# MiniMax M3.1 Flash

> MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it improves across benchmarks, particularly complex agentic tasks, coding and cybersecurity. It is built for multi-step co-work automation, understanding real codebases, fixing bugs and building features, and for visual understanding and creation, including frontend and presentation work. It has a 1M token context window and always-on reasoning with adjustable effort.

**ID**: `minimax/minimax-m3.1-flash`  
**Creator**: minimax  
**Category**: multimodal  
**Context**: 1M tokens  
**Input**: $0  
**Output**: $0  
**Released**: 2026-10-19  
**Web page**: https://anyrouter.dev/model/minimax/minimax-m3.1-flash

**Input modalities**: text, image, video  
**Output modalities**: text  

**Capabilities**: chat, coding, reasoning, agentic, vision, function-calling, structured-output, streaming

**Aliases**: `minimax/m3.1-flash`

## Usage

**Endpoint**: `POST https://anyrouter.dev/api/v1/chat/completions`

```bash
curl https://anyrouter.dev/api/v1/chat/completions \
  -H "Authorization: Bearer $ANYROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "minimax/minimax-m3.1-flash",
  "messages": [
    {
      "role": "user",
      "content": "Say hi in 3 words."
    }
  ]
}'
```

## Providers

| Provider | Input | Output |
| --- | --- | --- |
| hue | $0 | $0 |
| kilo | $0 | $0 |
| minimax-byok | $0 | $0 |
| openrouter-byok | $0 | $0 |
| vercel-byok | $0 | $0 |
| aihubmix-byok | $0 | $0 |
| opencode-go-byok | $0 | $0 |
| cline-byok | $0 | $0 |

## Model card

- Native context: 1000000 tokens [2]
- License: Weights not yet released [1]

### Strengths [1,3]

- Leading-model level on complex agentic tasks, computer use, coding and cybersecurity benchmarks (vendor-reported).
- Output speeds of 200+ tokens per second, with a 170+ median (vendor-reported).
- Vendor reports agentic task cost 70% below M3, at $0.15 input and $0.60 output per 1M tokens.

### Known limits [2,1]

- Always thinks; sending `thinking.type = disabled` returns HTTP 400.
- Model weights are not released yet; the vendor says they will follow later.

### Use cases [1]

- Multi-step co-work automation across SaaS systems
- Fixing bugs and building features in real codebases
- Code auditing and vulnerability patching

### API notes [2]

- Text, image and video input over the OpenAI-compatible Chat Completions API.
- reasoning_effort accepts low, medium, high, xhigh and max; default is max.
- Thinking comes back separately in `reasoning_content` and must be preserved across turns.

### Links

- announcement: https://www.minimax.io/blog/minimax-m31f
- docs: https://platform.minimax.io/docs/api-reference/text-openai-api
- pricing: https://platform.minimax.io/docs/guides/pricing-paygo
- Previous version: https://anyrouter.dev/model/minimax/m3

### Sources

1. [MiniMax M3.1 Flash: Redefining Agent Performance and Cost](https://www.minimax.io/blog/minimax-m31f) (accessed 2026-10-19)
2. [MiniMax OpenAI API compatible reference](https://platform.minimax.io/docs/api-reference/text-openai-api) (accessed 2026-10-19)
3. [MiniMax pay-as-you-go pricing](https://platform.minimax.io/docs/guides/pricing-paygo) (accessed 2026-10-19)

## Source

- Upstream docs: https://platform.minimax.io/docs/guides/text-generation