MiniMax-M3 — benchmark results

MiniMax M3 model for reasoning and agentic tasks. Provider: MiniMax. Released 2026-06-01. Access: Open.

Unified ELO 1779 ± 13, rank #163 of 1776 rated models, from 165 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LLM Stats (OmniDocBench 1.5)91.6Score (%)100
LLM Stats (PostTrainBench)37.1Score (%)100
WorldCupBench87Quiniela Points100
AA IFBench82.86Accuracy (%)99.6
AA GPQA Diamond92.93Accuracy (%)98.7
AA Long Context Reasoning74Accuracy (%)98.3
MedScribe87.25Score (self-reported)96.2
AA Humanity's Last Exam37.12Accuracy (%)94.8
Vals AI GPQA92.68Accuracy (%)94.7
Artificial Analysis Intelligence Index44.44Intelligence Index94.5
Vals AI MedScribe87.25Accuracy (%)94.4
Vals AI CorpFin v268.1Accuracy (%)94.3

Interactive version: theaggregate.ai/model?slug=minimax-m3 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.