OpenAI Compat

This document lists popular OpenAI-compatible LLM providers, their base URLs, API key names (for use in environment variables or config), and where to obtain API keys.

Provider Architecture

flowchart TB
    subgraph "OpenAI-Compatible Providers"
        OAI[OpenAI<br/>gpt-4o]
        COH[Cohere<br/>command-r-08-2024]
        OR[OpenRouter<br/>google/gemini-2.0-flash]
        TOG[Together AI<br/>meta-llama/Llama-3.2-3B]
        GOA[Google AI OpenAI<br/>gemini-2.0-flash]
        OLL[Ollama OpenAI<br/>llama3.2]
    end

    subgraph "Native API Providers"
        GOO[Google<br/>gemini-2.0-flash]
        ANT[Anthropic<br/>claude-sonnet-4-0]
        MIS[Mistral<br/>open-mistral-7b]
        OLN[Ollama Native<br/>qwen2.5:7b-instruct]
    end

    OAI --> OAPIC[OpenAIProvider Class]
    COH --> OAPIC
    OR --> OAPIC
    TOG --> OAPIC
    GOA --> OAPIC
    OLL --> OAPIC

    GOO --> GOOP[GoogleProvider Class]
    ANT --> ANTP[AnthropicProvider Class]
    MIS --> MISP[MistralProvider Class]
    OLN --> OLNP[OllamaProvider Class]

    OAPIC --> API1["/v1/chat/completions"]
    GOOP --> API2["/v1beta/models/.../generateContent"]
    ANTP --> API3["/v1/messages"]
    MISP --> API4["/v1/chat/completions"]
    OLNP --> API5["/api/chat"]

    style OAI fill:#f9f,stroke:#333,stroke-width:2px
    style COH fill:#f9f,stroke:#333,stroke-width:2px
    style OR fill:#f9f,stroke:#333,stroke-width:2px
    style TOG fill:#f9f,stroke:#333,stroke-width:2px
    style GOA fill:#f9f,stroke:#333,stroke-width:2px
    style OLL fill:#f9f,stroke:#333,stroke-width:2px

    style GOO fill:#9ff,stroke:#333,stroke-width:2px
    style ANT fill:#9ff,stroke:#333,stroke-width:2px
    style MIS fill:#9ff,stroke:#333,stroke-width:2px
    style OLN fill:#9ff,stroke:#333,stroke-width:2px

Loading

OpenAI-Compatible Providers

Provider	Base URL	API Key Name	Default Model	Get API Key	Notes
OpenAI	https://api.openai.com/v1	OPENAI_API_KEY	gpt-4o	OpenAI API Keys
Cohere	https://api.cohere.com/compatibility/v1	COHERE_API_KEY	command-r-08-2024	Cohere API Keys	OpenAI-compatible endpoint
OpenRouter	https://openrouter.ai/api/v1	OPENROUTER_API_KEY	google/gemini-2.0-flash	OpenRouter API Keys
Together AI	https://api.together.xyz/v1	TOGETHER_API_KEY	meta-llama/Llama-3.2-3B-Instruct-Turbo	Together API Keys	Tool support disabled due to non-standard streaming format
Google AI (OpenAI)	https://generativelanguage.googleapis.com/v1beta/openai	GEMINI_API_KEY	gemini-2.0-flash	Google AI Studio API Keys	OpenAI-compatible Gemini endpoint
Ollama (OpenAI-compat)	http://localhost:11434/v1	(none)	llama3.2	Ollama Docs	Local OpenAI-compatible endpoint

Native API Providers (Not OpenAI-Compatible)

These providers use their own native APIs in dartantic and are not OpenAI-compatible:

Provider	API Key Name	Default Model	Notes
Google	GEMINI_API_KEY	gemini-2.0-flash	Native Gemini API
Anthropic	ANTHROPIC_API_KEY	claude-sonnet-4-0	Native Claude API
Mistral	MISTRAL_API_KEY	open-mistral-7b	Native Mistral API
Ollama	(none)	qwen2.5:7b-instruct	Native Ollama API at /api endpoint

Notes:

• All OpenAI-compatible providers use the /chat/completions endpoint and the Authorization: Bearer <API_KEY> header.
• Some providers may require you to specify a model name in your request payload.
• For Google AI OpenAI-compatible endpoint, use your Google AI Studio API key (GEMINI_API_KEY) and the special base URL as shown above. See the official Gemini OpenAI compatibility docs for details.
• Together AI has tool support disabled due to its non-standard streaming format that uses <|python_tag|> prefix instead of standard OpenAI tool_calls format.
• For a more comprehensive and up-to-date list of OpenAI-compatible providers, see cheahjs/free-llm-api-resources.