MiniMax-M2.7 — benchmark results

MiniMax's open self-evolving M2.7 agentic MoE (230B total, 10B active). Provider: MiniMax. Released 2026-03-18. Access: Open.

Unified ELO 1646 ± 10, rank #378 of 1776 rated models, from 230 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LLM Stats (Multi-SWE-Bench)52.7Score (%)100
AA IFBench75.71Accuracy (%)93.9
AA-LCR68.7Score (self-reported)92
AA Long Context Reasoning68.67Accuracy (%)90.2
AA GPQA Diamond87.37Accuracy (%)89.2
AA SciCode46.99Accuracy (%)88.9
AA Humanity's Last Exam28.08Accuracy (%)88.4
AA Omniscience - Software Engineering (SWE) - Swift68Accuracy (%)88.4
Artificial Analysis Intelligence Index38.13Intelligence Index88.2
AA Omniscience0.68Score87.3
AA Terminal-Bench Hard39.39Accuracy (%)86.5
BenchLM64.1Overall Score83.2

Interactive version: theaggregate.ai/model?slug=minimax-m2-7 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.