Groq

Groq's OpenAI-compatible endpoint at https://api.groq.com/openai/v1. The Groq provider is a thin extension of the OpenAI provider — it swaps the base URL and otherwise reuses the OpenAI Chat Completions SDK call path verbatim.

Connection config

Field	Required	Description
`apiKey`	yes	Groq API key. Create one at console.groq.com → API keys.

{
  "provider": "groq",
  "config": { "apiKey": "gsk_..." }
}

Endpoint passthrough

Client request shape	What hits the wire
Chat Completions	Verbatim — Groq is OpenAI-compatible.
Anthropic Messages	Direct Anthropic ↔ Chat Completions converter (single hop — no internal pivot).
Responses	Direct Responses ↔ Chat Completions converter (single hop — no internal pivot).

Capabilities

Capability	Status
Streaming	✅
Function / tool calling	✅ — best on `llama-3.3-70b-versatile` and the larger Mixtral variants
Service tier	✅ — `service_tier: "on_demand"`

`providerArgs` — common Groq-native fields

{
  "vmx": {
    "providerArgs": {
      "service_tier": "on_demand"
    }
  }
}

Models

Pass any model id Groq exposes — llama-3.3-70b-versatile, llama-3.1-8b-instant, gemma2-9b-it, …

Notes

Groq's compat endpoint rejects unknown top-level fields and rejects tools[].function.strict: null. The gateway strips the vmx / __vmx_passthrough envelopes and omits strict when unset (rather than passing null).
Tool-call reliability is best on the larger Llama models; the smaller 8b-instant is fast but less consistent at function calling. Prefer llama-3.3-70b-versatile when tools matter.
The Anthropic Messages and Responses cells use the canonical cross-format converters in packages/api/src/ai-provider/adapters/anthropic-messages.adapter.ts and gateway/responses/responses-converter.ts — see the API Endpoints pages for the per-pair conversion contract.

Connection config​

Endpoint passthrough​

Capabilities​

providerArgs — common Groq-native fields​

Models​

Notes​