MiniMax M2.5 — Benchmarks

Benchmark scores for MiniMax M2.5 aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.

Overall rank: #48 of 123 open modelscomposite 52/100 across 6 benchmarks in 4 categories · methodology

Coding

BenchmarkScoreOpen rankAll models
Terminal-Bench42.7%#4 / 16#32 / 57
SWE-bench Verified75.8%#1 / 16#10 / 162
SWE-bench Multilingual68.3%#2 / 4#6 / 13

Knowledge

BenchmarkScoreOpen rankAll models
MMLU-Pro80.1%#23 / 119#65 / 259

Math

BenchmarkScoreOpen rankAll models
ProofBench4.0%#17 / 21#59 / 64

Reasoning

BenchmarkScoreOpen rankAll models
ARC-AGI63.7%#8 / 16#95 / 200

Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.