feat(model): add latest OpenAI model presets#416
Conversation
Add GPT-5.5, refresh GPT-5 context metadata, and add current OpenAI speech models.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c4ae257783
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| list: [ | ||
| { | ||
| type: ModelTypeEnum.llm, | ||
| model: 'gpt-5.5', |
There was a problem hiding this comment.
When this provider list is returned from initModels, users can select this new preset, but OpenAI's current public model catalog lists gpt-5.2/gpt-5.2-pro as the latest GPT-5 models and does not expose gpt-5.5 or gpt-5.5-pro (https://platform.openai.com/docs/models). In any environment that picks this added gpt-5.5 entry, the first OpenAI request will fail with a model-not-found error instead of using a valid latest model ID, so this should be removed or changed to a published alias.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
Pull request overview
Updates the OpenAI provider’s model preset catalog to include the latest GPT-5.5 LLM presets, refreshed GPT-5.x limits/metadata, and current OpenAI TTS/STT presets (including shared voice lists).
Changes:
- Added
gpt-5.5andgpt-5.5-proLLM presets and refreshed GPT-5.x context/output limits. - Introduced shared OpenAI TTS voice lists and added
gpt-4o-mini-tts,tts-1-hd, plus new STT presets. - Adjusted
gpt-5.4-prometadata (notably response format handling).
Comments suppressed due to low confidence (1)
modules/model/provider/OpenAI/index.ts:117
- The
gpt-5.4-propreset no longer setsresponseFormatList, while other GPT-5.x presets still do. If downstream UI/config logic usesresponseFormatListto determine available response formats, this regression may remove format options forgpt-5.4-pro. Consider restoring an explicitresponseFormatList(even if it’s just['text']or whatever is supported) to keep behavior consistent across GPT-5.x presets.
type: ModelTypeEnum.llm,
model: 'gpt-5.4-pro',
maxContext: 1050000,
maxTokens: 128000,
quoteMaxToken: 1000000,
maxTemperature: null,
vision: true,
reasoning: true,
reasoningEffort: true,
toolChoice: true
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| const legacyTtsVoices = ttsVoices.filter( | ||
| ({ value }) => !['ballad', 'verse', 'marin', 'cedar'].includes(value) |
…update-20260507-clean # Conflicts: # modules/model/provider/OpenAI/index.ts
|
✅ Build Successful - Preview Images for this PR: Changed packages: |
Summary
Verification
Full bun test still has unrelated existing environment/test failures in docDiff, Feishu, YouTube, and SDK tests.