MiniMax M2

minimax

An efficiency masterpiece combining massive scale with minimal active compute. This Mixture-of-Experts model packs 230B parameters but activates only 10B per request, delivering elite performance at remarkable speed and cost. Purpose-built for coding and agentic work while maintaining strong general intelligence. With an extraordinary 205K equal input/output window and thinking capabilities, it handles massive contexts and generates extensive responses. Perfect for developers and teams who need powerful AI without the premium price tag. Exceptional value for sophisticated work.

Try Now

Capabilities

Tool Use

Extended Thinking

Example Use Cases

Coding at scale with large context

Need efficient elite coding model

Large codebase refactoring

Technical Specifications

Context Window

262,114 tokens

Max Output

262,114 tokens

Cache Miss Cost

$0.27 per 1M tokens

Non-Reasoning Cost

$1.15 per 1M tokens

Web Search Cost

$15 per 1K calls

Code Execution Cost

$0.19 per 1K calls

⚠️ Legacy

Made legacy on

Reason

Superseded by MiniMax M2.1

Recommended Replacement

MiniMax M2.1