# MiniMax M3

## Best for

Long-running coding agents that work with images and video

## When not to use

Audio input. Use `minimax-m3-free` to evaluate the same weights without a rate card; this identifier is the billed lineup entry.

## Overview

MiniMax M3 is a 427B-parameter mixture-of-experts model, with 23B active parameters. It supports text, images, and video, and has a 1,048,576-token context window. Benchmark results cover coding, tool use, and multimodal tasks.

| | |
| --- | --- |
| API identifier | `minimax-m3` |
| Developer | MiniMax |
| Context window | 1,048,576 tokens |
| Supports | text, reasoning, tool-use, image input, vision, video input |
| Status | Serving |

## Rates

USD per 1M tokens, on-demand tier.

| Input | Cached input | Output |
| --- | --- | --- |
| $0.28 | $0.06 | $1.10 |

## Call it

Put `minimax-m3` in the `model` field. The same request works against the Anthropic Messages API — see [API Compatibility](https://inference.boundless.network/docs/api-compatibility).

```bash
curl https://api.inference.boundless.network/v1/chat/completions \
  -H "Authorization: Bearer $BOUNDLESS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax-m3",
    "messages": [{
      "content": "Say hello",
      "role": "user"
    }]
  }'
```

Every model we serve is listed at [Models](https://inference.boundless.network/models); the rate card in full is on [Pricing](https://inference.boundless.network/docs/pricing).