feat(google-gemini/gemini-3.5-live-translate-preview): add new models [bot]#1319
Conversation
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
|
/test-models |
Gateway test results
Skipped (1)
Skip reason |
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
There are 2 total unresolved issues (including 1 from previous review).
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit bb5bae9. Configure here.
| - defaultValue: 0.95 | ||
| key: top_p | ||
| - defaultValue: 64 | ||
| key: top_k |
There was a problem hiding this comment.
Missing max_tokens metadata
Medium Severity
limits sets max_output_tokens to 32768 but omits matching max_tokens, and the model does not override the provider max_tokens param. Merged config still caps max_tokens at 4096 from default.yaml, which disagrees with the declared output limit for this preview model.
Additional Locations (1)
Reviewed by Cursor Bugbot for commit bb5bae9. Configure here.
|
/test-models |
Gateway test results
Skipped (1)
Skip reason |
1 participant
Auto-generated by model-addition-agent for
google-gemini/gemini-3.5-live-translate-preview.Note
Low Risk
Declarative model metadata only; no runtime, auth, or application logic changes.
Overview
Adds a new google-gemini provider entry for
gemini-3.5-live-translate-preview, registering Google’s preview live audio translation model in the catalog.The config marks it as realtime / serverless with preview status, audio-only input and output, and supportedModes
realtimeandaudio_translation. Pricing uses per-audio-token input/output rates; limits include a 131072 context window and 65536 max output tokens. Sampling defaults match other Gemini live previews (temperature,top_p,top_k), andtool_choice/reasoning_effortare listed underremoveParams.Reviewed by Cursor Bugbot for commit 2983b76. Bugbot is set up for automated code reviews on this repo. Configure here.