MiniMax-Text-01 — benchmark results

MiniMax's open 456B MoE (45.9B active) with hybrid Lightning Attention and up to a 4M-token inference context (January 2025). Provider: MiniMax. Released 2025-01-14. Access: Open.

Unified ELO 1536 ± 18, rank #660 of 1776 rated models, from 18 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
ProLLM - Entity Extraction85.1Score (%)91.5
LongBench v256.5Accuracy (%)79.4
ProLLM - Q&A Assistant95.3Score (%)64.7
ProLLM - StackEval92.8Score (%)58.1
SuperGPQA45.11Accuracy (%)52.9
LLM Arena RU1021Arena Elo52.7
LLM2014 Logic 2025-0138.04Score (%)46.4
ProLLM - SQL Disambiguation32.5Score (%)45.1
ProLLM - Summarization72.3Score (%)42.2
ProLLM - Function Calling85.5Score (%)37
Elimination Game (Lechmazur)3.45TrueSkill μ28.8
NYT Connections Older Models14.3Score (%)15.7

Interactive version: theaggregate.ai/model?slug=minimax-text-01 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.