# MiniMax M3

> MiniMax M3 — MSA-architecture model with a 1M context window (inputs over 512K bill at a long-context rate), native multimodal (vision) understanding, and strong coding/agentic performance. Available serverless with 50% discount pricing.

**ID**: `wafer/minimax-m3`  
**Creator**: minimax  
**Category**: text  
**Context**: 1M tokens  
**Status**: Removed (since 2026-08-17) — this listing is no longer sold; requests that still use `wafer/minimax-m3` are routed to [`minimax/m3`](https://anyrouter.dev/model/minimax/m3).  
**Reason**: Wafer smoke failed repeatedly (2026-08-14). Use minimax/m3.  
**Input**: $0.33/M  
**Output**: $1.32/M  
**Released**: 2026-03-10  
**Web page**: https://anyrouter.dev/model/wafer/minimax-m3

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

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

## 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": "wafer/minimax-m3",
  "messages": [
    {
      "role": "user",
      "content": "Say hi in 3 words."
    }
  ]
}'
```

## Providers

| Provider | Input | Output |
| --- | --- | --- |
| wafer | $0.33/M | $1.32/M |
| wafer-byok | $0 | $0 |
| commandcode-byok | $0 | $0 |

## Benchmarks

- **GPQA Diamond**: 0.9
- **HLE**: 0.4
- **τ²-Bench Telecom**: 0.9
- **SciCode**: 0.5

## Source

- Upstream docs: https://pass.wafer.ai