MiniMaxactiveOpen Source

MiniMax M2.5

minimax-m2.5

230B MoE (10B active). 80.2% SWE-bench Verified, open weights. Built for real-world productivity.

Context Window

196.6K

tokens

Max Output

8.2K

tokens

Input Price

$0.15

per 1M tokens

Output Price

$1.20

per 1M tokens

Details

Familyminimax-m2

Parameters230B (10B active)

Training Cutoff2025-12-01

ReleasedFebruary 12, 2026

Capabilities

VisionFunctionsStreamingJSON ModeCodeReasoningTool UseMultimodalWeb Search

Documentation

https://api.minimax.chat/v1/text/chatcompletion_v2

Evaluation Scores(4 benchmarks)

HumanEvalFunction-level Python code generation

91%

SWE-bench VerifiedResolving real GitHub issues

80.2%

MMLU-ProHarder successor to MMLU

78%

GPQA DiamondPhD-level science questions

72%

Quick Access

curl pikaainews.com/api/models/minimax-m2-5

npx pika-models info minimax-m2-5

Get API Access

Third-Party Providers & Aggregators

SiliconFlow

China-optimized inference. Strong Qwen/DeepSeek support.

Other minimax-m2 models

MiniMax

MiniMax M2.5 Lightning

minimax-m2.5-lightning

196.6K ctx$0.30/1M

MiniMax

MiniMax M2.1

minimax-m2.1

196.6K ctx$0.15/1M