MiniMax M1 40k: benchmark results

MiniMax M1 open hybrid-attention reasoning model at the 40K thinking budget. Provider: MiniMax. Released 2025-06-01. Access: Open.

Unified ELO 1560 ± 1, rank #403 of 1392 rated models, from 22 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LLM Stats (OpenAI-MRCR: 2 needle 1M)58.6Score (%)100
LLM Stats (OpenAI-MRCR: 2 needle 128k)76.1Score (%)87.5
SpeechMap Compliance87.2% Requests Completed86.3
LLM Stats (LongBench v2)61Score (%)62.5
LLM Stats (AIME 2024)83.3Score (%)59.6
BenchmarkList ECI115.34Capability Index (ECI)55
ZeroEval MATH-50096MATH-500 Score51.6
AA Humanity's Last Exam7.83Accuracy (%)48.6
LLM Stats Score20.81LLM Stats Score (conservative rating)48.4
AA GPQA Diamond68.18Accuracy (%)45.4
Artificial Analysis Intelligence Index8.5Intelligence Index44.3
ZeroEval GPQA Diamond69.2GPQA Diamond Score41.7

Interactive version: theaggregate.ai/model?slug=minimax-m1-40k · How It Works · Data refreshed daily, snapshot 2026-09-05.