Replace decommissioned Groq model defaults

[umarjaved204]

Replace decommissioned Groq model defaults

Problem

Groq has retired llama-3.3-70b-versatile and llama-3.1-8b-instant. They no longer appear in Groq's /models list, and requests to them fail with model_not_found.

groqService.js uses both as its default primary and fallback models. Any deployment that switches to ASSISTANT_PROVIDER=groq without setting GROQ_MODEL would fail on every chat request.

Changes

• src/modules/chat/services/groqService.js: default models changed to openai/gpt-oss-120b (primary) and openai/gpt-oss-20b (fallback)
• .env.example: commented model names updated to match

Impact

• The default provider (ASSISTANT_PROVIDER=polymentor) is not affected. Those requests go to the PolyMentor service.
• Environment variables still override the defaults. If GROQ_MODEL is set to a llama model in any deployment, it needs to be updated.

Testing

• Checked the models available on Groq: both new models are listed, both old ones are not
• Tested openai/gpt-oss-120b with a chat request through the PolyMentor service locally, and it returned a correct answer
• groqService.js loads without errors
• Not tested: a full request through this backend's Groq path

Related

Companion PR in PolyMentor: [link], which fixes the chat outage itself.

[vercel]

@umarjaved204 is attempting to deploy a commit to the seno-quantum-coder's projects Team on Vercel.

A member of the Team first needs to authorize it.