MiniMax-M2

MiniMax🇨🇳 China
active

230B reasoning model with 1M context, MIT license. Top-ranked open-weight model on Artificial Analysis.

Context window1000K tokens
Input / 1M tokens$0.3
Output / 1M tokens$1.2

Version History

minimax-m2-launchmajor

MiniMax-M2 launches.

Benchmark Scores

Full leaderboard →
78.0%
AIME 2025
1346.0 elo
Arena Elo
77.7%
GPQA
82.6%
LiveCodeBench
82.0%
MMLU-Pro
93.0 tokens_per_sec
Speed (tok/s)
69.4%
SWE-bench Verified