feat: recognize gpt-oss family in getModelFamily / isSupportedModel - #362
Open
Luckycat133 wants to merge 1 commit into
Open
feat: recognize gpt-oss family in getModelFamily / isSupportedModel#362Luckycat133 wants to merge 1 commit into
Luckycat133 wants to merge 1 commit into
Conversation
The Antigravity desktop app now exposes OpenAI's open-weight GPT-OSS 120B model alongside the Claude and Gemini tiers. The proxy's request validation rejects any non-Claude/Gemini family name with "Invalid model: gpt-oss-120b-medium", so a Claude Code client pointing at the proxy cannot reach the GPT-OSS endpoint. getModelFamily() is the single chokepoint (its result gates listModels, getModelQuotas, isValidModel). Adding a third family branch unblocks GPT-OSS without touching request/response conversion — those fall through the family-agnostic path, which uses the same Google Generative-AI envelope the Gemini path uses upstream. Out of scope (deferred to follow-ups after observing quota / tool-use behaviour upstream): - MODEL_FALLBACK_MAP entries for gpt-oss - request/response converter branches for gpt-oss-specific tool_use / thinking handling
Luckycat133
added a commit
to Luckycat133/crouter
that referenced
this pull request
Aug 8, 2026
The upstream PR is now live at badrisnarayanan/antigravity-claude-proxy#362, so the "How to submit" walkthrough is obsolete. Replace it with a status section pointing at the live PR, plus the local-fallback paragraph that remains useful until #362 merges. Co-Authored-By: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The Antigravity desktop app now exposes OpenAI's open-weight
GPT-OSS 120B (Medium)model alongside the Claude and Gemini tiers. The proxy currently rejects these names inisValidModel()becausegetModelFamily()only recognises theclaude/geminisubstrings, so anycrouter antigravity-claude --model gpt-oss-120b-medium(or direct API call to/v1/messages) returnsinvalid_request_error: Invalid model: gpt-oss-120b-medium.This adds a third family branch for any name containing
gpt-ossand extendsisSupportedModel()accordingly. Request/response conversion falls through the family-agnostic path, which uses the same Google Generative-AI envelope the Gemini path uses upstream.Patch scope is intentionally minimal — no
MODEL_FALLBACK_MAPentries (need observed quota behavior first) and no converter changes (need observed tool-use / thinking behavior first). Happy to follow up with either once the validation change is in production.Companion: Luckycat133/crouter
b80e16cships a local sed-based patcher (bin/antigravity-proxy-patch) that applies the same change idempotently on a local proxy checkout, so crouter users can use GPT-OSS today without waiting for the upstream merge.