> ## Documentation Index
> Fetch the complete documentation index at: https://docs.modelslab.com/llms.txt
> Use this file to discover all available pages before exploring further.

# LLM API

> Access 200+ large language models through a unified API. OpenAI and Anthropic SDK compatible — use Claude Code, ChatGPT clients, and more with a single API key.

ModelsLab's LLM API gives you access to 200+ large language models through a single, unified API. It's fully compatible with both **OpenAI** and **Anthropic** SDKs, so you can use tools like **Claude Code**, ChatGPT-compatible clients, and any OpenAI/Anthropic SDK with minimal configuration changes.

## Base URL

```
https://modelslab.com/api/v7/llm
```

## Authentication

All endpoints use **Bearer token** authentication. Pass your ModelsLab API key in the `Authorization` header:

```
Authorization: Bearer YOUR_MODELSLAB_API_KEY
```

## Endpoints

<CardGroup cols={2}>
  <Card title="Chat Completions" href="/llm-api/chat-completions" icon="comments">
    OpenAI-compatible chat completions endpoint. Works with OpenAI SDKs, LangChain, and any OpenAI-compatible client.
  </Card>

  <Card title="Messages" href="/llm-api/messages" icon="message">
    Anthropic-compatible messages endpoint. Works with the Anthropic SDK, Claude Code, and any Anthropic-compatible client.
  </Card>

  <Card title="Function Calling" href="/llm-api/function-calling" icon="wrench">
    Enable models to call functions and tools. Supports both OpenAI and Anthropic tool use formats.
  </Card>

  <Card title="Advanced Parameters" href="/llm-api/advanced-parameters" icon="sliders">
    Detailed reference for sampling, penalties, streaming, response formatting, and more.
  </Card>

  <Card title="Count Tokens" href="/llm-api/count-tokens" icon="calculator">
    Count the number of tokens in a message before sending it. Useful for managing costs and context windows.
  </Card>

  <Card title="List Models" href="/llm-api/list-models" icon="list">
    Retrieve the list of all available LLM models and their capabilities.
  </Card>

  <Card title="Coding Tools Integration" href="/llm-api/coding-tools" icon="code">
    Use ModelsLab with Claude Code, Codex, Cursor, Continue, Aider, and other AI coding assistants.
  </Card>
</CardGroup>

## SDK Compatibility

ModelsLab's LLM API is designed as a **drop-in replacement** for both OpenAI and Anthropic APIs. Just change the base URL and API key:

<Tabs>
  <Tab title="OpenAI SDK">
    ```python theme={null}
    from openai import OpenAI

    client = OpenAI(
        api_key="YOUR_MODELSLAB_API_KEY",
        base_url="https://modelslab.com/api/v7/llm",
    )

    response = client.chat.completions.create(
        model="Qwen/Qwen2.5-VL-72B-Instruct-together",
        messages=[
            {"role": "user", "content": "Hello!"}
        ],
    )

    print(response.choices[0].message.content)
    ```
  </Tab>

  <Tab title="Anthropic SDK">
    ```python theme={null}
    from anthropic import Anthropic

    client = Anthropic(
        api_key="YOUR_MODELSLAB_API_KEY",
        base_url="https://modelslab.com/api/v7/llm",
    )

    message = client.messages.create(
        model="Qwen/Qwen2.5-VL-72B-Instruct-together",
        max_tokens=1024,
        messages=[
            {"role": "user", "content": "Hello!"}
        ],
    )

    print(message.content[0].text)
    ```
  </Tab>

  <Tab title="Claude Code (CLI)">
    ```bash theme={null}
    ANTHROPIC_BASE_URL="https://modelslab.com/api/v7/llm" \
    ANTHROPIC_AUTH_TOKEN="YOUR_MODELSLAB_API_KEY" \
    claude --model "Qwen/Qwen2.5-VL-72B-Instruct-together"
    ```
  </Tab>
</Tabs>

## Supported Models

Browse all available LLM models at [modelslab.com/models/category/llmaster](https://modelslab.com/models/category/llmaster), or use the [List Models](/llm-api/list-models) endpoint to fetch them programmatically.

Popular models include:

* **Qwen 2.5 VL 72B** — Vision-language model with strong reasoning
* **DeepSeek R1** — Advanced reasoning model
* **Llama 3.1 70B/405B** — Meta's open-source flagship models
* **Mistral Large** — Mistral AI's most capable model
* And 200+ more...