MiniMax-M2 — benchmark results

MiniMax's open M2 MoE (230B total, 10B active) for coding and agentic work. Provider: MiniMax. Released 2025-10-23. Access: Open.

Unified ELO 1612 ± 14, rank #451 of 1776 rated models, from 144 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA LiveCodeBench82.65Pass@1 (%)94.2
WebCoderBench - Functional Alignment93.62Score (%)92.3
Medmarks - MedConceptsQA Medium84.02Score (%)90
Medmarks - MedConceptsQA Hard76.72Score (%)89.3
AA IFBench72.31Accuracy (%)89
Medmarks - MMLU-Pro Health78.44Score (%)88.6
Medmarks - LongHealth Task 188.92Score (%)87.1
Medmarks - MedMCQA76.32Score (%)87.1
Medmarks - SCTpublic69.55Score (%)87.1
Medmarks - MedQA90.7Score (%)85.7
Medmarks - MetaMedQA76.52Score (%)85.7
WebCoderBench - Functionality Correctness72.52Score (%)84.6

Interactive version: theaggregate.ai/model?slug=minimax-m2 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.