[BOT ISSUE] fix: add missing Google model gemini-3.1-flash-lite (stable)

[github-actions]

Missing model

The model gemini-3.1-flash-lite (the stable/GA release) is available on Google AI but absent from the model catalog (packages/proxy/schema/model_list.json). The catalog currently has only the preview variant gemini-3.1-flash-lite-preview.

Google's models page lists gemini-3.1-flash-lite with a "Stable" badge, and the Google AI pricing page lists it as a distinct entry from the preview version with identical pricing.

Model details

Field	Value
Model ID	gemini-3.1-flash-lite
Format	google
Flavor	chat
Multimodal	true (text, image, video, audio, PDF input)
Reasoning	true (thinking supported)
Reasoning budget	true
Input cost (per 1M tokens)	$0.25
Output cost (per 1M tokens)	$1.50
Cache read cost (per 1M tokens)	$0.025
Max input tokens	1048576
Max output tokens	65536
Available providers	["google"]
Display name	Gemini 3.1 Flash-Lite

Verification checklist

• Cross-source confirmation: Model confirmed on (1) Google AI models overview page with "Stable" status badge, (2) Google AI pricing page with both standard and cache pricing listed, and (3) Google AI model detail page with token limits and capabilities.
• Recent commits check: Not added in any of the last 30 commits touching model_list.json. Recent commits cover Bedrock Nova, Mistral Magistral, and Together models only.
• ID format validation: Follows existing convention — gemini-2.5-flash-lite (stable) and gemini-3.1-flash-lite-preview both exist in the catalog, so gemini-3.1-flash-lite is the natural stable counterpart.

Verification notes

Field	Source
Model ID, token limits, capabilities	Model detail page
"Stable" status	Models overview page
Standard pricing ($0.25/$1.50)	Pricing page
Cache read pricing ($0.025)	Pricing page — listed under context caching section
reasoning, reasoning_budget	Inferred from "Thinking: Supported" on model detail page, consistent with preview variant

Fields NOT verified from official sources:

• supported_regions — not published on Google AI docs; Vertex AI pricing page lists only the preview variant, so Vertex availability for the stable version is unconfirmed. Only google is listed as an available provider.
• max_output_tokens discrepancy note — the model detail page reports 65,536; the preview variant in the catalog uses 65,536 for the Google entry and 65,535 for the Vertex entry.

{
  "kind": "missing_model",
  "provider": "google",
  "models": ["gemini-3.1-flash-lite"],
  "status": "active",
  "model_specs": {
    "gemini-3.1-flash-lite": {
      "format": "google",
      "flavor": "chat",
      "multimodal": true,
      "input_cost_per_mil_tokens": 0.25,
      "output_cost_per_mil_tokens": 1.5,
      "input_cache_read_cost_per_mil_tokens": 0.025,
      "displayName": "Gemini 3.1 Flash-Lite",
      "reasoning": true,
      "reasoning_budget": true,
      "max_input_tokens": 1048576,
      "max_output_tokens": 65536,
      "available_providers": ["google"]
    }
  },
  "source_urls": [
    "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-lite",
    "https://ai.google.dev/gemini-api/docs/models/gemini",
    "https://ai.google.dev/pricing"
  ]
}