MiniMax-M2

Composite indexes

• Artificial Analysis Intelligence Index#69 of 142
  19independentv4.3.2, reasoning (AA-estimated)Artificial Analysis: MiniMax-M2 ↗

  Leader: Claude Opus 5.5 at 58

Coding

• LiveCodeBench#7 of 56
  82.6%independentreasoning, AA-runArtificial Analysis: MiniMax-M2 ↗

  Leader: Fugu at 92.9%

Agents and tool use

• τ²-Bench#23 of 88
  86.8%independenttelecom, reasoning, AA-runArtificial Analysis: MiniMax-M2 ↗

  Leader: GLM-5.2 at 99.1%

Reasoning and knowledge

• GPQA Diamond#77 of 142
  77.7%independentreasoning, AA-runArtificial Analysis: MiniMax-M2 ↗

  Leader: GPT-6 Astra at 96.1%

• Humanity’s Last Exam#83 of 147
  13.7%independentreasoning, no tools, AA-runArtificial Analysis: MiniMax-M2 ↗

  Leader: Claude Opus 5.5 at 61.4%

• CritPt#71 of 139
  0.9%independentreasoning, AA-runArtificial Analysis: MiniMax-M2 ↗

  Leader: GPT-5.6 Sol at 32.3%

Vision and long context

• AA Long Context Reasoning#80 of 139
  64.3%independentreasoning, AA-LCRArtificial Analysis: MiniMax-M2 ↗

  Leader: Kimi K3 at 88.7%

Human preference

• LMArena Text#51 of 54
  1346independentarena entry "minimax-m2", 6,753 votes, checked 2026-09-25Arena (LMArena) text leaderboard ↗

  Leader: Claude Fable 5 at 1506

• LMArena WebDev#53 of 54
  1298independentarena entry "minimax-m2", 6,544 votes, checked 2026-09-25Arena (LMArena) WebDev/Code leaderboard ↗

  Leader: Claude Opus 5.5 at 1818

Other published results