Repository navigation
fix(groq): gpt-oss models take low, medium or high reasoning effort - #416
Merged
Merged
Conversation
Groq's gpt-oss-20b and gpt-oss-120b listed reasoning_effort as none|default (default "default"), which are Qwen-only values. Groq accepts low, medium or high on these models, defaults to medium and rejects anything else with a 400. https://console.groq.com/docs/reasoning https://console.groq.com/docs/api-reference
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This branch was successfully deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
groq/gpt-oss-20bandgroq/gpt-oss-120blistedreasoning_effortasnone | default(defaultdefault). Those are Qwen values on Groq. For the gpt-oss models Groq accepts onlylow,mediumorhigh, defaults tomedium, and returns a 400 for anything outside that set. So a client trusting the catalog would send"none"and get rejected.Sources:
noneanddefaultare only supported by Qwen 3.8 27B"; low/medium/high for GPT-OSS 20B and 120B)reasoning_effort: "'medium' is the default value"; "Values outside a model's supported set are rejected with a 400")Changes:
models/groq/gpt-oss-20b.yaml,models/groq/gpt-oss-120b.yaml:reasoning_effortvalues[low, medium, high], defaultmedium. Description matches theopenai/gpt-oss-*entries.I left
gpt-oss-safeguard-20balone. Its model page mentions low/medium/high effort in prose, but the reasoning docs list only 20B and 120B as supporting the parameter.Verified with
npm run validate,lint,typecheck,test,buildandguard:params, plus the modelparams package's typecheck/build/test/test:types and the Python package's mypy and pytest. All pass.