groq

Documentation

Overview

Package groq wraps Groq's OpenAI-compatible API. Groq runs open- weight models (Llama, Gemma, DeepSeek, Kimi) on its in-house LPUs at extremely high throughput.

Groq-specific knobs reachable through the namespaced OpenAI request extension:

• service_tier ("on_demand" / "flex" / "auto") trades cost for latency. See https://console.groq.com/docs/flex-processing.
• reasoning_format ("parsed" / "raw" / "hidden") controls how reasoning-model output is surfaced.

See https://console.groq.com/docs/ for the full API reference.

Index

• Constants
• type Chat
  • func NewChat(config ChatConfig) (*Chat, error)
• type ChatConfig
  • func (c ChatConfig) Validate() error

Constants

const (
	OpenAIRequestExtensionKey     = "groq/openai_request"
	OpenAIResponseExtensionKey    = "groq/openai_response"
	OpenAIStreamChunkExtensionKey = "groq/openai_stream_chunk"
)

Namespacing preserves provider-specific data without promoting it into the shared Core protocol or colliding with another provider.

const (
	Provider = "Groq"

	BaseURL = "https://api.groq.com/openai/v1"
)

Exported identifiers keep provider-owned names and defaults out of caller literals.

const (
	ModelGPTOSS120B = "openai/gpt-oss-120b"
	ModelGPTOSS20B  = "openai/gpt-oss-20b"
	ModelQwen36_27B = "qwen/qwen3.6-27b"
)

Current production model ids. See https://console.groq.com/docs/models.

Types

type Chat

type Chat = openai.ChatCompletions

Chat implements Groq's chat endpoint.

func NewChat

func NewChat(config ChatConfig) (*Chat, error)

NewChat rejects an invalid provider binding before the first chat call.

type ChatConfig

type ChatConfig struct {
	APIKey         string
	DefaultOptions corechat.Options
	BaseURL        string
	HTTPClient     *http.Client
}

ChatConfig binds provider access and defaults shared by every chat call.

func (ChatConfig) Validate

func (c ChatConfig) Validate() error