GLM-5.3

Composite indexes

• Artificial Analysis Intelligence Index#17 of 151
  45independentv4.3.2, reasoning, max effortArtificial Analysis: GLM-5.3 (max) ↗
  34independentv4.3.2, reasoning, low effortArtificial Analysis: GLM-5.3 (low) ↗

  Leader: Claude Opus 5.5 at 58

Coding

• Terminal-Bench 4.0#13 of 96
  41.9%independentreasoning, max effort, AA-runArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Claude Sonnet 5.5 at 70.6%

• Terminal-Bench 2.x#19 of 105
  88.2%vendorTerminal-Bench 2.1, Claude Code harnessGLM-5.3 model card (Hugging Face) ↗
  83.9%independentTerminal-Bench 2.1, reasoning, max effort, AA-runArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Claude Fable 5.1 at 91.4%

• SciCode#9 of 100
  59%independentreasoning, max effort, AA-runArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Claude Opus 5.5 at 66.9%

Agents and tool use

• GDPval-AA#10 of 109
  1769vendorGDPval-AA v2 (vendor table, labeled as evaluated by Artificial Analysis)GLM-5.3 model card (Hugging Face) ↗
  1646independentGDPval-AA v2.1, reasoning, max effortArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Muse Wonder Full at 1857

Reasoning and knowledge

• GPQA Diamond#27 of 144
  91.7%independentreasoning, max effort, AA-runArtificial Analysis: GLM-5.3 (max) ↗

  Leader: GPT-6 Astra at 96.1%

• Humanity’s Last Exam#32 of 153
  62.5%vendorwith tools, full setGLM-5.3 model card (Hugging Face) ↗
  42.3%independentreasoning, max effort, no tools, AA-runArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Claude Sonnet 5.5 at 64.5%

• CritPt#20 of 144
  19.1%independentreasoning, max effort, AA-runArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Muse Wonder Full at 33%

Vision and long context

• AA Long Context Reasoning#39 of 144
  79.7%independentreasoning, max effort, AA-LCRArtificial Analysis: GLM-5.3 (max) ↗

  Leader: Kimi K3 at 88.7%

Human preference

• LMArena Text#14 of 57
  1483independentarena entry "glm-5.3-max", 10,960 votes, checked 2026-09-25Arena (LMArena) text leaderboard ↗

  Leader: Gemini 4 Argon at 1525

• LMArena WebDev#16 of 58
  1622independentarena entry "glm-5.3-max", 5,801 votes, checked 2026-09-25Arena (LMArena) WebDev/Code leaderboard ↗

  Leader: Claude Opus 5.5 at 1818

Other published results