexW{*~3+W;o309Nz2i=sI`#ti^qhhPrIjMco(|ZmzLoLHq@?HTpE*ZW@Wuv;p zuY7O%X>%+&a87Z15yCeLz6tA}43e7btJ;K#r9Ab(m<{{%RP~@$D32k4$Hq30J)Q-1 z%+f`lsP$tfhl1h&xU0|J{`Z5Euih=*Wyc#dNzTfLhr*q|KAv~-0dsS$n}mi`u*5YP zgYM5@cf5Ne7gKE8NSaihs}=etZ@5sZK&JPfnICUI|C)ODg>c*_;KZb%FP{3A^~MZx zN}jUAbYK+ywR#J0su2_7cQ7oPRrkyzW^^%^wI$Lm=%nc;-(a=SRvOHnwKQk!`=jzh zjn>UbvFVI^HUW(oQ=<^x`BsPRQq8k3G5Wc^l^O6hl`N5v4%~zHMlr-r$4`>7sRuaeWMzH$J zhR)Ta_J60_{M}WOF(I2Cr`$zp4PM%fkANUVoot8l@`CnQji^~&g3Q~P06>qovS*77 zh2ZA7tRmR?wl0Y<;E3!)cKM=~ zY3^-+-3M1Z*Kp-j%$6(p!AQJJsDV}ENS4BMjZjj5Ch+AmDNU~Mtr$Oxd$ifAfF~)3 zPfvSTZ80Z@eEE6O^|otuG{dxo?}1#MwctE#4Ya3zuldiUkBMuC3D6Qy?*XMzsbLJ< zGuYA?xOw}zvdP!-^7y6|T>Gg-1uoUww$xUzh z+yI`@@r@D9>07~!?T0_@Y}4cq>^LfXX@#g_QIgy)+WV0wuUWa_6NQDlDT0yf1Lyki z*F6<`G=ZGeF(oojOlZ zh+3MT{*LV6#sGi9ze&I0Xjgj~IJJOTCfUGW3&+{IMc;T@wseDFQW1xEKEka7n01Qn`3j z^|oHaTqW3hg->^VmpE6{9cLdts@BONXECNlI{#hy!p$;QDU~*Ke`Ew=g*~fxVe1AY zK4=d?nhI>lagZO=nAy1?zB=`CE~4IEm(O;L3NINr-;eV6mP+-~;*Lq^`siYQ&=(l6 z?P|Ql>yCOt>ap%44rQqMQfk84ECkbTYx43*+CDz<5xc)Pc#}EbHFqX^GJ{s-ENoX@ z`oWjZi?*mp6>sYuUzL&6nDX{kuiu<|9CRr9_l?1p%8UlNfdZV$RRO>@+k3fY6z>X; z)19geO-c1-$V`dR*W~y~G9np}YVU)5y3Jty*juh)XZk=c)Mh&Oarox$)NBQALl5xL zsJtu)612>|`~BT{t%@LNKdLn9(P8_!jtS6X6F$^{O1sPmEmb`JwAZ{;4$-mjjAn=w z%`3AHw>Mk$JpWxVl(IfZ4Pd}aX!68~gB$jOaG6r;xn2RGq~C`n#{BRFkJ*=e134n$ zEG!S33GbsUr?Zt@F5Z*~HEy9G8U1Nyk|hv7qFlqg|K9fz4I^IIj*-dbyOy$))HmA{h&7B^2b+*2tB-r#M9>=oa>m2Ym+cwAS8Ajw`8v~O%U`4F$(|0wav z2=rW(dBmEYeukrt&u$mLda-WkOLGetT~v6BYLVG}bI0jz-Z|k-AxZgYxXQ)!%JpE2 zmc|U|EF3QG0Tu~F)wwvHH+z*4P*Pr3*jJXa3i#i(FJ;7l34?R;Vk_*r!*AkQxr!bn z@T5KH{e0PHgoiX&h;SJAB19*8Cfcu_Z+&lY3?YrDq|#yioHn*MdjpAI2Wbb3%w5h& zSr^quWhUEC$$1u{q_(>b#Q(C!MG;iYQ4!!bY_h#txJb}B*oT5HH zS(+&me9H9M%9dAK?Yv;~)b;4Vy+I*x`uRa(TQK2x{~IfZlnge1q*%gc-j*@t%#ZT zNvm9~jB*Q4hG0(M$!4N(sHdNu_m9@5*kOg*+VYZ?hGBfGZt!*(`twj_o(gfGhdDp6 zc#i=1cxo9qId(K`+2V0@*0}DmW_2@cYkr+5Ip^DAG{bRHNuzEO(N=C}ow#wC)_zX9 zc_kRHzZ8%AAeL ztTRkGE`pBa^J3f<15{{H>bXC>EljBFb*PXc2$Y{1KLp=VS8i*$AFig!I*&0;Nd?4A z1`0v~U+)jRi6#(>5KRJEC>|PJEoQ*S&%%jD9LvR9Y|4T9I}0JCBm_M8I?B@!^<^2E zGmc~7vO!wN%H<`L3$8gLgMu#-l=q@J;ETz~)ViV>89GzoU53G``~3|Wh87_w9ozgI z9$cOImg!fWQ%l8RZqjgX<;LSn^f1IwFuZXS9m%Mm)k4z@I$*y+@pgNWi)(z)(frlU z#t_CCV_t#V1ju$!a@F0rrRG!RHO4%p-DY>m;k9O(En|WWC_g=0CeAvp?>3bMGXsdp!j+-3 zaJzY^W>7yA$hUw39cV+rTs*cDs2eh0U9%1d(-J>1W!c zOXcZY^eHV_aIN({De0GCoTuhx)3kBgU|Oi~33Eqx9PIM%E0v*AA~EcROjV!nghJZ; z-?XA%5x}SPS)X~R7`>cize$$KTl*Iz0kntE9Z4-0myrU3LYqw~OK3`!XCB}w$L(CS z0Z4IrlOMdhdqh0heeaauuVo;V4f@Fg1p|QkZ}Bg^#zL>&7jsszT=nS&yad%hlZUiE zA;9~$ju$-|(%HZ>p5O-N`7)df!+i7U%n=W`SF)sw)35T(vEhVs&cnps#^aFZb&70oq0~3aGW7dADV0UKUlq052u5pJu4Vdy4pU|8Q zPNsuu;N21cL&X%$vIKd@>4m!-rsQDVbCUHwjL-CfBLK_1G7-^N#sK;(<`|jRKM~A3 z6%Z{}@JS#G88WOyEG;pYSzu{7=r*l>y4&2}erAI?QVnjyZAngBfg3lASCV;#BB`w> z?BRHQTofnKw6I8aVphuWq++C1Oit+n^1||D2tg~PyjRiP9SNE9&=rAvmagZ{dRHI3 zo;M#zqwSXOy{|X@#cOD$Vej>T$Awpu-qr$(6awq-WJHa+xj;;+(gv|PmDO#4=ow$k?yUbg>QOEw~Q^7|W zFz5VBjU&1#`9qte_tH9$#WNKSd}ZQawZOU>%i~~7iMb$39I(B4TFYN9Us}fm1lp?pn1rBeh$*qoKBoj) zcmCk5q{a4&$=@wZ4h}LsaY=xWvFlG5-qhD(T2?bJ;~kC60C%}Us^v*U_eVKOf=!2_ z3d*5c5QxZS1Fe*$xJp~7SmP3e=H-?D^89O~J3R2S!};sF?jk|kHO(r`Pca}{QR2n> z+_nY3#wj!<*IIoSEux>3<&Zi5(*uBa!i}+UzGr0rqT^)ia~$y8DOtR!d>A{mb5|Wi z;-J{Cfft02)HFg?XuiHr28+Dmcv4=b8lqR)W^4Z-?%oS)v!%_FFwvlb-oyV`&*yE3yg8Gn`l7@$540P2QVJCz=QotLBNC ziMmn_JPf8j=jnIcLJ6ZY^bAnl@B4j&-2EtITVLL8{{E0p`HC zbi6@J|JX;;+1n$=qwxai^khg&=te(jlF&JH*uym0>B>52E&t_fvu(dq$GG}1;(QDd z91%mD`GWSgNV~oxD>0#_=>b#^I&bcP%+HeAvGXI6Y?nc>Az55p>A_9t0j@6KYI@950w>z5eL8Cw?qupQ38!aM_titM$eOfZ9! zQ^O|f0iM?Km7Rn>|XW5Jl&`dgG7s^04kGak9pvh5Z&bq&5 zNO33iz*{zODsHycK~-fqQ^{p$+icVz(_8+cpa8YIL^YiFVg9(31Mm@7Rz`lAc(cE= zKX2mV;aHz*-;WF&nA(ewRHn;+CgVqpvJUynP>QMmdnRuR6IX9^b1zMHh7u+x==B@ z$vv*4Pq;Ypoz7R_X02VtLdN&iRdxtdxmVcxHO-^FPbM|0WGwF0^5b`Cr9;;o+xD(h zqLz%13fnR|F~J9871Cc+I8)U=)Ra`IAaH^GY9VJ^i}_W(CW=tSqo)2X2Wh!cw5ml@8+G74u>FZ(?u_}SQvbQ6<6L92pkeioEmAqP}@Dfw~vUI|K8 zJc5}Yr6!C|X{ojNOJ3#5i{&tMP;v`loZtLPFu5^lu8yy)n;NCa@+sMJ57EyQD=jV zqcsw5)U8ZB>so6k)?$deQ!TbI>TV=HoIq<&=l3EkH1Ovn>FD~jdQ}{+kWy{Q<;v7p z4u3Q`uS7fT#>j}BYnH~UJPdNuA50hV{aJa)Q)0(d*ZwVJC@VOP5d+lH1qyF#P;$9} zQ#eH;IF+h+9r`VUiSy-NCBs1B%-jcn6~T)Iov!u3Ry%ExoQLf}RslkK9S`NbkMM>X}p^ znc3)V3t#gzx(=!VrR=+4c_pprn<>kC7N7J%_n3a-34&?gorG)n%L|84OAv)qpC(<> z&O`*qfl~l01%tbFVr@sb# z{pfb@-iw0zZ#&hS#-sbMp8X*h5U+Sr_G1~QN(wxGJ~AmYhc;ynzi?_{6HY7aFZ|h? z#a6mqXdEV^Hj|Fq<;ZdG{rDmMqs56ws0d0yf{eZZ!miwf3hi`Eq#;9S*y z=^d%V#9(I#DE#Q7x)nOS=lpMOBzIzq#W6iNgy(0YXmGqV;O?m0{haetLZEusFvLQp zU5Qv&qLx2rYFa0D^nD#85!J&=hCVW5$Ij5Q-5;?f?4uBo*YPG@H87SKTSIEUuyD6;&O9xF0hN9{D0;bQx-b;5-67~;e8DyN|DC3g=Y}Q47qb1?y zsys`DrCSUtJmwv`z&8U_GP^g?EZpR*_mf8(8&fLW;olOI>aRn;@CWEqCm!g^0K(C6 zh#7AC$(x4cR(uRvY)l-d;3|a-; zj%(i+GYSend+Y9pIF7GX2&Ay!#6>-{6C^PFLHmX1iL}@cLrbf4c$cd7 z<9pnw50K!Ik<+l4b;g)gG|n3-e{>hi64_ktt|5^5if36I1H#Ed{^ZAsIO^RMKWTWL zu+wm!3o%+ay0n?a2e;Rkimc6C0&cu(5P6*@=(hLU3CWwH-B~X^y48c_fv5IKSidF% zQ@ssFGoL^LJpGKmesv=jkRxE_h$uAvt5SE?iqkUmu%S`EC>6+$rM(rYUOwLII@s@6 z@%UcLMP%}UpNFr&PZ*$;w)%P1i^imQ8d0s$Q z_Ilj^=m(txnoI?B-JqX9hc5IK3!U#6Efu2FJVXAjl`>ikPyJh{hmyYeW;o

    eXUTJE#=EVN1&M_X;n zo6)#JAM=MLQXOJXneM*0QbKVG^gFw{!ED44g(1R)McOQ%0kJ*bc~ zTvTaVpF57ENQO*O1L{w$TJt1jA4y2>ZCfSCKD0Ez1{OC0i z8!ye98MjUax3!E)K(lqT^w-2hEcMybK9cu)`fQBS42*rP@QTHWKZ2%hQ_>;Fe<*na zf8aNe44~9MxW@ld;t^;2iHSw1O~>bNKh;(CqX}=cA6zr^K3BmlCc2-)sb%v5mOV=J z1OicG;F}63zhr zPUTH^Bi){7Ie~fOF(v##Hwy+CFP^uKLBc9HlrMyfD&=JV2 z!C|0g28^+95Mht+(F`XIWdM!vh;-0%7I|b_z)mD}c3!`O%Yy$o&pBXm)LvJ$CAYVz zs?o00-b~*nHeurQYU9fzOQB$F0~E}^>5n4sRd~HLx+uAtFW*c?KC5MWiKmjiIOTww zMuqYk0^fP7PPo4Y!EOE+cOx0tY(+;_k3l8R`M-;Km8t`KSS(eD z^<4-i`a9-VzkSbqBfZ9AI8hGyY3%gs=`+~`+qgR)et&#@#WwWY@8oWsviHVF2SbTY z-rR1Kg9mbV)6w^odX=1ly!?yG`eD!J=1$=D0pFL28oN=s@e*yTTI+F17a0eaGHmc< z+VI{?eZbCzox~!67|k~zzg%Zm4C2BP` zr)MJ6lCXU59u3@}*X~bxXO6_450CHtd~0*v8EPZW=er$#mx9jSuz+lz=+*eo0jT=` z{(}3g(5^CeKHV(&7R7I*S#>wd(+S(#z=Jlq`Dv_+HcQ(y(g>y3yoI_wT`_O^^g&uG z5{f4)ugqN}&+hK#&hBOd6jKc?#8)a26tS_yjV5Bj(C*^m=x&}BCK=Q>A9$Ju zwAbn-qTav%!KGDdA<}L;oyV;v98VAI-SC6>yQrGTic+?HH|;Y9@!RwBIm{;PbXJ(o z#qeC|gu7*=TnMg6dKv{0 zF8a>P9dD#WiS_K&Z zjTMTOEA2NXd9>3CYds7E(t?MXJ34KvU*uZ)E?j+IIvRS#lS0Pybd@tmqoG%qn~qaH zr^jYMHg58cCr*qX_yrAZXZ|c~A~Py1su|R4>JAAt@zoV#N&$7+REOnSUJrY5VCuRQ z?KkMCM_EvG9L(Y8R@^*A8%`6?pk-$e4G^hHWsJWepFY=4`o7_`TOr9blpSb;L^@0_ zTt6j|v8-XaLQ@g!>F=MygF1WzS#qJfIYX5Tm3dY@AT-(frzQuU?20~!A=`M^yeE8Y zd>n(pa3=PI(a0C_oZ#zdSH`< za#|do_MQ(6Ka*aDKq%Y)P~?Z#Ac6_B_Yzmg5@Ct7oL9Pj{Y4|U zgsalsAa5m0p|mfFvNoS*So#vKeFuYYZ$-qWK6q6Fj0`~;W(Lf1;;3nb2(!NTVb?P+ z=2r_|Y--or%c-$gtIk!2`meW17V8X7O*H25h3kPB@80b32q0XHml|q$d9mE0riMvL z85tR6t`)2Ct>pytDe`yD(<*wdu~M<)oMY^VYbJ7iu`16L+3=QK%LfZ6q6-#g-?M_` zygXYo6e5&Z`yFD&5-0lXZ&8U?`x}ROHVb*QoB;a%0WZX&r>rpAZJ9s zlXWT`fuQE7>((0FHT8HOo~p*)4m?UhemjB9TBx6mV@I>w)?AlEH=A)dw#Ozns+mv6d2q@&Fl0Z87UgcP zj?o|+k)GZ3?&kDvNp;UIRmBiulVd>0*HllX)3to;V$^dkNaGsL*A#d)E!>_nF*c_1 ztZ^f!Rz!BM0<#zL_N~)5LjlOuP(R`JEl$6)YGitP(HigdA7e5(hU7vi8Z;4vge zhzmU{2|YiA1bN__sm-GXb#(gERZ3H(+;_Jb9EUAxeSSvQIub9Z6mCVSh0|4Tg&Z7( zntBMfhpfohPb*n>AJ@Ku%*>BF#?7{o!tE1P-k6nHR9g7r71-F>6E-MsP+a;}zI5tI z>F_kvW0@gre{@NO8|ppvahpU5B0d|YkMh#{B6nsa6(tT)F4)axD~ZgH&2)96aFAKl zBDvm$I5~gHxBLo;TuTS?e&(1AA?>g*=L1>k z#f1+;=x!tu9bR@J2C@|hVS6pb<>ih}erI2(nDSw3Z+>jGHw{yAg|9d99&6>9JT-D2 zi}vjDz4~V8a~w|Jf#}T!~x~$|%j*Hw3(YD{M1kM2Oh1Mpjwc)nYA5jfZY`yEbfn&ZudLrQ@@R8py zSewk3@h+A^fV-~V1#B9H@pUK68uYhc{kjgpq83}Ce28Th0s{T_ZBsHQe(!&yUqdSF znbilp(6XJcOF&u%(kNbJGct=>ZlAxWUm`+O(4(VMb9#FLAriY4mJPl5tKm>wVeF7n z;Orq^EW8F@p&PW)eQHUllyZ@A#kRRG*A82L^V8SS!He#-Hkb-cI@abB&!WuDvl?9I zAJroCvU_)+x${c5X?@nHBCg-2g~3jrO-!p?&MbSqQe3QypxO=IPa=En=~JXTc5)m; zA`9xWuV19Bk9EDFj|##(bV0L{ZVpCl`lVl zTu914s|O(^(ayi^jaY|-4z|!dl&^Eyd*4Q&4caLWy^0S#%eGK%CkKdjb5_X9QaQey zk-1NJnqdP?I4IK)U1MnbA{PoySbo}HW+=e1lM+q`O)>X62f&wJxw$g{>X68p^fbW- zHVL&Zr@K?>YqPb*E(pQ2KCf@z+=Z<}wa4p<>={^dME&-(I72ad|;zkjy#r}ZQ4g+rG%U#{meCd!R=W{BOZ zm73Jip$5jF(OF4JqpnaLOKI@a^rh&tqA4Z}l8qflE)J48?IfdBgUm5}gSi|xo(rut zGP-+WipV?K zLIDgz`)Gr8W0zBaDfV{Yt+YEG-V|V3l+pQDRcHh1-qGf z>qvu*GqE(9oR|?)%%c79*3Qn(&+a%GuBAB40P;RPjx12#iZl4B$C>v6A23`%Be=R$ z0Icp2K3!K2Jmgk-pG_s;5AAT;9-zgio*;*VmI^aOEQ3~8nV|u0<#b_sAa8|sy@Zcb0!~L{}*%D3jr(tQs6-|rHf#E{DMH#qbkYKwUNsd!n~#0qO$Ip z)ULwFIoYUbiCV1ODv%GNH7PJX4u-ho_A`dl<*hRwcCn6^R0KKfWKj=`fDb0f=Z!37 z#BwQlshG1l+e8)prw?@H55ircP!}z&Ka(_g7#OG10F~G}ctZo}!^`#UnWh#!AvMX5 z5XWW-H8L)46eI5L{hCm`u#-(dx<8s}YEqorO9K|Hzii?HgT2TudurX+ZRXe&W@IEl zABxL_rNsnZbfolGat>xRKm*@8osD8EOo1J*phKkV958?at-NOv7qZ0Vb{S5+B7T=RtiMl}ruz{mOO5*q8jPI9T@Kc?a1? z&DHusstp6Kl*kdcR`kOAf?PriE&jO5lyoUMylaMZ>q&2$r&Vf$>pk5ZVFrX3ym7AB z+TEQ!+nypblw8wkD8UaGGO-vw>ZD59u$(?BqKg2K=asQ@LG#wnAm_!KbfOkh6HdrJ z&c(%<4Y%dmpV0LXI>1hsT$sXQj(!SFy`K(3enO@ir!Le@B|vu+{5DiVgN~3pEu7$! zm2ERg&+{c2629fYufkBE3fnq~T$w+;3fRqIVBc~M#+m?cq6adSyr%Pl%ctD@He?cY zH$UgRD)HT$ZLnF*B0LJ3(eUWIOiAvwBX4<#Nyf*XV65naSWq9R^uc^M%bB$fjF-s^ z>5B3AalY<~Z30t{FQ6^JV1#SAt!K-pj%c~-VPSlOO5#Z$NjCw@Y-l zC@+~EqBQt17+#)~J3St}4Z0ElyRlL-8tOQEf?W3{hZ#egC3$n`F(LR|tvjiLb=+~_ zv;2BEYp=KSeB>b6h${NOhhp%s+1?b4&n~)_^!-OaIhw?!(vbGmRU1k(`rKf={uRlF zqEKq?xa7X0Rv5bs?LBHogX*aKg|}bbo!lMsw+Msr)WJWw$1%KvIq1cxW?Ro}_(UzZ zK>>MH>1&bxiszJbTi*UhYjI@rS3XT(DU7UL>+I|}!o_V-tLCD@U=Uqo3I+ozT_GPG zX?=XCX=vJ)4`Z|0*dJmvOUB6&0HU(&KEOd@GlJ^XnA*0|+9EifTp}=w_ESXJuIIhT zohQdMcEG@F&eoA@fIbjBS-aCFL3m}95OKFFhM^YP&*pxuKe>O}_0f=t4e~zvDNoS@ zGAt2?u&XxAG|ZRtU5vJ$c5#7f^4=eI3s$E5>1*Yo8EgqP6e)l=@q@vI_4>d!wq!4| zRb$4jN+_B#6p(|WQ*&ER9^!C$l7#S5jva*XV0MEm`y*B*f8ty-8hy~RV})T=ybNc& z$PFQ9pzgCl??!miMFMV@8V<(IGNWG!HH-es{{&T^&84sFDq44sQ8D39sQvQt2dBsi zW|}@ZHfJHkG%UYti19B_sVOKB0fUd`rgq3h(^ZWt9ppN}JQT4^Tb}Ymd1Ygr2zY*O zuFiVAb~UDDKlSFOVHpxx+*CTsq7TV{eyyUaM+&8B^D3Y*^1Kb?A|En1J~8cBqMKb4 zIE=a;B~OmQs+*d`AI0iALp>grpeid%t5GEld)tz)YK2@p1G5>tg!;4Qq)XwY)wL~- zz`z#H8{{04FS_HVV;)M365px7Q8&dnT{Tq1B6EY{z`?|G561y4Nf~9C>J&_?$<@n7 ztsA=r?W-%$;`3l#@m|ACt;lLk7y5t&dYIIBZ{vDY12!1TS|XAo4m~{k{Anx7nx)%T zT3IMWx;hsH^#X>?A^ z9It%$aJ5#Yi}dJ;vHIOi>-H5z(>}hm|01OIGVqa-2wb~6_b>KgswHuCRHIY^?5uKK0l7K@q!d3W>=_^X{ap2g|*@NQ_0=}AKCd)>a~vX!37w*3_bM>R`{Ue)cl zcb@GK-qT5Bz6A&?4kun*TlU~X4;ywwJL}A*Gb#l22{5|bOxA5!R;}NAgmUMWo*y#N~)8f_Z13-ev89 z79*B6i{1IWo|3{nfai2HsM^$ry+%si5S2iTcas}>dm7TOX_uW9_E!{2G{N|%JkH^1 zMt`&_fL~Bx@@UL+1^|RL$}iF9(k(?loFXROn_aGAg zi14-V*$ZLzadIQ``I_)DiP8<93ucvaD4*Mac;O_ykxz(%$!e{TAmk|{5S}pAz0BF6 zPv+Gd4DzEuxQ?HPQ+H+6bGp_tqeYtGc_#PSKLyh5Z$2MU5R+(vFUXyEwvR?M`M!QK z(p*|8P5JYuQP)dF>1vtvoA0Ib*4t-aysA37r}pyChlF47+-B0#XsDo@h5OJ zJpc*tcRCfAikO&Z<8p0j<2hQS2B?Ao>RSc=n_ooQUkbl%+=81{#;qJu&t?q>ZnTMY zd!N!$apP)?^i^i*4Y|NcFb{K36Etv?a@apd57e)1Tlq@p?H zhrjDfhd?pdXo`KNn-v>b6Ov*=5!$?DE9BLwF@i~*>c(Gsc1d;AJ+p)K)}Q-UNX3cg zai$99ah_4n;=B)y;;8X7agC__xE_Xs&-VO7zcKn>>$fi8F)rXTc!6;eIeHMLF-@{5 z1wC!aTi?P)gIm&4t3Rh`c0({lDcJW@pFpz6TwpgD*qk02UI_@PH1$y;2|pTNxrfph z;;|Jk5lH*o_)do6H-g6Wo!`M62e~se(#Z3+h6x1)-8V8uO-Zgj{hyxM|04h0Li?Z5 ze@_0N+He1l&AI=^zW+1&&&hw|=KpT{{SS=(kH~*x?*F;>pOgRQ;eX)!{|oZp9RBx2 z|Ka~XBLCg*{GWUOIr+~e`(Ik|{|owmw$lHz+$l|;kqbKhkvjg@#s6F#|4WtsFVyk> zZ4KlV{3GrBuZ#b`pq>Bq4+8((4E>i0`{!2a9|_ton?RD5&ZKC2@@G3)Iux2OU#V6C HUVr#sUm`&K diff --git a/docs/images/catalog-demo.gif b/docs/images/catalog-demo.gif new file mode 100644 index 0000000000000000000000000000000000000000..832ffd88223b78c41991ea79f96a018eceffee12 GIT binary patch literal 27756 zcmd3s_g52L7w;#@q_+vZ1V{o1QbSXuXy}NDp(!FtXo847p$Q^t=m;oLKoL-bfFc40 z1w};-y+h~-sG-=)6BQo2T;BWJ{SWS1-*whrYtEcmbIzPSXYc)S+3IY$DTsz>gJl80 z-(Q2jp9d|yX{(=JoOu{tf4{!u)}sTDmpLtef|~zw?*0mFTn)Uly077Hz^&CiwQKs_ zTIhGBETQ0<@N7}u$*aw2{F0RH)A3p7{cBeJuKnFr`Q7)*d(Yyx?80e})BP^leOD7) zPb9@0P03A2N{Kt38GSVMp9J1NF-O9a&mN3TJ`kN0!A;zMI3YADJ~%QqIQ*dV$tMo+ zH}{1d-g@X9Gp1-`lwecDWwTKJp5RD7PQ=c=VLp38z5RpP{-GO!&YA|~9%VBFJ#5*# z0zJHUZ}ZyY=DFL&W2f^rw!ND-%VozFM>iV>H*1!&Ih zFT>i8Zt0_Lv0ck>gO;9&mL5~h*hyaBMr*x^)_SJqnsnD>%+N8S>li8MS}ST<($*Pj zXc^G7_0=`?XqviIx~?isS4Cr;DotBiU0X?ATUkv@NljZxO-pf2)HPdEiK?lnrm0BP zR9KU$rm`ACk;+h1WhhV?3aV=cT~@E=QUEMnR3DpeCb0g%mVo$ES{nShmRsr6(|fJ zgZl?h+CqS|2#^YjphAQc2#^9EQoxcFh>$#CO=NjISsq80M-${}L<)f{M)P00I7*VX$Tc|Jx=2W)--OrMe0Qt!Wrpy?`EJd3zS5=~m|2aJ4g6 z$t-fNr=g-dpXQK5-FmySw?NmsCZP9rRsRKMNMG63JJ$wFEo0x!_1?KYe1(;UQ*&#q z9;t9G(A(SBcw_83yTa{~TT{)%O}{&l&-$8fPTk_P=csMFTl@4*#8}PV{=0Rv_qdCF zm$u!jfA)~t{`Oh_J>$R@9s;lKUi|WY=p7mTeFKTnw_`Ijw_SF>AJfV$c$N2b;QpQ0 zL$T{J?zlf_d^;9iP*l3?#J)Zy4Cy~(_@=)yrkAK%L+iHo6zB)m zQ@zVKJu+Ok?Sk>%xcTWj2XB>p_&LhG(-pTG?D07Hy2%}9Z}(RR%g)pMyMBDx2VHCk z1257;&TA-}rv3f=ntvRw9(u(bq2e%bLTPKrz{y!B-5~$`RtF6)$*EnJV{&cmpC5f& zhbCEkPls5T?>>m340xwmzBOX&;*iaUj_DzrFFXgnxl!92w#|w#O=9qYto*V zyVA2rO2?R_Xr_Y10FR3=@eq+%DVpo`@_IzK1q^q64EzFag61 z2-fNbP=;+sqWgg9OYvw5f-|l;z*LE~VVf>tVjWNfrdWjX1gwdsD)^9r`a%Haq%Co6 zCk&`+m<`Kec>=P@Xz^S_7ZxvT&=8Pd=$-I5JLD7qk4Pc^^Jr_xpxHCsukKY(46L`; zn`DOMD4sO%>S5%%uBLWV;b-rA3NBVP8m03SqKY@|Hmmk+mxZ0}B*5l@Z|gxuUrvt-^j zYWrG|ziF(xwpIBnuVTl)uIn_1iG&EKoopi7e4F`cx^LK~Q5xa`S1r3rHhS+qJZ#%L zfN(~b)CQ9N{XOnLG{a1Wytn>O;!@q>EP_7{rM^{i*^qoN&;xUuBf6e@q^7tzTVgg6+deX> z1*$~(RlteQy^oa@WwXHl#8|Q&$u>SR+gKWuF&K6xc9$GaUOJzfVvwdrOJosLxQOFo z`}kV#X`;Fyg=ikBRwQMisGPKf<}?|$u$6cXI1}$4qUJ43A8ySB9x~&xKBz%Q~xR$<~K`YrMa67Nm9Z(Du^3}T1uaS(8^Ifgefa+vmOkz zbA~Cd00=}dmwcp_flDn+HJb4xMY`JY4`rtkK0#z%W$8t+99E-|wH+I92Cz>|Rrt=& zi?9h%oqmJbC~P==n3JY@(A#-q?Y!N=#yYQc0}eV0XLA^wa)onBX$Dfe)eI^RponWw5~K@dHKpOBmhg&Z_k5iGQ0&tva73HqjM zJ4!~hYcVAVrXmL5JSLGWwSW=$DVzW5HMz;j)DZbFgPCZ=O@^ch!)mt;RTvyf`j648 zl%0tQe{J7>Tjmo@jje%5BF~vX-^7nzcfpskVdf4ZN<@A;qrwX{P&*Il#Syf z`OYHp_T=Dfb@=J%YE3w*8UViN^KeRQhq2Q?fbX%zuK_Pag}OM6CK#0&!eTcsXkHCLEZnH7agOiGAqg4bl+q)!K3}& z4M`gz(lh&O3q!-F*v`^TfYxgSG?s<-*m_}>R<#JHC}@|jX2T82IOIS!%(#q)R>-r% zTJ>ZlS=3N<)?Key$6zV5=R3|>EvdZ2q$2frFanzfm(x9hG=RYP00-KVQYKoRnQ@Z4 z8})B9e5YKF8%NVB#a=JPb%TVZOJs4Y7iZm)kcitx;CBdY-L z+nZ^`a90*uz|FK4rEY1LxMUpNPBhv)u?9hEk0E@R zAJ?r`#bH)a4)9Ev7rg2*-tf9C$RL<7{=LHk(NaEjcM7@PLM*V*#dhnd*O5Y=HN zFB_BcTrDX_`Lud_MJXQ=uXyJWzMKRjzb)f3<_!`fAp~RtS{Q-S^p_-R*vsPBsT8 zyQn=Pj~@*mC;y0M{qDm0s<tp~58(Y9k{JH{(t|2Y>aQE+UBPray7;XU_ z9o|6r$w62z!qg-%x3HK?*O^z);0G~GRScuDu^b+!i+A{4*Wpk98TU)bdW(QJ6S>bIeF0+s>EMztx zd7DL!{SR(mx4#xs2vdOrw%CMVYHkQp;m%~h4o}3EV(cS zkFUr>Sa9_u0+35U#)MKpw! z1@RC+`@Iwv+K6KdVX8b-1Z&Hl$uQX|7#2bn31CDHf&)2|pp*6A&yV;JpRgG}T=IV? z!afvUm=qG9@`y6T3a;MB)7m;=BaAkB3d-f|)#Qu@KFr(SESV^J4N3CN$3{ z|KZTTi^*>@$UpdGDUZx@BnvhX_v|Wp)UYp~iAxsbgVKbbig3qjBj5)#${`hTP!`fs z5w>hU8~ciF@gEI&VZ!y{+Sl9#6w(lA0A$6JH}W`t*+qoOY%~q=O?)i72{_eB__E6O zP4>0A*-rLt0+ji{us7+LFZbBQIr&AGF(Kh23!9M+?2_MYc|M59J;r-^g6@6ZTLgGo zxL}#l7%A+x#P7w^kn0@sXo9agWO^8TL21S!QnPr5XTTI+FcUb|Ph2>DgSgJ>>N-81 zPR-S)!aWW$q!b9n6{15RR0tDYBT)a^s=n78dd$=qlUCgST=8|J!bYVsOF`|(H_S~D z+76RQp@BtQYV*-_^4a1;oVA51vykW?0&o>{;s+kel8MZ|s&XAbZh_WY@=+i7gwiX;eNwC)4SXS3Gl>7Tq6gpAzN_>&6jC=iLp)Sb*T*zG z9}{6+_)c2fQg)HnHk<9~B@?S35 z)r<67LjEAi{z9V-079%jaSTB!ZOq={t-m55{}hKhX0@ASz@GGzqor_DYm`kA_}sp`3UBx?Izsoo8?z8k7~H%3f1XIGOeL@wyAa+ z9y({jLeNk4YxFP)yJMjfzcf^3HB~F_tHd_9^9jFU_w)Lby%=T#LK9_8Vkr(q+jU%n zU+ul8sac=yS?S<@_HNJFglrz>uEM>YdK9+`EmLdxI9Zu5OhQs3+GfGyF3 zKG%vVYh~E9X$7>cOK#IGYtw6KGni{L`rXE)wwv0tn+LRSNN%?*Yqug$!p>Db``pf= zcG%l=I0ke$M|H%9w~|NNUFJGGes^p)fNTfaD`1^F13I=EP~4L{1LivSxp(YtX%8-g zDpD{>BGi?$U6CzaQFC2~2b7}(=xlbvxG8aeUe`Zm-AOIoWeuoADf+RX0|z6`3$Vu^ zbaGkG&<4d7ps&vNgygl)OL0*w^a=4sRg1pLfWB)jm=i2? zloVG&wY*H|yBgJZ+or!_4s%j~4g|2N$q*|Y9VJF*u=?A|sMWQ7wFA(l-@WBF0|Rpd z?Jbz&OmwUWU)R!WHrE>@Mn7>H$dH!2dN&}k=_?jvk3*<#_pV#~!C3=rk~ij#e}8ZJqMxhYJLH5|DlL5I*DsXiO&g^hNy3G3rVKd-jf@8efs zqe)3TwAnMvW#QVyjN5-im(j#JmqzvLAj9L9f1>b5g{UiRT!Z_V&m;OJAv%^lsQj_* zwG8;vlXj@bfa|Y)!!-$kEkMFG|ptM z=Pc$n2Pr3gF2{oFGG8zNL<*N$+d?=e#!d;Ib+HCc)B0eap1s`ewV?K#vVEp6#DO=sdBKZKHRiWPbgJo09&4blbrvx=TRAwhuzlt-Q~bnt>GWyr z2|n6iZEii4)FMO?Y_WMfgMKwwPu-lfH7`SoJE*gxK!UbqBFtFG`x4|PDSQtfW6eZt z5+edHJt}4|REqMB7vOXrW5!si*SHw%)#v!F0A3QqA#pYjzm01w%YR#X z2v88*y&(pTM2K7|aufTV1NVjA`0L|lL${w{2ePpjY1j-F!Ho&lurV2%!7QUU$@>;c z1i1K*vjyDYS1fe((9ktL{D&`W;Q0p>AO4tzxxqpj3lK*oxGWK3w-lEl!5tL9-2`}8 z7^Z&TY$bEYE8^S5nwJ4X^hyYgVu5e?xMU9C#3LMYz*SyGmESijOoC+TRdhyBiSE0{rRBsl^eTej_y(K>HnTKca(U z2ps#>;4*Ar|1-;{t9vl!kC8XnpEU)bGY@IljBsMYNsvuUbc|4X zNrcRkzE0txw{qc+q{}CGWGf+JQHb6n9QxAqN!pBQVj)XU{HG|zJYXReg`go5ag+tR zNa2e7N8&kj0PW2Q{<6m(!A zD``ZefS)aZrKg4jqNK%}*YHBCtw4#h`-o5MquW1=a zzHR%qklkQk>QknBR)FwlbI>QoVEriiBc>_p8}&Xjt#36))@;AkZ$}+U zOYTdFQcO>eED?doAu~4%+VcD=TZ*Uegl4Min4K|bUkdBM7?>FmqGNt!H!}2mEYEfd z=byD43weGd`;3d3f!0=5~1=*>B6USw_xW6&Lx>m*TxGHY z;HIwhEUnwFn)!~VTeY*6@Q)pN_t&Swk1V{^@d$p}26}r;!*6Im)=xnv#t22J3Xx|+ z&rsT2on*GVaZ+>ff0cr>7(0~;&CH=g{d%}sa8Px+X|5&v_5t|YPLE_C+lg#V13?>nT~f0E zavu*F_ROJGVHGGCK5ZXG3o0jg;!IF?=|&M(FN7_%q{{8dE_sz7c1AboqV>gQ9;D+Y zT7c{C!_d0-*c!OY!ndIb|q{XOO&-mQAi| zC%n2Je&r;}Dx&<@{$1xBQ%{+uLN&j)7VcoC9m?9VnmwPIds3)N&y|t7R<gcOC|S)|DM zd%KeS85XMfdfS_<3{C&$h1#u**kFvhUymksIQP5D;RnZrv)N&9eC)DVnG>fs<|Y>e z$jnDnb5d~x$)#_&Y{GOXm3og+1yyqVNpEg=QN1?L{Fgiy9}bkabo%^zX>G zq-0Qu0;d7{u=bb~f~Wdhtz4VGyW27jd|{?wqpe0lKoeK2W10H?=8~M9e6o#D#I{Ma zqvSBtl9kHQk3^2@JEeCO8aa6F;#NY8t`4{KZ9kurw`Z(cQMMipr#)O6VQ_y+$`3yFloifsJ0^)*g4Lw z3qCt}4s8m}Z@POOx4?v?qlWpR8a!v%YNVm9mAv{1Niy$H$Yx66{b8mjY!m;$M`gh&Wt-4 zJoQaK(z?DA&*Z2228)Vn!OQv^8Z$KW1f(gy6bh7{XE)yt`fxk3j%J{V=oHpjwEF>o z*4Uc<@7|Zz8bAA(IV@m8ob_uJ323Xe@H0E70);yIYY~7)9A5T)36c0K0A4 zcXN^u5h~XvXDH;YHxQ=ot(G`CFfT^tn8ICuN0PPu((y5Tq*`7p+_gDVNc~cwx^w)h zVrPW2#ZQ@iwLzZ5qv^VrA-KjDM~uI2hnChz?!oL3Os=FA;+v$)4{%5w5%$OeVIKW~ zy=}2ls^Wi4aI0FY{KI#!y-5PG<~Rdum7)79qEaI{s5hJ!K(XI4y&9aoO?! zv&mlqm$PhB?9E7#=PU#`@ZKr(w1aya#h_V4t4yRxy5X|@PD5_0KCTs^mz{>&%I@@v z6^~zcj;~qm?@9H&AnsZj2$Ff-oW1D{>p~_IE@$NzhI6#`Cw!+VWW8~utc`C_*(P_T z(TVQg0P46`U;@O*Tt7azw|ETcA*hvq_;cbw`b339R^x{x<1sXSkWws7Bl)}J7&Ow* zvA-P1(>;7o?G)rDzYbcg1P;6$PxY}(wJZ~Xt?>Pret>KicQc*Y;)BH**wHgtXe)XL za_6)iK2uPwU<4J5O*i;hnQockfNPlF$8SUK5@#5>t~u1- z8A|kZ2aWx8u`(g{oi}=%G_&mpA?6)Q1`|87FWTXDHK)priZi?yzGhiPEVbq`R&UDd za#G;bd5l|6D`+4r&NQe{=;xFhCBM6HBYROkm)2$P+CT`QI?D7%IMKaQuBirUc0ZVo z&R0B(JFo=bZdvN^O)uum<`3onTxB5TovkBA_6N}nhLEN{t^9z-Xv(dKPDYXmRR9lJqU8euR%;RHxu0MfT(vLAulDQi=X~Q1P!YyWrQu z7w6t=><@Z8pRqCE)i0Bd9E%aO4SSC9nu%0>eEO83PwN|(Td69NkT01_bH;`*b_{H6 z2|l&j`uf=Zdk(abY@N7g@Av+j_DHqup9>p*za#}U-~Ug7ERIchmkXR3K47xl@nES?-4 z(itPmb@nW%KDiv7x2RId?9}m(CdLsezS^30z7I406?@c9e(`+Ire)r0R_Ue=xfbAr z^e8OHe(3r4b8k9c2PRTR%nB4*d7BL@ci;y!p12ghd9<=AA8V)>_}``Jb}IKu;+G3A ze!ouGYPpY+xj9rd{1ofxtHoQ#e=ZVai*fDxU4e5e_nsK#V$xYk;Q5gI0HI~7^vTaz zn-BPOV^^^4PIjmOVbi>&)g&F?!Mg4(LUwHVGlp3S+49Xy#gW!3Ma+vDp$6uEU)!JSV_}BpS9?^#Yu{=;vmR#80xs4`f{sQRQeWinlTT=%;koP z5F6g4A>7+E2yOJaEyARXwSZh7bkGP$-&~xwDcerfvEAIKef8@Y-cgL$pp<4=*uJT< z-Kw$Ox})7@++l+l;iy}$T87$~?qI9cVQ=ne8-TF+jtJznR54ZD7SN0dZ%Au=kCf*|HjTe2KCP4 zn4M*rdzL!`LFd2$TklHDK0;SWZD)WSHpH=Oe_U61rBj$s=Yh(u$i}XN9bHl5U5A#s z4lj5819o%ex}&waW6Zl_9lPUvy5mE;kHmE+q<1G4b|0C%2;~Z@lO9QqP&?o_w&E zFV}lktM{CFZ-HZPp-*p7Xz%&B-V5oy#f7~WD|<^CdrLcd%f@>zE%jbr?!5x`3FP|9 zwfe4__fT*h4x*G>${%bS6$e5qq482vF~O_U+s8b-BMrua^EemUntk#pw)ld zy#J14f1^)-Q)vI)xc+F=nNg`iW}%oALuC@=&c;+YaHnB7#J8I7+e|{S{@h%2gP!OBU*!_=7VF7gX2Df z6QP5Xaf4IogVTkBPuGIU#=+T+!MX9lXG??6mj_p0fI||wp?R&L1@oar$Dt*kp_idU zui}PYrw_d;9C}+h^saH}eaFy;@u81PL!Xw1{so7na>JjshQF8(e{~#Q_8DFY9sVzF z_*?q$_rl>HmBT+9hktbp{~jOyvo!p7d3Y5R1M*^+wis?9MmUKcoiJ%MtaJ5e?)hO@5TFJ<6~c)pQ!w@*UL<8(kMas*^FQTQs`9YE-XjRKIi7 zU}Dtp<*3ohs4;SkDL-bSJ!Wb#X67`uYVJE`5jM6Ver#jLm}Sw}rm8WkrZMZzF`J1o z+m~aTSH`v=$64~@cG~0i7UK?1T89%WrW5Ta!Vt3Vqf78UC&WV7DiM=l;_N`0= zA}2ZWlR?^(!4{JtPLrX&lVM?#`{O6WGbSU7CJ$6iMm9|z?3|36m^}1y^6<*!KgcPr z{8Y5|RE)(`tkYDS?^Jx))RFk9gp8@gqN$@*Q%Oxz$2zByC#H_SoI0^GbrLzvlb=q} zo=&xxPIH=0_npoNo6d}%&dQk1E}CvQRW+T{G@aWyoi{Ok`sMVQmFaxsQ@;Guv)WJ3 zSv)OpdRplFv?%Q9`S_<7GM*L}J-t}Y|w&RWmhBGdDYDYA0suUe467%-llG3gu@Tv}bQy%-(UD zZS;|k6z9`UYUJ@oD<2`Aul8pZQnQn)deu}1M6{ZhLAwAsPu z1Il$U!21=;&hCE{&$aS_*55yZ0T5=@e|h`Ey8{TWpqbY8@ZP^K9`9wP#7^J+yx3DP zCTD_1o>dc9Y`bD~36ryl5up0L{B>5!G)`W3HNsFVN=$tvuOzqQaycB0kjcq^!A?&9 zs)(N3HT2}k$#VmnT=vG<5-^wIr^70<^pf%%HMV_(3HI3|;yA4*mG&3!gkyE(=xo^L za1I{ziN`r2zruG?uI0D8DeqpU1u`ywhlzkIKVZFz@XiF3Rt}fVXYiM!7wtdW!uyqh zWJ{KoFk(b21HLl9)!>Zud%45s}4>+b#IE>+(-ur>5&lZ6lq}E!+5d;E~ zm|+`^8&d=vZyXyo0UQN^C}sCmT|QhnPXJi{N*t|$tAz>!(F@~?7(mHStY3{#=CIH$ zEfwU=KN}JKg(xM}o3{fcGW`@>nVub5%~1pZXTPtN&=p$V>Dy66H7rmqYU`_(<-2mC zM;n&Pvw&OA_V5)PsFemu$d^(Vd3ys-M~Zdf=gb!MAL}oXnIb+?gX1&6jh`x4@`UX+ zNbR7wes9Ru(9~>jRfr^kHHIraU(YarH~=ucbVB9b8Ua-!%4imc>@Wr}TcP70kB02f z=}>5lvNUt3eQxx#WszvqJK&IC-$P1=0mql`oZ;^sb~D=id^}2=wsQC5hV4gxOC8ib zRB1@7&mCzblzGSMRe0G;wYXxX&_rPx!A8#00fRN-zyUJGr|fcprq3LZWMpye zYDtb~(Hh6h%e&Q?a*@H^GM!oJG*{OJj7+07qxZclRoT*VB{7X`42Za~L;q%~m|1(x zs14bCgNP!EklMaY%l6!CzsZTj*nFMOPB8@22G>ogmR|^^hfKTKs~3}ww>XfE5c3N= zECata;7wat@C~N_)}|2#U4)$`XJENFo2?oL@RxPYjO2P$>ukQn5olcP*4q8(W2(}U zD;%WVv%#={=mOOzkGpac#deg>yiP@TOby-j#sLRG$`Zh8mXV_)%As^Hl$j+5NTB*= z7_fSkk4T-S;o&KO400i(oh85~_O%6-*`YM#Dr2pyVM^vH*Nh?;NT%#(<+Q8ov;%@E z=X#us_P@qhxCK)#u;IYCxzUyuO=?bah1__W*$#QK;)-9lZB%rH>v)>lAVQ=lDkmTR zu&CbRXIK4cgnY6wMD45C*Wld<*)7!pFGIV+sHBmSTNc%S@&Z3c1(BnbfPFZi&4?Qf zI#uH|{#0k_s7L1~w6ts7T?>4Erff^rN(;b%JcZ^nh2afdo_d8m;xFL z(=a4v?qT36Wh~p72Fjo~mI%4s-!sfCVJW+9(iD6lXA3DeCz7Cv8{#@!DELzi+9ZgI z#Wc6A-(X5y13T!gSiR5_AEHH5>}pOoi8%{|jN2<816W!6UXl0Z0lPRVmo@B=Jt0C( zs(87}I;nxYbEPfd)Em%x`7_3_u|oYJ8#E1`#2WUbsqOvfw1wxgc}rZI#&cc@qk~S` zI+LgP$F&rR9L+s4y+EiXpb1tc`SFemm`jN+rb7bE5%D5Ef7)ecDa*yFwnF3b_coiQ z6ugsGn)j0jziz=4F?^`BM%IV3j|A1MuuuI~*uV(@xggvo~6SQL*OZ z;b@vd}VY21}lKyZK;W z_}hvhZbQBwtwQ6JICbKAUoMiTsbP))jDM%#9WotgOej=~QJ&8g!4;ZCE|v!3$(Xu! zh28$G;ZfHnL$kYS%r|Kas}%gvT13=W2n_*DxS=jX=CY`gL8b2oz904nvYkBpi$un? z2#u>2b;aEQd>gI_Ibu@r?Kvc%G7Vq-=d2(F8*+1K_9`N=uJ<}#j&sTz;F!lk3{8I+ zF9M+9ho-5Z?q{*54MMJ5gkZXwJPQOwQ2*XrOn`;i1tdHgqCLWK;1c-P6f#V#3mN#3c9lwZ2` z6xN$`%TM$>Z6pzS>*h4rPk@v(-hSaraZFL&pKMB=K=x=8atl9*rwqWTe=uejH}3N{ zdr6m-j3YEmM`+WSN?Ds#IL}HABFoo8ig)Gpjj6oaWQ7L#Jeo+RJ_G?s=2gN3PWq|y z;T~6l<=U7XOkGbL%MbA|lj|&dm3^9Ri69RFFinHKqwH<3E03{OZWzNrrRJX#%;T;w z6p*=p+)9m`|7-RZEE6Utv414&_HL(~IWiF9U*;cKy=>|zT2=YrQrlGMK_2kbzQbs|H721&iq5#d#OcZLEL8bE(O2Ev8@LHnk_4Z>;KVT zAhdcV&eR(~hW{>5xJAJ9C*ZoQ>>1YnSyQ zB#@%NskN|8M*@&a^>(*sC&8{V=iMPBA4rBr|? z*@yrRb{h?0L_@`Lm7+p6AwHiXjck$&+0PN8FRoQ<(ye5(kEnlz@N6KBPm(f8F&9XS z0QoHogA@XSFFSh>K*<8;CztHcC#a1R&U=-?jSME)p-dGVru0r*RR z;TNf|G8((e4enod6?^5MYEJtqyLP+Y3CpvgUzGU*k3IpeU*IvtjeNI!mZp4;@#E}I zS)E-RbPNv_%RwKJpzI4Bv*<9~c}H1^qs2uIv-_XaV^(BHY+VIDBk^ z6vh&OnO-SSw9*Jb_S#hBeIO$1vvQ6Q|G<>Vg76Qe$g@KHDXylU@=kd4aRcbsLzZoy z9R86I?;fjlNrZ0^tks0^8{hz+hwl+50W??x5AQyTZWgipCRMJq!h~Y{apemD6LI)y zQZom?!J)W^WgGhxp9A0nkD?y(!ps1C+*_>uLMU4Sdz?o4k9oer>1_Q|4H^^l6h?w9 zgt-XVe+2CyEPePEeAHX|_$}D>wzTDG37%bm7U^i46k4!p+iAM#48-<}M`U2IIstK- zf9dHBVx6E&fu55RYyNqaZhS1v?#lO58o_{Mw{6rbI?_aajrsalP_#>l`11jD?O+iz&U!RDa$}{c$i1s|Qgan1a+yjYMKZqE9K>?kp_LibOSw@c8=zfjKOadX#?Ek z=cdZ8&sSEl0<0yi&QIW>OhNDq$WWDMNx}Q|it@D3Kw72agQ~{pN7tMgW8s?W?+$B2+wRyPXYmC!g~amkI62RBM`bc%l75LoT^?Sm>*YOW>KRF~A$ zJgBLisj2^2BUHJ0+v;Yc|INEeH=9dtK6r4mW#;DNcTlf#ZJSkXhktEXQf*I3ZEGZ@ zZ>DzWXKhasMQl~qwU*3Cs+%sUYp*5Lt*iEJnK1E}>-Zm>j&V`|BK`;r3dXDTXz=J#2Kx-(L3+^^fER^uBH*2fbEz zH=?OIi4i6egsL}#^M5HG@jP5xj6A|>#CkUEDs8gnU0Y`l25!3>x98p%9YJTVT@vEx zu){2r5fA>)EL?@%oOu&w#JqQ84`YAs-N;-hDz|AO1+iP7k$dcZ@xgnBYiMQTBycpf1;4QE0~{WT7<^uBQ0Jv$a$=j|ns{-E;U1Mvu)%7#(JaCOy(ZQIlw z`OPP#%`ONaf``fsc@U0ApTVJbvEdJ_TSiPCHr<4p_dFOo3*W{=8m&sLVc#1lL#Db1%LN5fAoWd_9wrO-Ot$r+H+{SJk&Y8 zocFVjf2pFDvy9@3Zvsu+m={sMpgdIGpVMqO>D z+uWUOK#@uxKd=H0vXGk}BFwnx3nJVNsH>0lBtg)$VeyVB6H;JU_O%F`f2)-^x^8=mhX z&={pH-6yF%X)T?5TM%|!L^uam1YxcTu{GYk2s^kF&%l~<6%YabH=7djddE%fSsEac ze&20rKxE!A5$X4h&BJ$bP<|~pRRA>j03la^3u__NaAvBJNapH^Xxwv3{zoP0|pB;DyFgohm1!tVP2&5ar$~agRAtKsF%>qEH z@Sx8&!@EJ`bvB`Cm5mp%#~kJ`x}M)&8$4QeYVd(EiuZ$7SH2t zY7^s3IT-ubxYNubCkZS}?^5c=CtB>8leV44{D#e?%_3DekA>H-J!LJ%g$V$6F+7Bc zG-iWp>}z8UV8`d7`+6o$QrLgV!{2}7{^X(G=^)(M#AiG-T>|sig)O6D*0YJ%S0UUK zZSA0hs}mE$I1auXWHbv1WuerB_YBZ6GNH!a-CPF4cbKm6Q2c-3T80c*@we@Vm!9$ZyCng^QupfiO^DKNj zFXF8Td-58rPD-F&M!q*hg$V`;9MnE(zntqO1gXuKjVu-;HQ5NN5R*xR{lEjoOkyE` z3j;87JnXO-7tg{KbBUKH?lLtG?*jdbUu>{)}Rd?XuKuwKLT zSU`|uC|Ha!Wy7MmNN)f+LBoV`EQgsm84-df1;-84OgZSg2j?2}Oq_+VuQv(VV#c4j z;njeNI+|xH1gCH@Lu_oO0C18LG9mN}ZKy&R5bA~>V-njXgeexrTS^#{684Lc&rT4Z zNMH&t3A@C&N;aDAgm-yl6>LN7=-B zF07SHtP*1;*u+XErWEB^1mS#X#9`W}rN4w%>bN<8kSQVzGx2o6Lr2l#<`WR>#3L0J zu}g@(#>AE~vA0)+n9cD(yF76(8~a3r?32dSqL{GCFDt7 z)n`L~J_uhPIsrlj0jOdYHvSI?05}l`i&jFX1Fr{YkUNtY55es@polHy(J+$|!XyVP z!^OGgql=i|M)@dRCQt?7yhX%aB5VZvM zN>FY4X?8ZOE&_(hJ)n}#ZdZ!F*;ZXbJDis4i}(E~VqPTZA~g$#_@V=qJkzB8J^lsG zbD>=8^sGJEK}8?ptV62_*nI2I=}i)!mB`7uy-zaL_WZXRHPdwHMc1W0-wr>2bbJ-1 z9?%?F?L}94%H8_(OW0DGHCndazA;kppow7uo&|6>w1ZgH@CP%qz&tKPVH=b}J|I{`~x9Ec$tsrmeE8SnA5i zG4MaFnV$;4R3jtTcggK`zwxr*P>u-?U}rYK_yCE~A?P)>}J7a#hZV-xR#9De) z$Z^AV$^*`ukbq*F5~_GL9oA8@dw+QbCCK0u3NWj=!6-4mmUt~Y$N_XVx-Qialz1WK{X0BltC5ZY-1+>0JA|W@L=X=66A%6t}1kJR<10yGH}iDxb`8huXUwGpsf;I zKfiAAfDl%NX{c?jkUPkV=ujEk%$R@BH&dKR0V)q<$Q*8ts3^66{r&>~Btm_{J%4^w8zZp9c=D3L6E1EXxLG0+4Akw(bf{kdPW-Z&F#fdwsc9 zp+Zh=L2LF6crbap*VBR>#v&D(9H&az>D0j)%}{_N4!MfK67F~LIsTNIvpcWS+f^Kx zHvE^TjK18brhL5W6!zlIn7E4Gxelf9T4=F22$^6`;b)&GHS;>#}D8&p9)?O9r6?%SZ_dqan#g zZWkqxhdHhDqIYq`yj)Pgyd&Jb4C|cdrn26lIEZ;51+wu%SVr(H+V#~P)5WPwO9#~s zT^F!F==5Pe0}ZW5{slhsR@rQOxLS&<1=|ISDuk0|vJuP=ZOpC%JW8>c_xnxwmfvr0 zQDlMiRo(fNpgI?#C%{_vxr5T9MWF3;VH+A7Iu03ls?NH)m;fS>{D#O;69YOV6PT_$ z=cKIe3&dJBEnFmq+LgNKiIyU`IzO@MhMW3w$MyQTI*e&`o5AII92sa+ZiJlI7jknp zKp!>LnqL?ciZH%J1R22rRa8??Va-!z0tx|7^)*l^`)uBMSBLdEVho#iRfVmap>&On zql5NJX%QJ_ZZM=W*s8gWm)4MaZ`;2AMlU=8_cO!I0EP;MvH815)@X#f69`7AN@0%0 zQg9t%RRA_(<^VkakdsS+DJiokdeG1)AOMXS_DIYc3B=OU1a(uPZ6Y3y53HG|3DrUr zy(Y8RDcyBjAB}hz(o{olcd>X7jr%|C-1}G3cO1uY5CkvjATcX-k%HMpGsCDXfyAt| zGjpi3s}{C2L)=i%P%Jg9Q`tqcRok$Mw75 zQKPVp7d}f!$H<*1QH0YzLDHvtZ2}9Zb=`V_;Fpl?W9g_88vb3l*qJ=D_sMkP_8`y5 z*J?D@EAS^vBs+tCkAJ|iPhws5myT_rzj>3*}t|Ofl zz6uP^NGy1d%s9Cz;~-vSjlQRdoRif42gmP$5o=*Rl7YyqKht9ROPSsJ%}~G z|K9q-n^SS5ezt!!f#X+oI?(43)706Le@Ik9w>$6LS!_!eVnaBiPq|KAk2a)otm|Ur z3{_cc(F2j>JB(#yAgAuzL38=}McJfK*xoVi7`}&1aB3E{T^0Bx#nqd_^6PZhR``|S zy70T1&RX@@x+}%U^edDK{H_x?E06kp0&#cu(R$dWRhaAF1Wi;t;<~naL8`J_Kj@7( zw))6-o!{JJ$`m&Y*AGdD6gK=DfsW(+H(7qQ#kQfnVx-%3D6PZ!@rbt)p$8-%x$bY+ zE6J`=)h8bp_xyIh%3`)DhIp!|$22cEttq16n={I-_In_|eQN?W%Ada@B zIHXn;=?8Wg?>}xc4Bx!2|1(K`yZvNjN?nJ3P~5oRFx((+7|{>Oh|*q5mIRrqc%*%f zDv6e9c$F`gn&_;wrYOVY#i@!%YYfE4{VIU@i~NcoCsWZ&wjTwU}s`34X5yOaznE5i7CGB%EzH1E5d|jxL_A`y|f0bvNOIaB&qKqFp8Z+PC zQe_N^jiW}Q?45>{IaTol%RaQJGT$nVe+|iHAD^;hA3BU3cKiFR}(yu?aaX=R0=YehPW{hcP>KXA#^_CAwuTmAf5%t z{1RkAHL?&z7U>W#6XI<~7UR`E6!j9C+LxhT`cK8<$5$^CshK(IPb(hvYLhy^tY+a+ zHU;HORy=4R7v=KMAU?WQga+rJ>k82ICFq7~Gz3LMb?8RZRKNRE8P>X;Lx~5UBwvp(S}71XL6hq&Jb? zd#Fj22uMeIP3SF@00{}HXZxK0yVg11&X==4=NGrITJqmaz4xf7XA?k}hd0lShDu8TWn z8HFX5n>)_8>zjzR^5ipeU ze}sYM_po0DfaRR-i{rrZD*MYXz;X{d4}AWAxeS-^KfsA0KH};TAwE7%_MxG*NZKK- zzM(-KtagK)`QM+bLKhSixOBZ6Tzy_qQGvtZx=9u`Ha6Byl2>}B)ymi`Rs!mlm9zbS zpS;TR!cFtdogLvI>6G3Z;btLJs{BS^*d%qKLP4QhZ~5xaho{~adv2hf;)Os!4_{O7 zEo`(W-ekZ3zi;-up&=+sQnFj`Vz#1=N3Dpaz|TL}){~T!vWq+L^|lIW!69#t_|*RQ z`*S$kA*riXWW=|eZ0npJF+7u<%)aDp3|w+7xL@vu-^V0Z-hx?@4DNt>-AF_tfLs7V zY;W}5{CPs_e9~_kmL>V0QTB(}nD)T?z14x1KVSU0I=Ml;raYMdu7Ze6H3OIVLOM}a zwoLx)g{X*Jd8?qV3j!By8ZRbrvM8lJx05d#Yg=9m&rcxjWIF~k6o(zAlyht0mh|-v zR8~dCCDPo?k9xX}kg78|$2uWDXvJ!32fMBNw||RIe;rD)e)?OC|M+GWn76na4>&Sw_fy;Fk%?eecZP*ZFw z^ZzN^-{9TZ?w)@jImI<1EguNxhWYXT3H*uByE+pm8ir1@tSAx!; zT6x0xXmlV9ah?5k+Nkv5-)D5+b=URE*_Kxhi0(@^#~mo?ea_E)ulRB$=`rs4ii?OY z_UY6ceYmaI+y@ZLqCYjQV?!;Tv$=q1bC11jeA_dg*9*H?1ZrjRE7}-_4v_DeCVEoF zfBA=P=MlozDZ5jv0R?voy7=7w658sGP)5;!uQTiKeM=4MKBlS|BylsDjOxEY=6|mw z`wtZGZRDlpcvNl|fsxW3tNNUOLAYh#%38WekaTBznGwy8_NGz69lZKY%*NmNqYfq4&VJErk^l}@g&X{ts^L36XmNbumV_D)WctM(1g zJG|QbUF(|Pna!nSsjM;$607+2{@p#}=fKVuLDHe2zZ%|F^XRN+a61Iz^$t@vTUdb! zw`CRV3I4=6@GOq`;YKdMPd9Md-(6Ds`Es!%OdWQsz5zQM&X^uwhpMB(}{^ST{mJ!OUc-~(sf1a}Fo&?qd)~106QBi_QTlMsOIb0Ti(^*-*HHJ-#CUR{ z@!2Uy52O1`8#n8@%WIi)r18TQXav5Yz^y1ALjQ&o;@)wx)H<9G7}sM_mhYcB?b73q zm|6WBJzcU3buM@y={G#oa(>g2@tsXgu^<#W! zJUhbou^yy3pH_K&3YRGE4Rv!{5k8y~zj(!vTMh4)gK8yeE)Nl(99OopFjX$a> z<39Cw38nWLT6 z2D8{*%pw~!MQYaXx*GVv=)ReMiwk$r5YLq{5IL}CPa<-Trn+L++KY=p4v>vcUq?R0 z&GKJPt?`|J*cb^NZCn1@MZ6HyaPxS0gsH~jB6TrVH68-2R*ZU|rX`5EBFS<}dEt|s zo;BEGXI^9y`;<2V;dLn_bH?wuCBbXnW`}!n^CL$JR&YKN5@K_$tQR@aDO!QpA-$)W zydoD7O|Z^=W^Iw%Tq1MwwCBy~JWshE6B855q?L`H zp8XnW&jF34AS3$6_wp=o*~B1znbRC=6Giwj^LVCJyR}XiH&wZVck@bb|MuRn$2ENL zwk$5MHtgV8aXu|NE?K)4C-yO!`y@sQE+Er7yM6G_4(7`XL&<=x5$MjkmI%9_gqePu z#+VIq-=D5?Jt|C&JAb4iO$rERzhkKP!8m8MP+NH6t8Q*?c@)nPqg?G2HfcJJGTmBe zldm^2%5_xd%2;_NB=8hAn;myddS@^{NNQDjC^JLk%=AF7VWHI#XT7IXt)Kv>J%T{Lwb%D~1VvN=jp(cZ4DvV{QfX8KRwZT%|9I7su-^w?(HJA`VnULfK+6S3@`E zQ{2>~_5VE0*YCFvUbF5FHq}1D<;rvRl8T0b>~1TCws>_GiP_AIqt8rQd++-rBr@@g zgA{MnZ;1ts7_Xk!yYdy~Caep?KRUkD@T2(3OK%}F`>&&6%A5a5a06U7yZUu00erQ*p_vb~;G4~)2pEk0rVAa8>0fEY70!98^J zHP@=E&!=_n_YB7T;ui9L01HY=!hrj7wzV6|7b$L{Ftgw(TAFH#5)2!8*<5q`Eiq2x zPLeT~NZ_!=AZ&<&LVt=Qx?}6OuwbKnsP7LE0%Inbjg12R9Ycl8aU=V={-9Kh%YlO9HJ63sPzcG?Wuzd0digV_q;8{zg)B!AN+ z>F(QN^eU5B*!KC?F1@*&>xsrVl|+8LKFrYS;JAcCm6tcW%!m$4*W?{f^sGeeghZeM z0+5DGv76|r**(0h8Z29dhqyV$ePYGtO3Y|UWawpg<7)EXPc6VtED6uF9Sm;@pl!ad zT@jmg;ymv;z5Dm!Xs7$`E%Ud!%yQ-TQf;W=AGQ}`7Pl#vd{Dd}%A=si;FgRD!R%{l zZzqYLS~Q$2EdppTXNVp6-$jF{7bs?9hsywdJ})vzdS>KD%i`;=Ygs_ZJy~^B8}&#l z;ERMF_Q-i|8{SWgy z_T1EFi|eLr30Kj)Ukjo}-*W9cZRBC_?ELW;dD@(yoR9B}9$d}2f(-NQIV#ymaqR-l zIayBcw^I`GhbdMio%n1ygQQc$uElq030)?}Y~;sVvN+AGP<7Lf$xblKX~^5b!|b`a zz3%H{QY{a(aQTgFoURkbxy?!soY#5)3uq24Sb+?}&Hzi;`US7IO`1};ulGqmh|@WInTSSl2MJty)#!wX z_Fg_U?*;<37Vb&eMEmt4v%~iH&MM+vmy0bHI!zP_&sTaxZ82;KuZhajBi&!l-x3m% zz9J!k@uOR3Wo02SjPMvC30~eVj2okk5Q*0WGIvK&s!q_fNY&9m`_W)_$bQMStn1bhVx0^U|9lK3<@M?~x`W zs+&>{7Jr{H=`xk#GCvzU!1;blJzv)n{y^t&Xum{A)N5Q4&0H8suKXL>u#(^djyorO zGQRZIPE`$O$Lo#DM9b*_ppX?2PDPs7j+A1@21nx{zrV%FfEd_AWd`sb?Se$|Xl z%4YF#^D7a>4!omR__$Tq?|PC#Y@!c!>VL5S$qT=Jy%f`T{kWKsta+gC^jN?NC}cab z<298?;xGha{^vqrJZEG#K0UxVwKyl@VaVqq46BB^OM~jz&$>LJM?x?(?F_Ne;^RTtjzZ_W!w{1b*#0D~H6KV?Re<*8`x{P{O_raSHN>0ns*Qz!Jj>Zbz z#hjNCuIiT5al>jB=SB^@mg}WMU1@yTA)dV121bRgo2vjFj_*;1l5R_(0==>?C};#) zF0QWPEp2A$#BF^}6Pb;Un>(}6tGq~=rJ9bY61cPxjP_Hf zmtH2CU)Bn+w&sh=1jf8q(4)in>eTi-01m-kVBlLlJKj)t=M6;2n1R_=Ujm)Luj30I zb6Z&vAz=`wuEMeT{KfJq5Pv>h*}~HDPFUZhS0!t8aqW^q$YC0tG7(gd(3|;ERmq-8 z=opOBsO4uK?}DaMXFo@ElkTFPqYMl}D~A$qmt3)~JFu$XF=5~Jd)auff1Pdz^IeHH z%ZrJ>GV&v$zd}R@9NvQfp~C{9?w-UJf8^|k+ZVI7R_ax7k`uEL$E9Qz2I$?GnPtpeWG zIgMyD?KTYueAOCW$XAqENDT~jn?B@Pv0)ed68imgl9*S=j$nLPZp z>-&_4-&H*r_%XVA9lMh~b zDrG&TtO9^r!7y9oE(XoT7nEi2S0%k3>|#vm#hN9iMwsczS&fqy&lqVreW%w~HX)~o zjl2n@4VP8f-~f44xcoJpwZvi#^dDxYZR^T$ObU3pwX05P2|im<7M5WymRF>ja^Hij zZ;dVoES-&l?3{pE29NBFzx${2ZJMOyHUNCJqEphRryIP$BI)Cm_L67r-{#~Gf(mM% z9Mu|v&{SI8facOxz(lz-J#W{^NzKoENRGtm3i$Qy)hU8N*VLMmACuyoV~!xn{GsNz zrH>d~%~~{kw*#42F1416h3OrPCN2_I#_j zHUep=)h+g1iRht={0AlwcY-qEzKbu*^9^xxpJ1L>6VYwl5;=S6UHR<{9RFia4CuBK zzMOxxRYdo$$4>6Sni8K@~V=fD5T2NxA67)sX{ic}T?ik=}$(AotP>m*uUc9KMl> zIX6BgvEt>KhmfQBbB5R^oIfusGRd7E%kwxq0TNOw@&5_}P#&|HP4|TD?}zm%NwnQ> z7F!yvpUeF3GsT>VA1{zJy?eeESQAZddzC#fo$hF51U1}MY?G1Ctcl6gX18N*RY}@g z+KVN88(KnQOW@(R!D9U6G{45}> z$b35IMp>)gVYZp!)rM;CEx1vko>7voo}B)*oQtdkeJgTj(QfMoJ|2S_TZ(<$$j~SW zbg@4|5()ZnJJ_HbPzy|r?H#4F5*xDv<~2U@d)yiWIuai_)Uu!UWi~uN+ILD*rE|J3 zXlZLlgT002grpxF(&?hkA;NJK3~vPHgqhpoiZoFdO_W+_`yg}`Ms6#GKy4PHOb89A zK3V=gf#a={xvUfsJwNid*NLRPYz3vosV09H=5Db?o^9niCUoJl0sFhyez%ECmhmp_ zmm=uZP*%5eZ?C0p#oR{3MEZD5Zp-y5akiFUAnma7G4_v%af=_*79V7@G+VS_w;nyR z2@VXd!1!*=Njfc{YImC5wnrE_nrLQzL`xd#IP2)FZ($iI@f{RrO`)WoI15*11&A#8aDbX8}ntDTH~N;!%#)6CaU z34G`@+U#$y;&(CLx^ z#S}5z7_^-7_JKL$&{y zQY{$;;6^OksTuVJCwBSV2Pu&j91S6HFxTM-g{7rxstghw^bqE zvLEPH23HP&Lec8h-F!<2DWM-@dmH-bC->=OS7Ryjg2(!2x(}V-w`p9Ef#MIa4oEEr z;Baz66$_eWTuD(x@2tDTV^&LK^rtme>Q&!#RS0TPKSd7eF*s%TqlOcMdLz@*=|HeZ zedgtbV-EKcFm&sMzeu!kg*CXkT3lJFyjc5RT2{kNc$Jji3JLJWiv?*Gxl`1-B+1KIHSJAl zPlRGh&uVKJOT9iQ&U|=bXGXhYy;ATkQ3uLpf)aFV*P{o;jvL;RO^gpWQ7Npivajxc zGF$`Sdo|S+^?G=1&{d`QGde)&mejqG8Is+Wqyy-W523t>KQcf8OT&vUH@WWO&0*_} z?WYJDPY{py|EX1Ycd*Oi8a|Dgfb126$KPhDg&4=3Gmhi0D}W@{J{bp4AR!*Eurq|{LB z4KAKt{pJAnCg$Vlgs@k4dIn2g1{YeUUC-Je4`=OCAn$c%tn8nGX5M@6X678tmJeWe zZ62Y@>p9a9(&J*I0xkWmzghe*ph?1~hdKEziTa^>;MV%BwE{Q7)l{X#Ua71$4S=;# zQ)_5_E*#q>gZy&+aGgWe`Wxno`I=$dBHGCt5P=fj195FTQqqkvA z=4`f-7%~{%1ZIL?eUO9N`eArmPOdT4ol*IUt$!eV6~%iU1Y`dwrT5QF9=YwWw;*Ba z3l%Bvc6@ea0~?brK28S5wXi<~b<1_*=OU$(NO#L~y(g<1a8tFbtXk|0{5U!h+Kqq!3kB&aokEN_TlBVWt!Sb@Qaxb&%$8MMqZCG%F z;H%HN*RGx=W_xbW9=BydNu|Ey06JWS^r|gEP zT0s&W<)qLtW3^$r@~XRIqcaM#&WGFTI+;q)6GemxRgsYC>XH~=|7Q=DINCBGcCzby zb&3p{I??Uha~LR(KayL3@gt$yoG&K)3UO)ulFK#ZnvM&Wf(^7dgh+PS0UZ;(-Ou-X zhGn1F-M3%k1^KVN`e*DoL@2g^cf_@^AtAS2$Xq}Fu)emsU=_c&7<5r;Wv6H(ulGmC zaZJ0_)?L55N}<)_8k;w-YS0d;EAw||seRi{?=Z6?-kkGdN}PVaH}8+6E2w#opXmQ* z$yV6yq;17|gsHp3fl+A@81z8XZN*4SFo;!~=Dn*Tkj&%pe8^y{{f7Yj0<^COM@DH0 z)!W$mD#olg`Lx^0#7p?Y)N5IxcR7u%6og4H8YrH^wNwdJj78J#r4ky@F!$~^mvoGL z*r_QQDMSKCP|$YZKyvabpT&fvd`y1IW+CCsPFr48e*X0&V>;cZ+0dYA#S6wFE4$xJ z+}I#Rn7<$Ti??=x@9b=G9!qo2A@m}FFju?J6RM9~n`>eZ%n%7*V?2F5S$#5m)mb^6 zr+O2Z)vo}K6VD$p6mVxU;N1)!q}@%v2a}}K3E`NsH-)xQQ{L90-ODxB zT~}>__L`SOH2$%JIlz#;>88Ox*!L_BG$E-hBhfYibwy)8s-1Hf7ib?^TGm)uCSy&S z_|RXJtf09RR99q{?Kd-sL4Q->N-ql_S~l`k<6*TCV$(SRY_P!16MvwaqCGoQa9Yxh%My?jbUkA#52_9nzNf-EcF(7KF zdXtyq9WZcC`ef11l(owno$A;CaFVTg|GN^|%?g(%Ek1`V3{$69oey)M&NKWS!93Ok zipqA-S?xSRipmTZtgy2f;In8txkJg|vgRQI?GRyf&A=OFG+{udkL|@U8Y6d-8yY!F z=dWD_>}dUk5`j*G>RD&PsiPSNQK9?+S%Py*a^=CQ4Jw!))$?!Zx51t^00rnzD15Y*;ReU=5J(34QFT5_{CR8wM& zthw;LxBZ^|_IK_8_xryv{@@oY_qY(gbtKeQXXbBgD=s7}%YKdH)>~^h4haeY9~_QSRR`L`3l6EQq(@od{zCN0p%EEiw+FngdV%qqsLOisaw?$+2` z_7}eNuxBR@v%8IkCV1>@Rk?3;1 zgBw9sfe*(rI4>@rHQ=>M=ej+zwM7Pc5B!mM6KanQ>>jU+3dP#%)*3v2pq^odRq9Lo zMAo|XzgWOrDN67#?ypD>7Xd7O!oL6m?w-J|LBzNQ~5 z7i4T~Z3@HH8yNNAf?5gVVcf*5-XRiMajUJ^>49??oAt4%L|Jd}##Hk7>JBMFePU`P z!&f|fa9?P!+;?`!c9>GnPRyQIsT1N0vK#RCI0&HwcgqKb<$$D znSteL*<6oqgP#r4Rr8HtQb z5akFKCE-P$<^xcO=hDBeY-o4rdT2-OIjI0j%lUF9kpc9wpV3=;Ae{dS3NjGfDCE4P zF36kEJ8bpgU1addNiv)f$qs=kNidrkPx`pMvF_mnQ+qNzlmh0b@I`qN*;)~elg5XH z^_*xZ>gC-`E6y&hU;LCSc(>Jp zUUgh#*hZW4AEyu;H{8-RbrSMGPCd@Sm&F}Fc-dkB@t(ZoJ|5B#M%$ocj&hb958qf{VvS7qm)A1E%DX`Q zO=}{sQG2GJ0V3p0tH(t;)~9U=0+sFjP-!LyVzFA|L-@!y=J=udc%>+-e+OW8)hJU)I!A+SPZ8 zf0EF;P~srf8g>BG2_U9DH$OCzoh_*MJVbZF?fOrwt>-X%1gP*AY0J|&^^K@{ZCI@= zcrhK)#?2t00~`$b!lOvNA=O0o{U96{k;~Fg7

    I40_yUAo><%`f3v1sf z`!4;BX#4$E_W0e#z?)S%O2qqKBhC0|N95A9n%S7OTApD_{zz2%wBv2-TNf<>3^+6Nc${oZ>9(T~gLrdqNBDt${5T*ViJ-QfVsL8`9p~BedslMr&2qa{*KjL! z3x=nDxKhun>xHwI$n3HqV{FLRNbS^Fdbg@OjQYbvhJbkHwZ8oXYZUklTA&*ZsB79Dy!`Y|{#^kOQ3 zwQ^u0%iuiWkb9zyF2uD3u2ZI{GzA7`@6FG=`DD>3y9#2+?+EIcdIK3yXQXfVItNV1 z3}~#sfsR9U1FO#j0qi0uc{`quUe6rlY{AAM{zRrUNqO^&6|uG4XH0dF!+{wN>Q;zr zoKf0vO9oUr)>Zji;07T7n7j9w*sR?K**UW+GN!Yg4x9700vg=7jg0HvDd_Cky?aM9 z(26Y8&)U}e6AeVZPN*9o%%OX^?7q!17TWr7?qqvxasy|^b{g&PMXWj4bcgm$*wUXS z)>=zYK(NsO@`Zl6@rI}mVf$}BmP+c~oN8<@uC1+92R}EpDlLC9x^8N3pFkQ=`SmCQ z+gOFyFdOS+1x7Fb^B^vx-hD6!KzK7G24_MsvN z^;7cWeFJbokEQnTlL!bMA|)$jJW>Pi_Tle88N>-5J62^DAmbjF<`A|vGUcA^%Rcon z^65v(J2#rSd-xTJysK09eLBmzHdT+4wj7pZOhbGZ{&j_gUjEh4!pP0d(t6~BteLfq z+nqGmgV1wlj!zY9qz$e-kE$G)uAU;~RWu$XH60i9*W~4aMdKgczPN9K;D#-=LmtIv zt1Nssi8cJ~L@a{0uL@yL0BF+u*!}s*-oY$mx^1gu&bd70>KxNdZC|Rfs`n=RXNfhV zUG{9P(+$aI*oR|hPW{v+O?h4OYs;L7Umv`*07J*{&zuN~OffzfjzO>|Ejmm(g{{B4 zPv}H&0V^+W0U(NVB`Ey7tXM20w z>@caewzj3EMYK2N2K&wbk>3%cMXJh6eSN2q+7&VGTbo;xDTB=CQlEb+Ah(WRyZh1g zj+&ZUx8C{eTRL<2atR}T)Eha*kNE$p^c-8Xv}1%o&Nj~3gA7>x-6?HAUk0#L?U)Gw zFU`I+6?65!KU5`}>pwa^9@Y0@OA@E6>*G_w4k5q$|iKQa7o0B`iF=RZDw%W`-RWo`@r|rd7uWc51&lFCF$v06n{H;-hV< zN_X#dL4_+1-U4LCCU9(TY)~j?F+Xl@l_=}TDJUq&DOHu1Z*6YQj!rU%Wz6^l(Dk#qz3@_YHNe z@--7pDI6id7RoFO+iGu4twn?Hj*3V~VZR3$C9&EB>iDY7ITzb3bDOBp9m>aPha`&xKg^Eb|J7 zI>4PQru3W!FG$UYopg-+wo) z5fzp;Sv^%zTkAPD-JqSVkTf7zR-)s_zZ@f3&;MoQfe6KEqB+@#_At51M)Rgja@5hgN`^-cF zlP(n1en2VWW7vdKnil3BvJ{5)J^aCCf{U(bDn_R+Vzu6I|HuN%P~2nr>S_?QWA zZ#?~w>)X83YDW3v+0}45SEslVd)O{VR`D616+H{u@N}G2j_9*bA$YB3#`tnNL}s#g zrS&k*m%=$Gf94KE&dIp+^z=%}Y*1zG=O>(3=_PqyajBa7)KE4w6_088td)^5u7ZBt z+jF!nVI!sU+2ydtCEF3aSs9g)T!8=>Xz1h#cuhFGW->;KKLofzkR__ zO65t2Q#cXWJDt9EBmLFa&*Igx=d*!fH%5X2idx>2|v|{ILt?8X*BC>Y;1da zJ?7YTO?Lz>6~0To+lrq3xOv`13tWM4s%vuP*18nYC3WEKiDjF<5m=wE!y1QW2%d-O=3U*j=GCq`E|H{^*d=(4|0*ZeZmE8@@b6-;HoNWY zP4jd|OUqjH+%%c9&jRori@XPfg<7!h-aLUnhe(!3D^8uQjG)ZeR5%{R=5SrZRol2u zyEsbR%np}0*S)h&p_GxiBm3KdC_3c1B01qQzND?T{JK-NIsKC5%7#7?`|wu(-A1G> zcoa-mvmT!SH)0vIrxK+SC*!O_Uo1g&LI++1$d#3*XJh~*@am00y3%C80KsoKzuUmw z%X2z#TLho7Wk!^^LShy+>g-MC5!n4zSmwC-L~m?O%DDM8*yQLmE;hy@|Kuu}C5z>~ zyM-<*EoHTX9WCwc=@nHNx7ctXGl)d26XDSh!T5uUDw*)Fq;QCsq~B1R=`eeU7%xV7 zU}5OakQUKpm3*6JiNF2u=B zexXwWeL!(DN!eN&OvWwz8>@605f8=eLmES)Q$BB_T*t=70MN|a)7vzdoQ?Y#amQwg zUP!-<2#(nM_i!NlkQCw3()#(crycS8z+rOS;bHb+;k&a^JF`tK=>08k`?;2PPQB@D zn@V!3`AT?WKyDXz?P|!E7XgVcu2C+LJ;hZ%SK{>qp9ml}*4J;Dw{1J^mQ;Zb=eZ0z zoi63SY_W^x=S1*E_;Vv87oz`&RE_CRf=grUy)h0k9a5LjwHd)J_>c=bdOEK9N5 zb-3jE+YcjfrbYV4d@IqhwDdr9gpUy_G@(Rcqqp$(NR;>C36YRa`?aeP>+nH6?r2Mp z*P2uywIBs{#4RJ#d9J)^@UmHp(FwB4-Jm%W7NL{BM|4#*_T}(RYZ2Q84T;+n%Eq>d z(6`vpFQu2CnQ$eglZ*=l@V7{I;!Lqs^5S{yk;M7zz;SflAYol-;lxF3(0VWP^aY|a zkCnp@Hup+YvrgkPi%BUxS_$zOzP_H;-35|hRl_e{IjI8|tPboBn3+AT zzOM`$StaB;`VC*JPQuRF@Wn-#crX6434a@*NYyxYB0C`6Gdd*LVGk-kt;Agb? z6e-foP0nZx*?jomlG5Y50L%G0@uq@9OJK)-Kk-Le-63WDnUbt5pp#-5Fvt^AQ}$L4 z5}nKNP{qg9pPIDVe?@rt08)LoJw>Y`GZC_o9ImX@%dZ~mK)KPDS`aH1pP`(ao4X)_ zoV)Csrs6fu+`9nePkns1Z_&4H>*k0`TwVAh)KuGPwJcn7^1y5=t9D?AR3OpXsWvKD zX4)9&79P+&*-~3H5$Nr`J@03GYT(@2h{Q;3m*dK>2CZT4`&+NF;vMbD=cy#Ds`nrTt<3TtU?M0QaZ4(5+D zbp6OUhgmfBT1@MUJBe;=NK5)@yBPX}`xjj?>7L+i}M~?LQm+%L0O{One9C z1jlyiY*8no7XD&d>Ak(t)7vYeyG1=OZ5X9Q%-Mz4ZQBxPQ9!&LQX?H@Gh*Nd;!;9z7T#x#-n%mac$N=cHQ!f7jq4?>!&?_ge z!L|KD>Z$j-azciE8fWY2)t;bJa=DeII@Z4}KH&F!7h^F*i_fG-IDKqG!fL-5k945VK9W@NcgY5)!COF&z@<&Fa_Tu|*o(^19a6)<%udTbC!<(r>3q z%e>nE&|*z45uY9`E*vY$^0x`SfB(A?d_jjjYGqY4sKMZzC0d?Aq*_5C2MhwY-z8oz zf_c8`tVADyM#z!V8|(#F=z3Jt5M$Ax;n4f(GLQB>Cb1RV-4erHJ^1hMafApa%^mD9dHp& z2R()Km>n>hMxBoG%DHV6)wp}sc|jO0joz!vk#|()GZW28ETbZ`j2iM5m-tPkE34YO zd-0qG*6xmb4)W`42XDn>xJeVNSOvAza++3Fo| zItKP;xb{7)hp44{$h=@+z~PONFSG6$f&v^7h(R^-QE)0A;M<8;j<$;Q&KE*8xo@dk z*Zf}A`}6K2P7zMQ?Oz6k^&|1331-alvr5vmIM?5YUV^Ua>{=8QWMxS%h#jOAWy@y| ztiF!+QhBd$lgpC&jq&eGV_8jdmZ?(t*42UnGU`7t&!lL#$6 zfD zM(8gp1}6DRD}`r16G0;#zkC4$kQa3=glTh282*X9K;Tdu6&kfY>-Hgh1w0xZ4 zN!c=TAEso?B$$>q>%*Q3>JcE$BjMB(QP4$%?Ys1nF|J z>VV$blza9Wx%HpL+RS_6xxmhHdpHk}FS@Tw?|u#2*xcj{^U><%E6TqM21&QAR+9iF z^e=0#K1egT4>pu9^{=tQ!orE_hoqgj)kk<7F0TydW7TpqOIFsEoe%*taWFJ6v$usp znT_-pCui92P6fGyL$9>_>E%EP6?1}y~pkYYON3+g0wbY4sz4kz)y)(z+}SQIpN z1j*LBKsqgx<2SS+Gx*GRNB(AS^~Khf)>!xMipol79JNTm6fqMr|IIMlE)50$?Tm$V zgklzI=jeEfka136j|Xzje?}Kc7R8SZ&+2H*bfh?6x=p;ad$5hdzPxVhV9S_8V5HkH z58|$h>Ou7**<~C8A*4aI!pD?p| z@tBzN^Np4UuVCf7x3}Z3I4y{9sR9f@-0+xb#VZdOltatCbs?taHhWl&6q>1FaUy~ZC??C^yL+cXT>E~oM3Yg9aPFoBSNh2`?yU}H}_dRwI3Ri#@y0A1SP4hxJhr8 z8iGfyZM{R$FPaZlqizh1k55lFsmQ3lgZyIBSMmrM4qR&6F``$RqWG`-m`MhY?f81YI_iG9^>R&nylX||s04Ijn*++`B zx>Hsb1VHN&8t?9(IsvdjjiD{8l%gQCeO9In|Bb}h+scjs(g-}GkB z};277vX%XuV44e$#Og2Ja?|_zgZ88e>#?LEca>%{4wI%kaMVE}YzqEGb z=nDCV#yT=-iO^`V%g)rqc*rFwnOS5k=s@(q-`xC6)?D-T#Ga?H>i(nRym%pu7TJHM0)QmUP9Bzut9vt> z_Aa~$_Z8#Pra{8yV(}JA!s6|n{OS|??vSG&0cXW-9+%;S1418Kej-$I7Wuok#x}?U z(z5;G@R9?8eWCxHwYBxaAPYM_{?(yf{CA!E)Mkl{Iz(wFBRir2cft1r96X~U_dH{fkFya@W1_bpzhE&fp^e3rSlKo)+W z9Vtftn?K6wD-oIzm!@+S8J~-Yx zuLH;ZY{D}`_C30!FztIX*%wbDCL$O!a{v`{2YEl2|0#9o*g=_ky}a|7ovn3^_jaKh zMcqt7yY2n5;Y12l+&TI7b-J}%;sYBc`zYSo=8=K}Q;{;AuY0bosDk1Z@L+Ipweyo`-a@W6n~ zC4#9?W`OIuOSBSkUWaqmp-qXf*bm>L)e4?&kG*1Jl_b(2oAlAm5*Wl=4!BXG?{U-W zmXdt)o@}vCIy0dKBU|})dVNbGZik+qFl5sWK)&l zPxgw0WAA$3R?0s6Pq8&H9({a*ek)t0hB;p*V1D4wM=`b_YhQk1JmUYOw?A~;V3+=6 zy!T+Rbsq?dqBn&z!-q82U?I%&h9scQ>c%HQ_tjQzoK2~;ulM$`lZGle)3*}Fnj>*7Xa&yo7-uu1J`#jHGRw90Sv%2CU7u#-Q`Kw%l zmM&f##UElTwX0VQy!$?CSTLv7{CyIPIm9j&z2?S4mhUUnpsH{d+-ZPjh^l%u1F2_g zHAA=NxS{ZR8LDblpbFdxnDV%590qoS4zb^*Qb6YeSPQ zKaAjZYAOut@~1WVQhlC@uw!?&lmg4AnZM*!onk!6BaJ)uu3-DCDRAd4DACT6tn;l_5MIkY%l*^ zr$|9VW*2+gVN7d!cZMny7u48)i*3Q@x#ncv=xNTuX~bHlnFjprk%MRnN=IPqw~ZyzAZ>ih?HG+%@l6@?$cG z<&4&TJ{(tkS*7s`@@P!s1?Oj$G0)X^H1m~6WA?og)0(aT{fhlHpTwu#{ggJ+-@kpK z!f!sks4%FwkFlC8_gJ!uMTb&F+!1zN1Xq!E_+W)>Gx`3a1;i@IV81+E%RLMQYprE{ z^9@Xl?K1=*9q>E>YK9(zHU;dWP?kv}_4VZ1jfHuC*Chi=ar zt|h5>f~u3P4O{cMF}`F)GhsfqqnQ}$J^@>)x{()VX>;LBD>~w>e66S(B&W3eXb^Rk z6QfgH3?z?oXy%JI+f^*lXJk0lTQ!rQ!>nop5`d>!VVJRfX(EAuSyf%#W+r(88gJu{ zrxGaIzC34nn%Y!BzUHMWkboq&x5&$8)r4(jWPP(9pd< zuEisuwwn70TbQD6idpFz+b7DfhKTi7h1rz!Qn7Dd|FZ%uL;oM`CIzXDS;6`~BXPnc z#1upheeEG^aTMUZ(Q5T{5g_?LzQL&+L+}pDS-u4DVNleLGMFxouG;h{)0ECm+X2>IVAo0C!{Wx`f&V2u@ z;Ec>UO$S@s)B}1{?HGfvox8ilqjuJ@-51M(S9>m|8kLvP9-DZy&1}sTcQ$eq)MfRD z2tsKc;6sq|%^@0j1kBPnLf-kHu?Z-m>g&)6t;0?Z6;~paq1$#>+?V$bVkRIs77ax5 z_C}X9c%=J7KeYGSLbtuV-%upTLn$H-A-*aMN$RyG08()dF2wW>mr^Uffs61+Fg)Xj z)Gy3DwkLz?5$uk>J#LB=XF(T6q2qk*z7e)!^1Se?L&ko>$v_Jj`!Rb&kSNyaip~oh zwuwrf8?aUP0%Rme^tWHMI3XlarF@@j$NR+rSVap2#s5?ExSNxOOUOBzCJbn;A)Vde+T8LW&Le}+cqE{9WydrZok#=Wrm!v zC$5K;O(X-&lB0#VLm_!ye=Mt5Wj3OcuC`nN;|22P%~vs!#=W7av` z4ZoDUpjYVB4=201($c4JuOn)MfBSj!G!Arp@RF4%1)S`!1vTMaFD@X}#63!j1$lX& zZC)L3R)zZmgbV2Jt*qWE661UsxtJ=HlXZ4c5vW~#!(flo>@P-l%$}nyo(J2p0h>~a$z8?21-XwH#0K4;~e3W0qfL&5L6gAa_ zN-g}%GQ%@PTg(wK@gc%oJ&nCS3q^(b-UDR^`qiUE9Z@x7#M*5cWN&XK=s5`V<+i_n z@8_p7!dNZ6ffEP_S&=yr(m1Jn``CECPRlkIlSN2*x+|Wod*ip)+LE6-d+5Vzwc9~U z1u3*xIi7K%y7XcSi67bKc81QMVwkUxE3s-$1F+DS5tZ#G#>Sv(EmxA+cCL9k;88f; zMFBH};VzTOBfjQMP4}Mq^M@A5_~3%7Y>$}}nQey`Z>qfHtWi*s9LaM2=INDwfAPHT zn9>2^$!K~{TcLTFW%glySEuElbTr=YD5X9uagAFTFS!NT9e%Hx=b1V;yWXlMMN~no z>`_*ZHyVnzBw{lzM+Wr@C)K`&W8Uzpo#Z1HUys-zCPDMa5J>;WP=f5pbadEQJcZo%d97kg$ffkAt#J=1#Z%{66A z-6!4mRsL|qF8K@pN*re|OqbsK;G3?gFj)N*i%|KhtV}oajQhb^N|13x?Zb}IqB|t6 z&okm%raRkIuNvdl0>2E?Q8fL|XQE>4>09|b?}(D`MW~4;Xdz|G^Wi8g#GzE4^<&GaB1Xio&Ub1A_y3dHMCl!b~DPeePyfQh5U2qZvp3&>VY8 zB1KRLV=(2_=dp7fjg!h;(-JqZAR8i}ff1qOBAV|&d)bj{5*sqtzU*#oU*M2h6Z_+( zOMa^Uda*8LovRO4k|L|WPAn+ap+r= zP2b^eg9lnvxJ>qZ~w>T|HSC+bZ1P4C;T80avHEYS{oBN)j4aB0)u zE`7AzlJ2J2!bK#03G)fD1cnIbY{*Rs_nX}SA;rgQsP9!y&IJOQqv4GTcD!1AU0=Mu zIt>Enj^;YYrT4Qdgt2e|`VhmE>^1}CYu>(|o_^=xTbSNxLQqiJK8`BacD=>9 zD-2Mu+M@)`$BQB_RhDF#oov$IR{ zstO8pg(oryITV(zPvXLFbJx&ldYINHOmCV4teUzb3_A#2XUv?1i`VH>tgq&o)HhP9 zRL#W4@MrsH!qR_i>$Y**$1=$KSTosD&7A#!1%T-I{T%&1)_^w#xiORfYe}>blN&L) z5tF~08SF-IZ3NdwaQzo9${UHZkvJQPvynI(760#IQvXL=UX|ay4Frb@prq#rd;|~w zKic$cOv%4F>47EbFWI~U6vCU1N)SY?c9lQPZIw>g6`OUAtAI2`YUrVN=^#~lM-WiDKL8U&;_AwY!C zTj(v)+ugtOemUd4=Z^6oBojTKVqsCRnHXvV z&5`5)PaB`R*tzbK0RxgPW%?-pA6^bN6*J~fNevBZcVRkU};gB9-A)1?gzfZ~VFOcQH z!8Wk_Ub7BLnE7a?3V6ngo$hxT@1)s8eRrVgvZ6MX85MpKl#DK{k*Ml)?`(DWclFA&(kuFj3>-wuWjIh`%pF>X{KT>|~Erl-644Q6sPR3ph|*w&swU zgZE|BdvYOSc`Om;Gu5^3BQ(_af-SNkal~bEU%%&>ikyxs`PHORLpt+@48Z*ZO^({m zDk5|4LZ6`TtFYIFMc<9R9e0aW?DGrvOor!9&%vFC!U|L}=&UNH$()M#jGmEB9wtFNuN{%&fNygR2M zG8+iEfGdH5az|mvX;cJ5JMuVSJHU%`q(1g-7mLD|OhL-7dChqUxO@5V=qOrFkXnO< zG|E~sF_t|k(SG7I2&6ue*G8wjZ8BQwU?lv^!}L^g0ZLi`t#;lXr-+-*PV2v4tUgKo z`B{`)baxo^$V3CyzOM53`rY&vMf+h?LW-V_661->#?8z;;^O4UH4q-uRAn6v#bLyUj_z8klvY8TcVhOpSMw?7aCVE=Hges^*u2! zpzde~zNvBUuww3yM>d7z^H;icT0*K#>z$|iIo!z6BcgAL#|R?GNOsW8kN@| zSds7Ar65oCgti$!u15x*EwSD0`YIz=D$;@sY<5lxgRW+}47IJqAfm^Y4K#uezF5r2 z<@jgioYX?WjKV^KxA681@BTT3zpL1`ijc)#^R2)w^HLYm0uCGNYI}8M7yPG zX&@tqs;6tSC`ZvFo26OwY7ApE{$AS42EOy*_6^J-WHj}5a_9@w(PS?N`@$4^LLxEK zTK}h%u*KrLcz9yBVU+n$QuFV{U|ey_URh(}9L>&|Dk;orzC=#P-d>C%vtG=r7lx(Okfwz;KDKW z!=%W>)sp%3G>l7M@zEi^VXI)WT0&Y8#4C|B=IcLk*I?KoGLRfPpt*2K(lnX@#Wd4!1JuRN%&Ev>~7- zpTGT8S%-p4>rMONRpHdB)I44>g^#%E7vUP%hqG|(jf9(^UIFBu`r@-ryk6nCFHmCH zl-1dkaUTt&FTEQdm(6Q@i_B7Eflex|aAL0`DvGg+@1!Q~qhFw!bFX{;!TacOco#%r z0c{p{-EE#~o-Y4l7(AFFBb@{ESyDQa`@B;X85*$Ziz!C6pRuEml89j4Zx};&-C6?O6_m;>wVMXMz4p2z&R&vZt8olI? zQCus*Hd`6Fi12Y=B0O?CD?>i*=r_IK;PK9{mc4(R}g%`BfBBrdLeKFdCx zFVN>|BX!oO=l;zf%~iSw>ER6xvNn(Bo^2CQjPLEVS(2 zbp95Y`=LQmCe(jz-4&KPC~a9nW3w2KY(+StZZ(Uw2RLAoZeZh7;mY&Ziw9O3K{YG$ zqXJzzkfD$TeafJX(-QOJ(!Jr^uDQUn=ek4{jh*owjtCHDyT)6g2d;~ zWf$r%?}}<<^rXMMF4KU`=#jvhb8lUgQOifvTu*OSRhmN6j?onl?&5tkM>TW3sw^(z zhGWz9(VO6QL_5X4(FQD6g)yVG*T=BK@u8yzy) z2`a?R8ykaqhNg+6-DHTulBxwe{CKk_2|p2ciYN3~CMvN=Nssuw!z7MZS4Vli%04Ws zs2H#n%Jx7B^9*{M_TpNPqN4+OvFGz)@-d$yA5>4rYJ2il(A`CGupTYht%=IGhOG9S z$pkJ4)QY~0%g?hTOyYvT5ngO65_@;=_x?C`wFyh0im;(btX%017{7^aycrfaRbMtM z9V3ZbOBu)43fJrjTb7$_^|C=>tpg+yWN?+45ovlONJ%rHuHmP7RUGCdGNSaAV3CTz zg&T%jaiNMM@iZ3NtU1p&4VK}Mj<%$p`-9Sl3lueWpQ|88Eu6MhbGSAQ4^h$mHuKz| z!4i>A{Bgn+jgx$<@~=5iP<-3@14`V&X>YT9xQ6id`o&s9+d*AR+blmTRFf%4RByU^ zohXE~s`QZHS8UQ$xeJ$r(d^;%s9+Zp6ueMs=-EcXJiT}Gi72>V6)H4L92x+`9HFk% z&@SOZ{~KM(pp#&(H?SAb#MHrk;R2YMcjIA!V_Wt()8XbeT#aEt#NQsr8gk+5;L-L3 zfAY(6L+WU&^|s8c(wE1S0<74vflP{a57)`7ZYZgIpQo4NH zG5H9kJXdWTv7h>~M_uxA&Pn4T!jdEYA>K!!b+ENf)b*l1dI)LH!~S8eUP+mj@eX6K z$EC>WwOhUCbA{HOj$YUE*+}e3v5F;lfU~Dsb(>`-(O@|!oPXg`tpI|=DeC>zkat7u z2SJGosI!>HWR9~n4&Ko*4pg_pvqm{6q$+h(z|>K4TU*;iJKJK@Gan)HPz$O`A=?r2 z#*J>udL2DIK2rn)RCkzBVlMq7Pf*WL*K2JOZB#rt=e69ui*Z5jKJDVy@T8w>T-wC! zHF9NegI;yB(+B)r$IK@yv4mXw3{pwub+xy5H^qiBTz_?#ArzgXiR)&#oPPV+T-x4n zDx!jdb4?7%w1$RrUtzoIx1ZID3PyF0ko^dR%{)07E4~ z>O5N;;ClJzr~dYU`{jxAOvnZ^iU#v!olBmb9nDdm9;94r!Z3e3bZ}$W6Ti#jSgE0I z%-s}%RKEv}rJGx~!PTj_Ru8_9A|Fn>*858Iz+1XvGaND?RkAxH^F320z-=_1UAn;^ zwGjASOi&B6m-{)wX9=;^w>9OXSU{ z>-pwqNAuu=b}sBtt`~~ILsXH98LC9BGu%>8+f?j&8JQPUeKg(P{NeBKv4N&E+A&g) z7)({!g0)=u)3aWGC_cUwI-FboEEzluBZ|`hy_#e#!IX% z_qwVZYd7ttFi<$HqKnVT1Uz5n9=D&heG2Mg_IKt2WtC8|TE~9<01KpWwpeqXc3Mcg zwKG?=V|y>axKiv@MT|Nk%gVV5QPLf*RqwhYu`$(K9gK}5El^y}i|vv5)n*)?UZ|7w z-Jaz}A`|fzltt`kYd2!8tc)h*di-#d10ae?Y*fBo8|^nQ|D-K%o6;{vCg6*y ziFY7bVQ_C#W4n*$6GFD4>#>k_bE+LkNP*>Gjk6;!#=b~;{KF6Kf$@n;t%XG~%9&DQ zsgza!+S=f?%h^XM;)f7Mp;y7yD#xAjapik7GuInUkfJ0=L%$rFcS%AB9-vVjoJ+}2 zmR)E4_h2OP)946t&1Feqthza=_wN}>5L<l$9MG9?t&i8+>>|8@c2+a&Vx@P48;&d@t+qfM z)p8qnssNsExmK*yZ!7t5{4l!EbPT-vtKl|T!mx~_Kd7blmC^h!x}LUH3XQ0rIH9-E zCqYl_x;IUKN031Utn^I_t}p)D4$4c22H72J52D*vn;13x+Y!Qo#N~ z3qXai?Lu&F2dBBT%j7C2-|D8pY^2UO1PmoNjPtogHN2t^glig=<;^@yczENSS-()d zpZ_;E>>?L`YV+cytZENik(1kzc@5KT1vMrj-KVV2l$q0i3Te3Hd`7WG!LII^J9S}# zs{MQRXT}Syqo-Tk+v^%Mn$vIHzxcWx5{-M;Q!zcvQ+f9~=r~Tz_@$I+TU`o7PcLpA)ECzF}JB`n{=&mNLP|2hb@8^G{FqWR}M$FC!nRuEg~3Kc`c zvBa`F)ep-0uNDXYML?2#cb~`_TKtu(DC>?*1HIv}BAaUwNsfQ+>u@Q$P-QGooP`*$ zm*%XYt!e5%Z;fPq|^ ztP1XkyjoEVz9Aa=0oxyHyz$_Y!JTeeeagByX>Rff?pObHaIt!aBJ?Z%Po#H?c#KCSTE`8zI;VwSyAH#5q+#ani$q;pbAP3HKuw+YW$lXRPH0h z3{6z>`()eablZXG%Hp{8EO3d>sTd*jndCO0t#Tl{!?9ONAsyv zi2(R}lC0pfq3?yw{m{pSUlbhF54v3rHFh=@xH5pnll#YY-n(tvs-f7jL^xnix;FN<{4b(~e9-$s8D1HVp9uvkre7dv!>cbV_KDpo&nWM$KK|yWXdH$04 z|MY$c{@WE`H=NsY-eech(vr8(80s;PM&%4!KS4<@M81r*#&smgAD)^nT;(uO_&$k zyv1DEhZ%f`z%=2F^J z+9S?_is=A~uimnzrVR3X4Z;y*v~i{gI|lxDHGWQoT9nS z#IPrP#PXIY!$1zF6Bfr1Ab@FD;QQk1)fLuqTOZQf#iJiAnN;>P`=g-q)ds;VYky>h zNDQ1o;wn(>k`T9dHLKZsLWw*vD=328@gWeZy6uW!I5WWtCG4E%El`>vg3o#Mn7pd% zl|jR?@$@sotsz!jWrIMd?N%z2MzyhO{I zE+O#Rt7bP-Bj)4U#Zz_9UH5Ubwblp%nS+F`hf@uv*VN23;?uhH05BRzfk_hZp5`79 zB-`^n$>`!H+t9BoC+#wplA@0z2C>5AK&3hp(N}~vU9;`+#Fh%k;w+g$xtYZFM`02i zSwF09(=L4kD!Zbus9|KyhzMfkKCY}{4tXjo_%>9Z?RhUk7597aAzkm7uo{>;p%(^O ziu?#2Pcxf!*=R1_rWLKQgDMn^{qVB;c}`&bQn}O#Q_p?57PC(s3t*{Ra2Bw;i_@S%Nf2P17;| zvahHs)Y>Fs(BIow2wa#sSazEzJ|cYCIX)@_`wj)LIh$&$Q2w$ldL=AXxsFwIH8?{r z*volr#q1>|#kh7frz~~yYr0e&fgUa*ifK$HpP4H!(Re9p(tu*9Z_nVM8U2vDGtqk zJEgABmwfLHyw0}_Pc;>+U%$%pZ*=X|xSH*Y54rtyb$YEkn9?xXdU((jq0Ydmr8>fqzO&z6a@D$_px0r9~eD_#p1=OWVR{Z>otzB#Ut^V0I_m^&>4P z1nnh6wwaVv1lihHdK)DdmR;t@z(7kVI>Z?M%M#&s#2xq%>VD+aYkAnex9H6x%ba+y zq9pvv=GN8}L_P93Lfaey9;{A*G{ebEWo}9HA#x~#u|t^~cB%Hu&_oCO>fU@gYzyPpsp zWLI(}D}ELhnWS&dcT@)KxQ9->3HSVB1>=YdiuiZrcE>Rl)&)M8PluV*rS(m3=`d{Z z``-x?zYe~Mr$ghFacqVE2sX8px}?~L;Leu1Ft#F-kOZPrv=uNSLG_1 zl=9%M?5O#CDGeZq2wK5_#e1CjuwyR4F5p#C96=eS8EvIK=^Ft|O?yc1CkuQTYIBa{ z*~uY;>my^bx|(BqhYzd1!5bStRT4DSn8BNR6&7nbh%#HgH2NLiR`(Gb@Ih|nEv?wJ z_4O=JXZDu8zpR&e`2&|9sq2$g$6W+HY5!}`YOBDaJKb_^0XFO^qs4u?zQ!xeXLZaH-^y|Q`+KeRPHQ8pbx+99>%*GxBAu~B*^8<7S4jv*P2 zB>{(Zh9}L&!Z=WJi}$))ZKs?$nAthN*%`$c3QP zs-lhh-&BPPSlQc~lHPm>6Ep?GGf0*w9Q z?d{SW>iom63$*}J#AlbtBKE;mioY`$pEBBE$%W&uX*<{-{?W^wixGVFx>;H)+Fo^< zZBoR0Dx#x-?VZ{lr<~HCs@FNqY9C3#`1xD0CB;y!o{91X+`sfeVYm{g^JHVo`p((v z`m0#T%X8%cPuVEatJQY3AXzX7)n|?N?KiKZTyl^QZa)gA$%iT_Q!yv!P5gBWXm@aOL$M;r$WwLVU#hbBVPan_OOT~;msZ8oO%>a9V=JhJ6nt${4rWu-% zE%Qz-{z2j2)!L|6*sWz_>L2QA0=XHHTGzN3xd`)#Sf$#6G0sY!Cv-T$w~$A3KPU$N zC7efSe^6QJPEMW&zQGr;S=%L)PSD9L?p{GGCH9+SVB2`KUW4q_P<xp`1)df@0Y3bHa`F4*7ZJi+ zO#AVb-H-fmq9j+*2|dQ{wB3A;Puw0XoQsYDf`WI6XjFQPK#y=-Ym8A`#6vF&^GjYT zLKu`d4B~$WJ>zzlt=Qp--GTm@;%lW+UMIXec#FdixM3_-M#>JQHq)8#c8Qxk-Dvy~ zsnKDqJLK_2v+;U+fWli6#_;{i^FZE%Yf-lpT`&o$1`AW-clb7Ct2x`qBHy**N8C@3 z>_F)%XlM8{yr0G!bqxcFkeNiKvevMNp;*6c9O{`a2WxwDIX4tyisl&{C$tK5n()d- zC-a+m2nW|GyhaIk2|Hs(Yy_vB){kR9LcplnqXABE)OKUG2WYKnDrv5bw2i@k@`5JBbXy`22(3}C(Obox*P?BRP1%jnva_O|CM4LI%=x25_xO z;}Dqxz`KGvstc;YN``${5jB8llZ7p+idw-yTr@(=B{wlWgc3bL8jsaM0ql7l2YyA+ z(ET+1NtQv*Y}t6+Yt;NTt#ZU&s(W`>g%eeu9f^klT9DWC>onXT8wbM-hI>ns_D49g zpPdL`x3#6 z_OEITL8CY|p+r^nA`~{WQygv256;S>)vq#P3onXw#vsZ|8b*H;!gR#q$nL!nqZIVB zU;3aHlw>}Xn2Y5sFB-xUL|I2|i(wxdKn9dp20VGvM(i`61?LmDRy5zl6Wxw1cjF?y zYM1=rfEhP91sq&1e5iIPcCgFcb|Yq!kv<2{`*H>)+kGYGgiQ#0U+*$FIoU6U?z>&A zzC9&6KBX!?ACPN};kP6ZrrBi3vpdks%YmUl& z@aCs@1GnSRz?PUij9g1j^hp*4FEXP|kV^O|U}5(RD;5u^YT(;T1S{D>JpeJ9tt7EcEi<^?{!K{CP;SF-{gP2XlP%DSo~j2qi3HdMd>?8#`C+JaWEz zky-2zY0-KiMa`RubXOSl7qg?ES*4cT{ZyJ@9h;-~z~@4vw=|S+fzyuAcgnd^>b-EqLvltaFUEUXU?qo4TjG1+Ul0B%hW?c@7mZC$(}lS= z4TVMQ4x>tBuFpojx(4Rd0Q7c3>1m1!a-r#QKskjy%zTCt-mP%7wqqa! zlWR4*=+qB0@VS4|D%G*E@rZ+y$%Yd*J7g*n;U98XK|RKhp3}2_xSHvq3-&xEmSjqE zcS9=q=!&Z^W`d{1bj6#LX_M=20zr2Ucsq&1;OPt}XIhso5s% z8dwt~c*dLSf7VFk+@S>twWr$Q+Qa2-ItuP*y1eLXX>N;BLTu118#M?S3UMD zX3hJ5-XS$7sC2{m!7vC^lTn_D9*|0WVS;LooL@!6tn-gFqk z+;^lw2}?;<%Ho@%yI9C`tKD2w@kxRqU}}1+vsJ>lH~_~!o9-ft)Sy6eQCXo)LRO}R zM~R3C6_~U}{N&|r@(tXg<<(G+u$GyBMka(M% zs4z-Ud8uRWZo=tKz{Sm+vFP*B&49~l=EAP5jD}Ym%_a1wAdL+VfQa*H_VP8@& z6`Ya++oH`fL049k!7IxN0?k1b_kx4ozV(xj*uN~%B{2O-ojqh0y9u~7y~J4gHXDpP z*{eB$uI|a(>`dXu|Fpy5$f6=^*WNwm&qOfP3d$J6Yu>U4qvJD5^*X26?om*T2hQYv zrm1G?>}O*mMnJW`vYxsbCD(~F8LVa+Gw@PV^HPh$`Ynwrp=#HY#sx<`c9mG@6M)f* zz*#JH`ZHzJa>o!SDY1cI#SW~=tnh6;9lfbZGy$Vv#?aE+&-hx90f*sXn2pV$T$q5K zp`ospmVmQS40K#@*w`gxbU?QpDDYi|vPC3lVsfb8Y8Zb5XmfKDb{}vQarGjBT2#eK zoylp-d$bkp-gRUSO&l!#A|n_Tf6oL0=71rwrXeVvxjAS=2JQncW?LRe+PM*Rvah#y4roBjeZr+rl@H3>j zQL6t%fA8l94_(?<+K~HMsmFhqi4H$5c=rYnLzGOg9+}+oGX)#8|4+U9HPS zOG9Js{~^FnzWeGv0R`h6ySmftkDD^bs&Y8oB6fH8PmLW~eXx4ox=vfHMjm=PF&Y9P zMHlGZ2gFv1UmTn1(@WMfpiHW$=x$40wc*zx=4@F4PgdAvkgm7$@xI*BZI#aNzO`3X ziKc)lV!CrA%!Hr31v_9);u>a)xBXd2O>ZYwpucHdnw+hDIWoAXYaFrlm09qXrS+sF zhy>~MkDLCZqk$N{mOx*(7luZ_$j}3<^wY3+(?==D{kF5t`h<{W?c0;%5VLBh`&Lf& zZtu>^Vv2}?skQm)gXK+F#AhP9*7GS0DJv_KxQl=9Ji#@qxas|Icd^5hQ8!uC<<043i_f0DgM(qMuQG4i z^^_=b3XN)O4B7--fQ@Z@;k?21dGNUnY$B{X#>fa$?~3W9Y5_v19Eefwt(BF7@@70` z{qA&Kax3fUsTLEg089c(A%!ThHYS;X^PwzqfDV(X*#(e1Dp(;Bg0IN*O@fABn=@EJ z9V&u0tF4tcZ{nX~;*|jYFVWs}dtL1`5WB#|$z^P8C&zTMj?KgTovj&6tqNSmCIs)d zBT=*COyN4ws##N#0VAz@aNdD*7b*fwTK4gQS6pS&s^DPts|5O{(|`pYzA6qL9w71Y z=*=~C)gtoVy38eS-MD{VU`$#=CwlOYmZv1IPQB(=Sz=jXwZXOOW{beZw+wAtGuybP zKsKQde>W_;gxR%6UWeQ!c71;3y+7OF<74PB9eJ$bI@h|Hq{1fytkr)KyBw<;-S@7} z2h*x%+kN(b{`|S7UlAk6WZ`w5b#dpLF#g2Dz2*&br2kyn?bTM*Rd?MjwjnKOxiDs6 zD2Ko}hN2Wgj|}d?XY7|7_i5X9X)J9E+T7Uai(2WWb9t*_5cyZs2A?3|}2H3FudlOyg-Od$4rKSOK+=?Qph~9d?8+ zLiT26+cjFDaC+!hNHEs8>+L5%`r|$#qPD%e3*u5yb99MfCk0m9wA2(>Z!d52SuqvMClV_j&wQVOf$rEu)aj{1=KA8#j|W%z9Ua!yK;o7k z8obx?_U$JyP_J>`bA6xrQcyB*wa;Q8-DWxqAk2U)nS!#?tUmC5fEVGw{m~^&aCp+2 zpLcox0o`T%XCx@E1&(ceN+brG@R)-Q>OUcT5`0Jy61?woiyN=g|KqKL-xSzgmLpLg zaMNGy*i&mFou-~(S2Jz$(imct2{>7XDp0oWU%vfxx5>%Y$!Vy%VkM{dSsfq3`4aG? zW;nr!*INdf=VG{n&!0ah&`ZgJsHt`|0P%jH1tx~i$=GYGJLVampNcIXIipUL3FE?t z+IBnpE7)79xle_6nO7oY^X=SX`#*Z*LAILrqID_NicsbkgY9PxRoAG36V^H}FX0Ny zEJz~VIAM(@M{GhIup(?W)5mltMXeTN^H3RZj?(WH*3iA_w&cjkYuA%wE|}6UTU`on z>g2=k{DdymukZWCot3=rhq=%Iek60%aVvRVlP7&Sgc7nFUAj{RFKx1%n@o^ALjkvj z&j`4U8Y;8FdD_~T1O|n2(nQ$`Qa2bpXP})FRzn{j9wJn2nWUu6`ql=xxM1WWB6Z!C zOE5ajqkMouD(E}-{nZ&0&S8oM7|jBN!R5HksO0&tf|C<_7C}M50s&i$x3Y3SKv8kQ z{#j_?&hl6yk5NZ#T$)kY(e)mh_%P+v^TcSCZC@a)6hn@FUONlfnFUVI_)3yI^Ib}N zf%lV71`zx_QIJr4%pCE;Rtug^o~_A5{y}QNI-6f=F^sgp@dB9f&X`t*Kqsif%toJ* zv*}#LbPZY5ugZo6#7;qsKHN-7nExw1gQ!|fTafZ)5^PM#)7Q;yeMcO~r|Mm%oiM4A zhQsm&h;$38ScEUr8<5BoA=0?x<6~fcS6@^Q>YLWC|IYvP>F(W_G+y(d%{52fripk3 z4Nvc9M-ic+cSb#=&?2hI#C6|;lnpuSJbF*2>N8G^1#U^a5tPX7miwqIDyiCl4b9%) z`SD}Sbn5=U_?|UlK`<(-CMS43GPt8t-eaA&4~y%zwAehNNVW0VnRdn`_D8Q<3j!-t zW;wj=rLGwe@)E1Zq0Wj5k&$6P+fO2Cs~Wd5FD@bhE!L;YQy$|cLHw(mP^7)^dqpMy z?~h62mclO4+6vPU1cm^osHdB*39z%W(iWx(z%hS=3W^8{OL`YMezS91ZAvqnuW#_^ zQu%!9^HdVyWzL(&4(0?RFD)HiDI5+YMu*EMK!MvB{}#XmxVTTH&vlAz(u<)?G;s<{ zh_S@P{QNf~e&*&?E>zdoL0zBT0(csOb4N z4U$_3NWWx6sTF|GhQBj9cQI9&BEH=jM zRY4AlOpz8GrNU}lt2iW9MEIri><=Uy8A?2k@&bh69wV3j0U7X1WEv-*jOV-d?N#1>+yqTgQ9WllY&=StFvNAD6V*qqNQn42v zpo-~PM(%4|7r_x7h^x?0OGCj|m6ArDbD7s;;i4Y-o2oJIJ4zX-3Y^ z=kD>EimJ~~tAAm7|L)s@m(<9oj*dG(2_<%UvX=<1Q+1W{-swZeOLcX<4gY=-*)tEg z2%JpuDZ`P8&q=l)WFXC4`lsL5jC86U2K)CeaCqrl^=1>$bE&o%6f#&J6LHlf*qc#E zq9>4+l2Tt=Z|i;@4wFIn*`5A2u-pH-5Z2NQvQ$pLbxcCuTT~`@LaI)GG}Aa)8iA>$Si&- z^TT>tMY9*1#rxbxXSxX_R=+6NOg!D2_=_h>$ znv0jf;wq>X;L;_uvSJkp;32?-!Zl?VZvuN?91Kce4i^W%H37}z#=dLjl!>ls(6X-` zI%>trirHDnmNluL`SVj4%E5U+Z8L@@r{gly?i<$6fd8FfQ z_=?|Bu58z*Pl@){>pRoz@jfn?W>|F@fKC5BGmw{uLEZx>cKvP}U4ZY;F3ff3d08MY z`ju5rUteDj;poKFq^>S{RghW@l^e9a>FLj%L!T!S0C&bbB--P97BlVSZEbCUJ|6l7 zyyScvzG-vE$3HneIdP+Z4ixe^^bAxRj`mJvrKNx~Z2al_g1rPnSZsAvl&3$Fsg_og z-KjcgZSm)8;xpMV9UUEsiHT)fnH{0QrYO6S+?o*cUZhk2+k$WMidFi98DTuUujW7X z#Me@Ek=PvXZdkal$)Eb5!Ls_nnF5Iu2Ju(<4D=_8HSw#V!DZEUNAX3L2z#m?wT5{l z5D$S&5Jt`_r{rLmTI_(_$Gvy~wa>JjCRxCBZY`F^$iTNVi{F57l6<65_ujk+|KW`J z@EmmCYN(z$oEYr|)maKu_QkZ5%aV|u1mSaqnxONc?=z+AG(sT*{7+ETJqnTLxAx(p zJq`8sfjd?VM@<#rl&ec31^~v)6>USVA59Lr(=+UBUd;QnGe}Ek*Cc`sQvr-`MJ%&8 zL)*v@0{&MwmLPR>u(9#O8IVYLW*1mwJ5jGS^D++b4;`pH?-h8`C9DQqFV-`A`J8A$ zSZ(C)D2~G>vp7eOQT9o@sHGGo~oOfRAt?DH7zsZSNYU7 zd48-ds!&2yGMj!2?J4#APN_~~E3K`>D6SK(#UAj?mp*?2Sad&lcK!Cs@z@KPUww1` z4e=S0iAC~6e|+hFPDirZkHBNa_J+On^7qXF>?~<_yBvBCh1rD2Izn!X>^!*(w^cy} zz@D-+K{i;dTp<&_2^tm)VihSi{1-AiOU?C{l|rw{hMz*Xjb`U(Ug92}2BWKYeq8FG zt?P#yqyigXAPP(*qd0u6VuzgfDoY+C%!wPNqOq)Ui_P?`awri~(ZJD|9!t?V|E%K8 zM5VvJSVSD{lnUrcm*n)^Kz#&pE;h56;@%y`Xozjz`?i-}WQ;uyXozTqDw%d1vw=97j2h1xwEtGy*i`7T-xlfny=P8D4pYc@e`z7EKMwV`|~i`Tl5sdr%G*PQgg zPL#r8Mb0dH!rE2b&1m2?rT2qJjNq~6nWS`BxTXb}AedSULh_P->+FWh!D-7_wAISW zJrt7bZhGJyQH>U>d1@hx~$%;nFIpD%}#jY~P!E<<(3e;@xy z7~5tEPjB8ij>#{Y@C0_&TKn1`Tw$&OevFIL%XHLQP}__>CG2xsg?i!UE0pKa%Z0KN zw$T@iEcoF%B{N%OVD-8}QO}cS>R8pt^#+IMo%PJxX+h?Af z9K#pQ7EnueFk+K3<9=p-yCZMt*JNXq3q&t`f>yT&Arv9^078Z**+W|T73aU)ZgAoq z$0OHeC1IKQ6o7@j!n;6X$z$SI-X%W=AT-&ZnawxHE*CtKhIVSa-}kU{{50TF0=m#r))TQ$ z?akOc(&#epe>0ES7+~Ym^@S}jd_QT-m+b9g(D8qDzyGsc;s0;fzeDh!|0&6T fsih874sP)Zg-rhZLYVFTyV5AjtINS<--Z1@Vra?h From 5fdcdb88f073ee7f515a53b49575f77e6c34f2e5 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 12:08:05 +0800 Subject: [PATCH 293/311] =?UTF-8?q?Add=20Model=20Selection=20Cheatsheet=20?= =?UTF-8?q?=E2=80=94=20bilingual=20quick-reference=20guide?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/model-selection-cheatsheet.md | 105 ++++++++++++++++++++++++++ docs/zh/model-selection-cheatsheet.md | 105 ++++++++++++++++++++++++++ 2 files changed, 210 insertions(+) create mode 100644 docs/model-selection-cheatsheet.md create mode 100644 docs/zh/model-selection-cheatsheet.md diff --git a/docs/model-selection-cheatsheet.md b/docs/model-selection-cheatsheet.md new file mode 100644 index 00000000..be900691 --- /dev/null +++ b/docs/model-selection-cheatsheet.md @@ -0,0 +1,105 @@ +# AI Model Selection Cheatsheet + +> Quick-reference guide to picking the right AI model for your use case. All data from [AI Models Catalog](https://github.com/i-need-token/ai-models) — 4,587+ models across 95 providers. + +## 🎯 Decision Tree + +``` +What do you need? +├── Cheapest model with tool calling → ling-2.6-flash ($0.01/$0.03/M) +├── Best free reasoning model → DeepSeek R1 (92% MATH-500) +├── Largest context window → Gemini 2.5 Pro (1M tokens) +├── Best coding assistant → Claude Sonnet 4 / GPT-4.1 +├── Open-source with tool calling → Qwen3 4B ($0.03/$0.15/M) +├── Free model with vision → Gemma 4 27B IT +└── Cheapest for production → bdc-coder ($0.01/$0.01/M) +``` + +## 💰 By Budget + +| Budget | Best Pick | Input/Output $/M | Why | +| ------------- | --------------- | ---------------- | -------------------------------- | +| **Free** | DeepSeek R1 | $0/$0 | Best reasoning among free models | +| **Free** | Gemma 4 27B IT | $0/$0 | Free vision + tool calling | +| **< $0.05/M** | ling-2.6-flash | $0.01/$0.03 | Cheapest tool calling | +| **< $0.10/M** | Qwen3 4B | $0.03/$0.15 | Open-source reasoning + TC | +| **< $0.50/M** | GPT-4.1-mini | $0.40/$1.60 | Best value frontier model | +| **< $2/M** | Claude Sonnet 4 | $3/$15 | Top coding + reasoning | +| **< $5/M** | GPT-4.1 | $2/$8 | 1M context + vision | +| **Premium** | o3 | $10/$40 | Best reasoning benchmark scores | + +## 🛠️ By Use Case + +### AI Agents + +Need: tool calling + reasoning + low latency + +- **Best value**: ling-2.6-flash ($0.01/$0.03/M) — cheapest TC model +- **Balanced**: GPT-4.1-mini ($0.40/$1.60/M) — reliable + 1M context +- **Premium**: Claude Sonnet 4 ($3/$15/M) — best agentic performance + +### Code Generation + +Need: tool calling + structured output + large context + +- **Best value**: bdc-coder ($0.01/$0.01/M) — cheapest coding model +- **Balanced**: GPT-4.1-mini ($0.40/$1.60/M) — great code quality +- **Premium**: Claude Sonnet 4 ($3/$15/M) — SOTA on SWE-bench + +### Chat / RAG + +Need: large context + low cost + fast responses + +- **Best value**: Qwen3 4B ($0.03/$0.15/M) — cheap + 262K context +- **Balanced**: GPT-4.1-nano ($0.10/$0.40/M) — fast + cheap +- **Premium**: Gemini 2.5 Pro ($1.25/$10/M) — 1M context + reasoning + +### Vision / Multimodal + +Need: image input + text output + tool calling + +- **Free**: Gemma 4 27B IT — free vision + TC +- **Best value**: GPT-4.1-mini ($0.40/$1.60/M) — vision + 1M context +- **Premium**: Claude Sonnet 4 ($3/$15/M) — best vision understanding + +### Reasoning / Math + +Need: reasoning capability + structured output + +- **Free**: DeepSeek R1 — 92% MATH-500 +- **Best value**: Qwen3.5 4B ($0.03/$0.15/M) — cheap reasoning +- **Premium**: o3 ($10/$40/M) — SOTA on GPQA, MATH-500 + +### High-Volume Production + +Need: lowest cost per token + reliability + +- **Cheapest TC**: ling-2.6-flash ($0.01/$0.03/M) +- **Cheapest reasoning**: Qwen3.5 0.8B ($0.01/$0.05/M) +- **Cheapest coding**: bdc-coder ($0.01/$0.01/M) + +## 📊 Quick Stats + +| Metric | Count | +| ------------------- | ----- | +| Total models | 4,587 | +| Providers | 95 | +| Free models | 81 | +| Tool-calling models | 2,350 | +| Reasoning models | 1,306 | +| Vision models | 1,487 | +| Open-weight models | 527 | +| Structured output | 829 | + +## 🔗 Explore More + +- [Interactive Catalog](https://i-need-token.github.io/ai-models/) — search, filter, compare all models +- [Free Models Guide](free-models.md) — all 81 free models +- [Tool Calling Guide](tool-calling.md) — 2,350 models with tool calling +- [Pricing Comparison](pricing-comparison.md) — find the cheapest model +- [Context Windows](context-windows.md) — largest context windows +- [Model Comparison](model-comparison.md) — head-to-head comparisons + +--- + +_Data sourced from [AI Models Catalog](https://github.com/i-need-token/ai-models) — first-party data only, updated automatically._ diff --git a/docs/zh/model-selection-cheatsheet.md b/docs/zh/model-selection-cheatsheet.md new file mode 100644 index 00000000..407cf496 --- /dev/null +++ b/docs/zh/model-selection-cheatsheet.md @@ -0,0 +1,105 @@ +# AI 模型选择速查表 + +> 按使用场景快速选择 AI 模型的参考指南。数据来自 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 95 个提供商、4,587+ 模型。 + +## 🎯 决策树 + +``` +你需要什么？ +├── 最便宜的工具调用模型 → ling-2.6-flash ($0.01/$0.03/M) +├── 最佳免费推理模型 → DeepSeek R1 (92% MATH-500) +├── 最大上下文窗口 → Gemini 2.5 Pro (1M tokens) +├── 最佳编程助手 → Claude Sonnet 4 / GPT-4.1 +├── 开源工具调用模型 → Qwen3 4B ($0.03/$0.15/M) +├── 免费视觉模型 → Gemma 4 27B IT +└── 最低生产成本 → bdc-coder ($0.01/$0.01/M) +``` + +## 💰 按预算选择 + +| 预算 | 最佳选择 | 输入/输出 $/M | 理由 | +| ------------- | --------------- | ------------- | ---------------------- | +| **免费** | DeepSeek R1 | $0/$0 | 免费模型中推理能力最强 | +| **免费** | Gemma 4 27B IT | $0/$0 | 免费视觉 + 工具调用 | +| **< $0.05/M** | ling-2.6-flash | $0.01/$0.03 | 最便宜的工具调用模型 | +| **< $0.10/M** | Qwen3 4B | $0.03/$0.15 | 开源推理 + 工具调用 | +| **< $0.50/M** | GPT-4.1-mini | $0.40/$1.60 | 最佳性价比前沿模型 | +| **< $2/M** | Claude Sonnet 4 | $3/$15 | 顶级编程 + 推理 | +| **< $5/M** | GPT-4.1 | $2/$8 | 1M 上下文 + 视觉 | +| **高端** | o3 | $10/$40 | 最佳推理基准分数 | + +## 🛠️ 按使用场景选择 + +### AI Agent + +需求：工具调用 + 推理 + 低延迟 + +- **最佳性价比**：ling-2.6-flash ($0.01/$0.03/M) — 最便宜的 TC 模型 +- **均衡之选**：GPT-4.1-mini ($0.40/$1.60/M) — 可靠 + 1M 上下文 +- **高端选择**：Claude Sonnet 4 ($3/$15/M) — 最佳 Agent 性能 + +### 代码生成 + +需求：工具调用 + 结构化输出 + 大上下文 + +- **最佳性价比**：bdc-coder ($0.01/$0.01/M) — 最便宜的编程模型 +- **均衡之选**：GPT-4.1-mini ($0.40/$1.60/M) — 代码质量优秀 +- **高端选择**：Claude Sonnet 4 ($3/$15/M) — SWE-bench SOTA + +### 对话 / RAG + +需求：大上下文 + 低成本 + 快速响应 + +- **最佳性价比**：Qwen3 4B ($0.03/$0.15/M) — 便宜 + 262K 上下文 +- **均衡之选**：GPT-4.1-nano ($0.10/$0.40/M) — 快速 + 便宜 +- **高端选择**：Gemini 2.5 Pro ($1.25/$10/M) — 1M 上下文 + 推理 + +### 视觉 / 多模态 + +需求：图像输入 + 文本输出 + 工具调用 + +- **免费**：Gemma 4 27B IT — 免费视觉 + TC +- **最佳性价比**：GPT-4.1-mini ($0.40/$1.60/M) — 视觉 + 1M 上下文 +- **高端选择**：Claude Sonnet 4 ($3/$15/M) — 最佳视觉理解 + +### 推理 / 数学 + +需求：推理能力 + 结构化输出 + +- **免费**：DeepSeek R1 — 92% MATH-500 +- **最佳性价比**：Qwen3.5 4B ($0.03/$0.15/M) — 便宜推理 +- **高端选择**：o3 ($10/$40/M) — GPQA、MATH-500 SOTA + +### 大规模生产 + +需求：最低 token 成本 + 可靠性 + +- **最便宜 TC**：ling-2.6-flash ($0.01/$0.03/M) +- **最便宜推理**：Qwen3.5 0.8B ($0.01/$0.05/M) +- **最便宜编程**：bdc-coder ($0.01/$0.01/M) + +## 📊 快速统计 + +| 指标 | 数量 | +| ------------ | ----- | +| 模型总数 | 4,587 | +| 提供商 | 95 | +| 免费模型 | 81 | +| 工具调用模型 | 2,350 | +| 推理模型 | 1,306 | +| 视觉模型 | 1,487 | +| 开源模型 | 527 | +| 结构化输出 | 829 | + +## 🔗 更多资源 + +- [交互式目录](https://i-need-token.github.io/ai-models/) — 搜索、筛选、对比所有模型 +- [免费模型指南](free-models.md) — 81 个免费模型 +- [工具调用指南](tool-calling.md) — 2,350 个工具调用模型 +- [定价对比](pricing-comparison.md) — 找到最便宜的模型 +- [上下文窗口](context-windows.md) — 最大上下文窗口 +- [模型对比](model-comparison.md) — 面对面对比 + +--- + +_数据来自 [AI Models Catalog](https://github.com/i-need-token/ai-models) — 仅使用一手数据，自动更新。_ From e09281d3a791d9e524bdd9113f1a25ee8ae70aa7 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 12:11:07 +0800 Subject: [PATCH 294/311] Update index files and cross-links for Model Selection Cheatsheet --- AGENTS.md | 1 + README.md | 74 +++++++++++++++++++------------------- docs/model-selection.md | 3 +- docs/zh/model-selection.md | 3 +- llms-full.txt | 1 + llms.txt | 1 + 6 files changed, 45 insertions(+), 38 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index a371f062..2fc6dbaf 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -37,6 +37,7 @@ A structured catalog of AI model providers and their models, stored as YAML file - [`docs/data-schema.md`](docs/data-schema.md) — Data schema reference ([中文](docs/zh/data-schema.md)) - [`docs/quick-start.md`](docs/quick-start.md) — Quick start guide ([中文](docs/zh/quick-start.md)) - [`docs/model-selection.md`](docs/model-selection.md) — Model selection guide: free, best value, large context ([中文](docs/zh/model-selection.md)) +- [`docs/model-selection-cheatsheet.md`](docs/model-selection-cheatsheet.md) — Model selection cheatsheet: best model by budget and use case ([中文](docs/zh/model-selection-cheatsheet.md)) - [`docs/benchmarks.md`](docs/benchmarks.md) — AI Model Benchmarks & Leaderboards: key benchmarks, leaderboard landscape, interpretation guide ([中文](docs/zh/benchmarks.md)) - [`docs/migration-guide.md`](docs/migration-guide.md) — Switch providers: pricing, API compatibility, checklist ([中文](docs/zh/migration-guide.md)) - [`docs/api.md`](docs/api.md) — API & programmatic access ([中文](docs/zh/api.md)) diff --git a/README.md b/README.md index cc2aced0..7b8f2c07 100644 --- a/README.md +++ b/README.md @@ -476,6 +476,7 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin | [🔍 Interactive Catalog](https://i-need-token.github.io/ai-models/) | Search, sort, and filter all 4,587 models in your browser | | [Quick Start Guide](docs/quick-start.md) | Find the right model in 30 seconds | | [Model Selection Guide](docs/model-selection.md) | Decision framework: free, best value, large context models | +| [Model Selection Cheatsheet](docs/model-selection-cheatsheet.md) | Quick-reference: best model by budget and use case | | [Benchmarks & Leaderboards](docs/benchmarks.md) | MMLU, MATH, HumanEval, SWE-bench, Chatbot Arena guide | | [Migration Guide](docs/migration-guide.md) | Switch providers — pricing, API compatibility, checklist | | [API & Programmatic Access](docs/api.md) | Download models.json, code examples in JS/Python | @@ -501,42 +502,43 @@ See [`docs/data-acquisition.md`](docs/data-acquisition.md) for detailed guidelin **中文文档：** -| 文档 | 描述 | -| ------------------------------------------------- | ------------------------------------------------ | -| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | -| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | -| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | -| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | -| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | -| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | -| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | -| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | -| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | -| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | -| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | -| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | -| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | -| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | -| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | -| [大上下文模型](docs/zh/large-context-models.md) | 2,195 个 128K+ 上下文模型 — 397 个 1M+ | -| [小型/边缘模型](docs/zh/small-models.md) | 1,153 个 10B 参数以下模型，适合端侧部署 | -| [提供商对比](docs/zh/provider-comparison.md) | 按模型数量、能力、定价对比前 30 个提供商 | -| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | -| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | -| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | -| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | -| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | -| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | -| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | -| [智能体模型](docs/zh/agentic-models.md) | 1,080 个工具调用+推理模型，用于 AI 智能体 | -| [代码模型](docs/zh/code-models.md) | 189 个代码模型：生成、审查、调试 | -| [OpenAI 替代方案](docs/zh/openai-alternatives.md) | GPT-4/GPT-3.5 替代方案：定价、免费选项、兼容性 | -| [聊天模型](docs/zh/chat-models.md) | 2,350 个带工具调用的聊天模型 | -| [多模态模型](docs/zh/multimodal-models.md) | 1,519 个支持图像/音频/视频输入的模型 | -| [嵌入模型](docs/zh/embedding-models.md) | 5 个嵌入模型用于搜索、RAG、相似度 | -| [模型选择指南](docs/zh/model-selection.md) | 决策框架：免费、最佳性价比、大上下文模型 | -| [迁移指南](docs/zh/migration-guide.md) | 切换提供商：定价、API 兼容性、检查清单 | -| [术语表](docs/zh/glossary.md) | AI 模型术语的关键词和定义 | +| 文档 | 描述 | +| ------------------------------------------------------- | ------------------------------------------------ | +| [工具调用模型](docs/zh/tool-calling.md) | 2,350 个工具调用模型 — 最便宜、最大上下文、免费 | +| [视觉模型](docs/zh/vision-models.md) | 1,487 个视觉模型 — 最便宜、最大上下文、开源权重 | +| [快速入门](docs/zh/quick-start.md) | 30 秒内找到适合的模型 | +| [图像生成](docs/zh/image-generation.md) | 28 个图像生成模型 — DALL·E、Imagen、GPT-5 Image | +| [音频模型](docs/zh/audio-models.md) | 118 个音频输入 + 34 个音频输出模型 | +| [视频模型](docs/zh/video-models.md) | 167 个视频输入 + 4 个视频输出模型 | +| [API 与编程访问](docs/zh/api.md) | 下载 models.json，JS/Python 代码示例 | +| [代码示例](docs/zh/code-examples.md) | TypeScript、Python、Go、Rust、jq 实用示例 | +| [常见问题](docs/zh/faq.md) | 关于目录、数据和贡献的常见问题 | +| [结构化输出](docs/zh/structured-output.md) | 829 个 JSON 模式模型 — 最便宜、免费、带工具调用 | +| [模型对比](docs/zh/model-comparison.md) | 旗舰、高性价比、免费和开源模型对比 | +| [定价对比](docs/zh/pricing-comparison.md) | 各提供商和平台定价并排对比 | +| [缓存定价](docs/zh/cached-pricing.md) | 1,374 个支持提示缓存的模型 — 输入成本节省 50-90% | +| [模态矩阵](docs/zh/modality-matrix.md) | 视觉、图像生成、音频、视频 — 各模型支持什么 | +| [上下文窗口对比](docs/zh/context-windows.md) | 最大上下文窗口，各层级最佳性价比 | +| [大上下文模型](docs/zh/large-context-models.md) | 2,195 个 128K+ 上下文模型 — 397 个 1M+ | +| [小型/边缘模型](docs/zh/small-models.md) | 1,153 个 10B 参数以下模型，适合端侧部署 | +| [提供商对比](docs/zh/provider-comparison.md) | 按模型数量、能力、定价对比前 30 个提供商 | +| [免费 AI 模型](docs/zh/free-models.md) | 81 个免费模型 — 工具调用、推理、视觉零成本 | +| [开源权重模型](docs/zh/open-weights.md) | 513 个开源权重模型 — 自有基础设施运行 | +| [提供商概览](docs/zh/providers.md) | 95 个提供商按类型和市场分类 | +| [推理模型](docs/zh/reasoning-models.md) | 1,306 个推理模型 — 链式思维和扩展思考 | +| [数据 Schema 参考](docs/zh/data-schema.md) | 完整 YAML Schema — 模型、定价、快照、提供商 | +| [数据采集](docs/zh/data-acquisition.md) | 数据采集指南 | +| [设计原则与陷阱](docs/zh/lessons-learned.md) | 经验教训 | +| [智能体模型](docs/zh/agentic-models.md) | 1,080 个工具调用+推理模型，用于 AI 智能体 | +| [代码模型](docs/zh/code-models.md) | 189 个代码模型：生成、审查、调试 | +| [OpenAI 替代方案](docs/zh/openai-alternatives.md) | GPT-4/GPT-3.5 替代方案：定价、免费选项、兼容性 | +| [聊天模型](docs/zh/chat-models.md) | 2,350 个带工具调用的聊天模型 | +| [多模态模型](docs/zh/multimodal-models.md) | 1,519 个支持图像/音频/视频输入的模型 | +| [嵌入模型](docs/zh/embedding-models.md) | 5 个嵌入模型用于搜索、RAG、相似度 | +| [模型选择指南](docs/zh/model-selection.md) | 决策框架：免费、最佳性价比、大上下文模型 | +| [模型选择速查表](docs/zh/model-selection-cheatsheet.md) | 按预算和使用场景快速选择模型 | +| [迁移指南](docs/zh/migration-guide.md) | 切换提供商：定价、API 兼容性、检查清单 | +| [术语表](docs/zh/glossary.md) | AI 模型术语的关键词和定义 | ## Design Principles diff --git a/docs/model-selection.md b/docs/model-selection.md index d415dd54..d3a465fe 100644 --- a/docs/model-selection.md +++ b/docs/model-selection.md @@ -1,6 +1,7 @@ # Model Selection Guide -[中文](zh/model-selection.md) +> 📋 **Quick reference?** See the [Model Selection Cheatsheet](model-selection-cheatsheet.md) for a budget-by-budget guide. +> [中文](zh/model-selection.md) How to choose the right AI model for your use case — practical recommendations based on cost, capabilities, and context windows. diff --git a/docs/zh/model-selection.md b/docs/zh/model-selection.md index f1386fe5..7817d877 100644 --- a/docs/zh/model-selection.md +++ b/docs/zh/model-selection.md @@ -1,6 +1,7 @@ # 模型选择指南 -[English](../model-selection.md) +> 📋 **快速参考？** 查看[模型选择速查表](model-selection-cheatsheet.md)，按预算选择最佳模型。 +> [English](../model-selection.md) 如何根据使用场景选择合适的 AI 模型 — 基于成本、能力和上下文窗口的实用建议。 diff --git a/llms-full.txt b/llms-full.txt index 6a0113de..bada8717 100644 --- a/llms-full.txt +++ b/llms-full.txt @@ -315,6 +315,7 @@ by_context = sorted( - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds - [Model Selection Guide](docs/model-selection.md) — decision framework: free, best value, large context models +- [Model Selection Cheatsheet](docs/model-selection-cheatsheet.md) — quick-reference: best model by budget and use case - [Benchmarks & Leaderboards](docs/benchmarks.md) — key benchmarks, leaderboard landscape, interpretation guide - [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq diff --git a/llms.txt b/llms.txt index 4e0e67e5..a2059496 100644 --- a/llms.txt +++ b/llms.txt @@ -315,6 +315,7 @@ by_context = sorted( - [Quick Start](docs/quick-start.md) — find the right model in 30 seconds - [Model Selection Guide](docs/model-selection.md) — decision framework: free, best value, large context models +- [Model Selection Cheatsheet](docs/model-selection-cheatsheet.md) — quick-reference: best model by budget and use case - [Benchmarks & Leaderboards](docs/benchmarks.md) — key benchmarks, leaderboard landscape, interpretation guide - [API & Programmatic Access](docs/api.md) — download models.json, code examples - [Code Examples](docs/code-examples.md) — practical examples in TypeScript, Python, Go, Rust, jq From 237f85150764219591df8efd7fbb88ba30121b43 Mon Sep 17 00:00:00 2001 From: i-need-token Date: Fri, 22 May 2026 12:14:43 +0800 Subject: [PATCH 295/311] Add social preview image and update OG images across all pages --- site/ai-model-benchmarks.html | 4 ++-- site/ai-model-comparison-chart.html | 4 ++-- site/ai-model-pricing-calculator.html | 4 ++-- site/ai-models-by-provider.html | 4 ++-- site/best-ai-models-for-agents.html | 4 ++-- site/best-ai-models-for-coding.html | 4 ++-- site/best-ai-models-for-image-generation.html | 4 ++-- site/best-ai-models-for-vision.html | 4 ++-- site/best-ai-models.html | 4 ++-- site/chatgpt-vs-claude-vs-gemini.html | 4 ++-- site/cheapest-ai-models.html | 4 ++-- site/context-window-comparison.html | 4 ++-- site/free-ai-models.html | 4 ++-- site/index.html | 4 ++-- site/llm-pricing.html | 4 ++-- site/multimodal-ai-models.html | 4 ++-- site/open-source-ai-models.html | 4 ++-- site/openai-alternatives.html | 4 ++-- site/reasoning-models-comparison.html | 4 ++-- site/state-of-ai-models.html | 4 ++-- site/structured-output-models-comparison.html | 4 ++-- site/tool-calling-models-comparison.html | 4 ++-- 22 files changed, 44 insertions(+), 44 deletions(-) diff --git a/site/ai-model-benchmarks.html b/site/ai-model-benchmarks.html index e31555f4..e68e1b53 100644 --- a/site/ai-model-benchmarks.html +++ b/site/ai-model-benchmarks.html @@ -21,7 +21,7 @@ /> @@ -29,7 +29,7 @@ + + + +

    +

    🤏 Small Language Models (SLM) — 2,000+ Models Under 10B Parameters

    +

    + Complete guide to small language models for edge deployment, mobile apps, and cost-efficient + production. All data from + AI Models Catalog — first-party data + only. +

    + +
    +
    +
    2,002
    +
    Small Models
    +
    +
    +
    928
    +
    With Tool Calling
    +
    +
    +
    557
    +
    With Reasoning
    +
    +
    +
    48
    +
    Free SLMs
    +
    +
    +
    689
    +
    First-Party
    +
    +
    + + 🔍 Search All 4,587 Models → + +

    What Are Small Language Models?

    +

    + Small Language Models (SLMs) are AI models with fewer than ~10 billion parameters, designed + for efficiency, low latency, and deployment on resource-constrained hardware — from + smartphones to edge servers. They offer a practical alternative to large frontier models + when cost, speed, or privacy matters. +

    +

    Key advantages of SLMs:

    +
      +
    • Lower cost — often 10-100x cheaper per token than frontier models
      • +
    • Lower latency — faster inference for real-time applications
      • +
    • Edge deployment — run on-device without cloud dependency
      • +
    • Privacy — data never leaves the device
      • +
    • Fine-tuning — easier to customize for specific domains
      • +
    + +

    Cheapest Small Models with Tool Calling

    +

    Best value SLMs for AI agents and tool-use workflows (first-party providers only):

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/MOutput $/MContextReasoning
    ling-2.6-flashling$0.01$0.03262K
    klusterai--Meta-Llama-3.1-8B-Instruct-Turboklusterai$0.015$0.02131K
    granite-4.0-h-microibm$0.017$0.112131K
    llama-3.1-8b-instruct--fp-16fireworks$0.02$0.03131K
    schematron-3bfireworks$0.02$0.05131K
    + +

    Free Small Language Models

    +

    48 small models available at zero cost — perfect for prototyping and development:

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderContextTool CallingReasoning
    deepseek-r1-distill-llama-8bcerebras131K✓
    llama-4-scout-17b-16e-instructcerebras131K✓
    qwen-2.5-32bcerebras131K✓
    gemma-4-26b-a4b-itauriko262K✓
    glm-4.5-flashauriko200K✓
    glm-4.6v-flashauriko128K✓
    baidu--ernie-4.5-0.3baimlapi120K✓
    + +

    Small Models with Reasoning

    +

    + 557 small models with reasoning capabilities — ideal for math, logic, and step-by-step + problem solving: +

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    ModelProviderInput $/MOutput $/MContextTool Calling
    qwen3.5-0.8bqwen$0.01$0.05262K
    qwen3.5-2bqwen$0.02$0.10262K
    qwen--qwen3-4b-fp8fireworks$0.03$0.03128K
    qwen3.5-4bqwen$0.03$0.15262K
    deepseek-r1-distill-llama-8bcerebrasFreeFree131K
    + +

    Best SLMs by Use Case

    +

    🤖 AI Agents on a Budget

    +

    + ling-2.6-flash ($0.01/$0.03/M) — cheapest tool-calling model with 262K + context. Perfect for high-volume agent workflows. +

    +

    📱 On-Device / Edge Deployment

    +

    + Qwen3.5 0.8B — ultra-compact reasoning model. + Gemma 4 27B IT — free with vision + tool calling. +

    +

    💻 Code Completion

    +

    + bdc-coder ($0.01/$0.01/M) — cheapest coding model. + Qwen3 4B ($0.03/$0.15/M) — open-source with reasoning. +

    +

    🧮 Math & Reasoning

    +

    + DeepSeek R1 Distill Llama 8B — free reasoning model. + Qwen3.5 0.8B ($0.01/$0.05/M) — cheapest reasoning. +

    +

    💬 Chat & RAG

    +

    + GPT-4.1-nano ($0.10/$0.40/M) — fast, cheap, reliable. + Qwen3 4B ($0.03/$0.15/M) — open-source alternative. +

    + +

    SLM vs LLM: When to Choose Small

    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    FactorSmall Model (SLM)Large Model (LLM)
    Cost per 1M tokens$0.01 – $0.20$1 – $40
    Latency (first token)50 – 200ms200 – 2000ms
    DeploymentOn-device, edge, cloudCloud only
    PrivacyData stays on deviceData sent to cloud
    CustomizationEasy fine-tuningExpensive fine-tuning
    Complex reasoningGood for simple tasksSuperior for complex tasks
    Best forHigh-volume, real-time, edgeComplex, nuanced, creative
    + +
    + Related: + Free AI Models + Cheapest AI Models + Tool Calling Models + Reasoning Models + Open Source Models + Best for Coding + Best for Agents + Comparison Chart + Context Windows + LLM Pricing + Best AI Models + State of AI Models + Benchmarks + ChatGPT vs Claude vs Gemini + By Provider + Multimodal Models + Structured Output + Vision Models + Image Generation + OpenAI Alternatives + Pricing Calculator +
    + + +
    + + diff --git a/site/state-of-ai-models.html b/site/state-of-ai-models.html index 27283fe4..5da2ac0d 100644 --- a/site/state-of-ai-models.html +++ b/site/state-of-ai-models.html @@ -715,6 +715,7 @@

    7. Key Trends & Predictions

    Benchmarks + Small Language Models

    Data from diff --git a/site/structured-output-models-comparison.html b/site/structured-output-models-comparison.html index 9f3ba26e..31059bfa 100644 --- a/site/structured-output-models-comparison.html +++ b/site/structured-output-models-comparison.html @@ -725,6 +725,7 @@

    🔗 More Resources

    • + Small Language Models