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

# MiniMax

> Connect WednesdayAI to MiniMax AI and the M2.5 model family.

# MiniMax

MiniMax is an AI company that builds the **M2/M2.5** model family. The current coding-focused release is **MiniMax M2.5**, designed for real-world complex tasks including multi-language coding, web/app development, and agent workflows.

## Setup options

<Tabs>
  <Tab title="OAuth (recommended)">
    ### MiniMax OAuth — Coding Plan

    Best for quick setup with a MiniMax Coding Plan — no API key required.

    **1. Enable the plugin:**

    ```bash theme={"dark"}
    openclaw plugins enable minimax-portal-auth
    openclaw gateway restart
    ```

    **2. Authenticate:**

    ```bash theme={"dark"}
    openclaw onboard --auth-choice minimax-portal
    ```

    You will be prompted to select an endpoint:

    * **Global** — International users (`api.minimax.io`)
    * **CN** — Users in China (`api.minimaxi.com`)
  </Tab>

  <Tab title="API key">
    ### MiniMax API key

    Best for hosted MiniMax with the Anthropic-compatible API.

    **1. Get an API key** from [platform.minimax.io](https://platform.minimax.io) and store it:

    ```bash theme={"dark"}
    echo 'MINIMAX_API_KEY=sk-...' >> ~/.openclaw/.env
    ```

    **2. Configure:**

    ```json5 theme={"dark"}
    {
      models: {
        providers: {
          minimax: {
            baseUrl: "https://api.minimax.io/anthropic",
            apiKey: "MINIMAX_API_KEY", // references the env var
            api: "anthropic-messages",
            models: [
              {
                id: "MiniMax-M2.5",
                name: "MiniMax M2.5",
                reasoning: true,
                input: ["text"],
                contextWindow: 200000,
                maxTokens: 8192,
              },
              {
                id: "MiniMax-M2.5-highspeed",
                name: "MiniMax M2.5 Highspeed",
                reasoning: true,
                input: ["text"],
                contextWindow: 200000,
                maxTokens: 8192,
              },
            ],
          },
        },
      },
      agents: {
        defaults: {
          model: { primary: "minimax/MiniMax-M2.5" },
        },
      },
    }
    ```

    <Note>
      `apiKey` is the name of the environment variable, not the key value itself. Keep secrets in `.env`, not in committed config files.
    </Note>

    Or use the interactive wizard: `openclaw configure` → **Model/auth** → **MiniMax M2.5**.
  </Tab>
</Tabs>

## Provider settings

| Setting      | Value                              |
| ------------ | ---------------------------------- |
| Provider ID  | `minimax`                          |
| Base URL     | `https://api.minimax.io/anthropic` |
| API          | Anthropic Messages compatible      |
| Auth env var | `MINIMAX_API_KEY`                  |

## Available models

| Model ID                 | Context | Notes                                          |
| ------------------------ | ------- | ---------------------------------------------- |
| `MiniMax-M2.5`           | 200K    | Reasoning; recommended                         |
| `MiniMax-M2.5-highspeed` | 200K    | Reasoning; faster, lower latency               |
| `MiniMax-M2.5-Lightning` | 200K    | Legacy name — prefer `highspeed` for new setup |

Use model IDs as `minimax/<model-id>` — for example, `minimax/MiniMax-M2.5`.

<Warning>
  Model IDs are **case-sensitive**. `minimax/MiniMax-M2.5` is correct; `minimax/minimax-m2.5` will not resolve.
</Warning>

## Verification

```bash theme={"dark"}
openclaw models status --probe-provider minimax
openclaw models list --provider minimax
```

## Troubleshooting

<AccordionGroup>
  <Accordion title="'Unknown model: minimax/MiniMax-M2.5'">
    The MiniMax provider is not configured. Fix by either:

    * Running `openclaw configure` and selecting **MiniMax M2.5**, or
    * Adding the `models.providers.minimax` block manually, or
    * Setting `MINIMAX_API_KEY` so the provider can be auto-injected.

    Then restart the gateway and recheck with `openclaw models list`.
  </Accordion>

  <Accordion title="401 Unauthorized">
    Verify `MINIMAX_API_KEY` is set correctly in `~/.openclaw/.env`. For OAuth users, re-run the login command if the token refresh has failed.
  </Accordion>

  <Accordion title="429 Rate limited">
    Check your Coding Plan usage via the MiniMax platform dashboard. The CN endpoint (`api.minimaxi.com`) has separate rate limits from the global endpoint.
  </Accordion>
</AccordionGroup>

## Related

* [AI providers overview](/admin/providers/index)
* [Reference: config](/reference/config)