# MiniMax-M3

> MiniMax's multimodal coding model with a 1M-token context window for agentic reasoning and tool use.

## Key facts

- Model author: MiniMax
- Input: $0.60 per 1M tokens (starting tier)
- Output: $2.40 per 1M tokens (starting tier)
- Cached input: $0.12 per 1M tokens
- Context window: 1,000,000 tokens
- Input modalities: image, text, video
- Output modalities: text
- Capabilities: reasoning, tools
- Released: 2026-05-31
- Model IDs: `jd/MiniMax-M3`, `minimax/MiniMax-M3`
- Model page: https://www.realrelay.ai/model/minimax/minimax-m3

## Overview

MiniMax-M3 is a multimodal coding model from MiniMax with a 1M-token context window. MiniMax positions it for agentic reasoning, tool use, coding and long-context tasks. It accepts text, image and video input, supports reasoning and tool calling, and returns text.

It has a 1M-token context window. It accepts text, image and video input and returns text. It was released on May 31, 2026.

## Capabilities

Multimodal input is the main step beyond text-only models: MiniMax-M3 reads images and video along with text, and reasons about what it sees. It also supports tool calling, so an agent can act on what it finds.

The 1M-token context window holds very long inputs, such as a full repository, long transcripts or many documents, in a single request. Answers come back as text.

## Pricing

MiniMax-M3 is priced from $0.60 per 1M input tokens and $2.40 per 1M output tokens. Cached input reads cost $0.12 per 1M tokens. The prices shown are for the starting tier; requests with very long prompts are billed at higher rates.

MiniMax-M3 is available through RealRelay's OpenAI-compatible API.

## Use cases

Use MiniMax-M3 for coding agents that need to see: send a screenshot or screen recording of a bug along with the repository and let the model reason about the fix and call tools to apply it.

The long context also suits long-horizon work, such as analysing large document sets or following a task across many tool calls without losing earlier steps.

## Call it

OpenAI-compatible chat completions:

```bash
curl https://www.realrelay.ai/v1/chat/completions \
  -H "Authorization: Bearer $REALRELAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "minimax/MiniMax-M3", "messages": [{"role": "user", "content": "Hello"}]}'
```

## FAQ

### What is MiniMax-M3?

MiniMax-M3 is a multimodal coding model from MiniMax with a 1M-token context window. MiniMax positions it for agentic reasoning, tool use, coding and long-context tasks. It accepts text, image and video input, supports reasoning and tool calling, and returns text.

### How much does MiniMax-M3 cost?

MiniMax-M3 is priced from $0.60 per 1M input tokens and $2.40 per 1M output tokens. Cached input reads cost $0.12 per 1M tokens. The prices shown are for the starting tier; requests with very long prompts are billed at higher rates.

### What is the context window of MiniMax-M3?

MiniMax-M3 has a 1M-token context window.

### What inputs and outputs does MiniMax-M3 support?

MiniMax-M3 accepts text, image and video input and returns text.

### Can MiniMax-M3 read images and video?

Yes. MiniMax-M3 accepts text, image and video input, supports reasoning and tool calling, and returns text.

### When was MiniMax-M3 released?

MiniMax-M3 was released on May 31, 2026.

### How do I call MiniMax-M3 through the API?

Call MiniMax-M3 through RealRelay's OpenAI-compatible API by setting the model field to minimax/MiniMax-M3.