Provider Compatibility

Provider Compatibility Table

Provider	Type	Auth	Streaming	Tool Use	Model Slots	Notes
Anthropic	DirectAnthropic	API Key	Yes	Yes	Yes	Native, no translation
MiniMax	DirectAnthropic	API Key	Yes	Yes	Yes	Anthropic-compatible proxy
Google Vertex AI	DirectAnthropic	Access Token	Yes	Yes	Yes	gcloud auth print-access-token
OpenRouter	OpenAICompatible	API Key	Yes	Yes	Yes	200+ models via single API
Grok (xAI)	OpenAICompatible	API Key	Yes	Yes	Yes
OpenAI	OpenAICompatible	API Key	Yes	Yes	Yes
DeepSeek	OpenAICompatible	API Key	Yes	Yes	Yes
Kimi / Moonshot	OpenAICompatible	API Key / OAuth	Yes	Yes	Yes
GLM (Zhipu)	OpenAICompatible	API Key	Yes	Yes	Yes
Groq	OpenAICompatible	API Key	Yes	Yes	Yes	Ultra-fast inference
Mistral AI	OpenAICompatible	API Key	Yes	Yes	Yes
Together AI	OpenAICompatible	API Key	Yes	Yes	Yes
Perplexity	OpenAICompatible	API Key	Yes	Limited	Yes	Built-in web search
Cerebras	OpenAICompatible	API Key	Yes	Yes	Yes	Fast inference
Azure OpenAI	OpenAICompatible	API Key	Yes	Yes	Yes	api-key header + query_params
GitHub Copilot	OpenAICompatible	OAuth	Yes	Yes	Yes	Device Code flow
GitLab Duo	OpenAICompatible	OAuth (PAT)	Yes	Yes	Yes	GITLAB_TOKEN env
ChatGPT/Codex	OpenAIResponses	OAuth	Yes	Yes	Yes	Responses API
Claude (Sub)	DirectAnthropic	OAuth	Yes	Yes	Yes	Skips proxy
Google Gemini (Sub)	OpenAICompatible	OAuth	Yes	Yes	Yes	Gemini CLI credentials
Qwen (Sub)	OpenAICompatible	OAuth	Yes	Yes	Yes	Device Code flow
Ollama	OpenAICompatible	None	Yes	Model-dependent	Yes	Local, no API key
vLLM	OpenAICompatible	None	Yes	Model-dependent	Yes	Local
LM Studio	OpenAICompatible	None	Yes	Model-dependent	Yes	Local
Bedrock (LiteLLM)	OpenAICompatible	API Key	Yes	Yes	Yes	Via LiteLLM proxy

Provider Details

Anthropic

• Base URL: https://api.anthropic.com
• Type: DirectAnthropic (no translation needed)
• Auth: API Key (sk-ant-*)
• Models: claude-opus-4, claude-sonnet-4, claude-haiku-4
• Get API Key: console.anthropic.com

MiniMax

• Base URL: https://api.minimax.io/anthropic
• Type: DirectAnthropic (no translation needed)
• Auth: API Key
• Models: Uses Anthropic model names (proxied)
• Get API Key: platform.minimaxi.com

Google Vertex AI

• Base URL: https://REGION-aiplatform.googleapis.com/v1/projects/PROJECT/locations/REGION/publishers/anthropic/models
• Type: DirectAnthropic
• Auth: GCP Access Token (gcloud auth print-access-token)
• Models: claude-sonnet-4@20250514, claude-opus-4@20250514
• Setup: Vertex AI Console

OpenRouter

• Base URL: https://openrouter.ai/api/v1
• Type: OpenAICompatible
• Auth: API Key (sk-or-*)
• Models: 200+ models from multiple providers
• Get API Key: openrouter.ai/keys

Grok (xAI)

• Base URL: https://api.x.ai/v1
• Type: OpenAICompatible
• Auth: API Key (xai-*)
• Models: grok-3-beta, grok-3-mini-beta
• Get API Key: console.x.ai

OpenAI

• Base URL: https://api.openai.com/v1
• Type: OpenAICompatible
• Auth: API Key (sk-*)
• Models: gpt-4o, gpt-4o-mini, o1, o1-mini
• Get API Key: platform.openai.com/api-keys

DeepSeek

• Base URL: https://api.deepseek.com
• Type: OpenAICompatible
• Auth: API Key
• Models: deepseek-chat, deepseek-reasoner
• Get API Key: platform.deepseek.com

Kimi / Moonshot

• Base URL: https://api.moonshot.ai/v1
• Type: OpenAICompatible
• Auth: API Key / OAuth
• Models: kimi-k2-0905-preview, moonshot-v1-128k
• Get API Key: platform.moonshot.cn

GLM (Zhipu)

• Base URL: https://api.z.ai/api/paas/v4
• Type: OpenAICompatible
• Auth: API Key
• Models: glm-4.6, glm-4-plus
• Get API Key: open.bigmodel.cn

Groq

• Base URL: https://api.groq.com/openai/v1
• Type: OpenAICompatible
• Auth: API Key (gsk_*)
• Models: llama-3.3-70b-versatile, mixtral-8x7b-32768, gemma2-9b-it
• Get API Key: console.groq.com
• Note: Extremely fast inference for open-source models

Mistral AI

• Base URL: https://api.mistral.ai/v1
• Type: OpenAICompatible
• Auth: API Key
• Models: mistral-large-latest, mistral-small-latest
• Get API Key: console.mistral.ai

Together AI

• Base URL: https://api.together.xyz/v1
• Type: OpenAICompatible
• Auth: API Key
• Models: meta-llama/Llama-3.3-70B-Instruct-Turbo, and many open-source models
• Get API Key: api.together.ai

Perplexity

• Base URL: https://api.perplexity.ai
• Type: OpenAICompatible
• Auth: API Key (pplx-*)
• Models: sonar-pro, sonar
• Get API Key: perplexity.ai/settings/api
• Note: Models include built-in web search capabilities

Cerebras

• Base URL: https://api.cerebras.ai/v1
• Type: OpenAICompatible
• Auth: API Key
• Models: llama-3.3-70b
• Get API Key: cloud.cerebras.ai
• Note: Fast inference platform

Azure OpenAI

• Base URL: https://YOUR_RESOURCE.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT
• Type: OpenAICompatible
• Auth: API Key (via api-key header, auto-detected)
• Models: Depends on deployment (gpt-4o, gpt-4, etc.)
• Setup: Azure Portal
• Note: Requires api-version in query_params

GitHub Copilot

• Base URL: https://api.githubcopilot.com
• Type: OpenAICompatible
• Auth: OAuth (Device Code flow)
• Models: gpt-4o (default)
• Note: Requires active GitHub Copilot subscription

GitLab Duo

• Base URL: https://gitlab.com/api/v4/ai/llm/proxy (or self-hosted)
• Type: OpenAICompatible
• Auth: OAuth (Personal Access Token via GITLAB_TOKEN)
• Models: claude-sonnet-4-20250514
• Note: Supports --enterprise-url for self-hosted instances

ChatGPT / Codex (Subscription)

• Base URL: https://chatgpt.com/backend-api/codex
• Type: OpenAIResponses
• Auth: OAuth (Browser PKCE / Device Code)
• Models: gpt-5.3-codex, gpt-5.1-codex, codex-mini-latest
• Note: Uses Responses API, not Chat Completions. strip_params = "auto" strips unsupported params.

Claude (Subscription)

• Base URL: https://api.claude.ai
• Type: DirectAnthropic
• Auth: OAuth (reads ~/.claude/.credentials.json)
• Models: claude-sonnet-4, claude-opus-4, claude-haiku-4
• Note: Proxy is skipped. Claude Code uses its own OAuth directly.

Google Gemini (Subscription)

• Base URL: https://generativelanguage.googleapis.com/v1beta/openai
• Type: OpenAICompatible
• Auth: OAuth (Gemini CLI credentials)
• Models: gemini-2.5-pro, gemini-2.0-flash

Qwen (Subscription)

• Base URL: https://chat.qwen.ai/api
• Type: OpenAICompatible
• Auth: OAuth (Device Code flow)
• Models: qwen3-235b-a22b, qwen-max

Ollama

• Base URL: http://localhost:11434/v1
• Type: OpenAICompatible
• Auth: None required
• Models: Any pulled model (qwen2.5, llama3.3, etc.)
• Setup: ollama.com

vLLM

• Base URL: http://localhost:8000/v1
• Type: OpenAICompatible
• Auth: None required
• Models: Depends on served model

LM Studio

• Base URL: http://localhost:1234/v1
• Type: OpenAICompatible
• Auth: None / "lm-studio"
• Models: Depends on loaded model
• Setup: lmstudio.ai

AWS Bedrock (via LiteLLM)

• Base URL: http://localhost:4000/v1 (LiteLLM proxy)
• Type: OpenAICompatible
• Auth: API Key (LiteLLM key)
• Models: bedrock/anthropic.claude-sonnet-4-20250514-v2:0
• Setup: docs.litellm.ai/docs/providers/bedrock