# MiniMax-M2.5-highspeed

> MiniMax's high-speed variant for polyglot code work and precise refactoring with low latency.

## Key facts

- Model author: MiniMax
- Input: $0.60 per 1M tokens
- Output: $2.40 per 1M tokens
- Cached input: $0.03 per 1M tokens
- Context window: 204,800 tokens
- Input modalities: text
- Output modalities: text
- Capabilities: reasoning, tools
- Model IDs: `minimax/MiniMax-M2.5-highspeed`
- Model page: https://www.realrelay.ai/model/minimax/minimax-m2.5-highspeed

## Overview

MiniMax-M2.5-highspeed is the high-speed variant of MiniMax's coding model for code generation and refactoring. MiniMax says it offers the same performance as the standard version with significantly faster inference. Its documented strengths are polyglot code mastery, precision refactoring and low latency, and it supports reasoning and tool calling.

It has a 204,800-token context window. It accepts text input and returns text.

## Capabilities

Low latency is the headline feature. MiniMax-M2.5-highspeed keeps reasoning and tool calling but returns results faster, so interactive tools that wait on every response feel more responsive.

It reads and writes text, with a 204,800-token context window for source files and logs. Cached input reads cost less than regular input.

## Pricing

MiniMax-M2.5-highspeed is priced at $0.60 per 1M input tokens and $2.40 per 1M output tokens. Cached input reads cost $0.03 per 1M tokens.

MiniMax-M2.5-highspeed is available through RealRelay's OpenAI-compatible API.

## Use cases

Use MiniMax-M2.5-highspeed for interactive coding assistants where every second counts, such as chat panels inside an editor or review bots that comment on a pull request while the developer waits.

It also helps agent loops that issue many tool calls in sequence, since faster inference shortens each step of the loop.

## 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-M2.5-highspeed", "messages": [{"role": "user", "content": "Hello"}]}'
```

## FAQ

### What is MiniMax-M2.5-highspeed?

MiniMax-M2.5-highspeed is the high-speed variant of MiniMax's coding model for code generation and refactoring. MiniMax says it offers the same performance as the standard version with significantly faster inference. Its documented strengths are polyglot code mastery, precision refactoring and low latency, and it supports reasoning and tool calling.

### How much does MiniMax-M2.5-highspeed cost?

MiniMax-M2.5-highspeed is priced at $0.60 per 1M input tokens and $2.40 per 1M output tokens. Cached input reads cost $0.03 per 1M tokens.

### What is the context window of MiniMax-M2.5-highspeed?

MiniMax-M2.5-highspeed has a 204,800-token context window.

### What inputs and outputs does MiniMax-M2.5-highspeed support?

MiniMax-M2.5-highspeed accepts text input and returns text.

### How is MiniMax-M2.5-highspeed different from the standard version?

MiniMax says MiniMax-M2.5-highspeed has the same performance as the standard version but significantly faster inference and low latency.

### What is MiniMax-M2.5-highspeed used for?

Use MiniMax-M2.5-highspeed for interactive coding assistants where every second counts, such as chat panels inside an editor or review bots that comment on a pull request while the developer waits.

### How do I call MiniMax-M2.5-highspeed through the API?

Call MiniMax-M2.5-highspeed through RealRelay's OpenAI-compatible API by setting the model field to minimax/MiniMax-M2.5-highspeed.