Claude Haiku 4.5

Composite indexes

• Artificial Analysis Intelligence Index#81 of 151
  17independentv4.3.2, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗
  15independentv4.3.2, thinking offArtificial Analysis: Claude 4.5 Haiku (Non-reasoning) ↗

  Leader: Claude Opus 5.5 at 58

Coding

• SciCode#70 of 100
  42.2%independentAA run, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Claude Opus 5.5 at 66.9%

• Terminal-Bench 4.0#73 of 96
  0%independentAA run (Terminal-Bench 4.0), extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Claude Sonnet 5.5 at 70.6%

• Terminal-Bench 2.x#66 of 105
  44.2%independentAA run (Terminal-Bench 2.1), extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Claude Fable 5.1 at 91.4%

• LiveCodeBench#33 of 57
  61.5%independentAA run, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Fugu at 92.9%

• SWE-bench Pro#31 of 33
  39.5%independentpublic set, Scale AI leaderboard entry dated 2025-09-19Scale AI SWE-Bench Pro (public) leaderboard ↗

  Leader: Claude Opus 5.5 at 89.9%

Agents and tool use

• GDPval-AA#75 of 109
  719independentGDPval-AA v2.1, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Muse Wonder Full at 1857

• τ²-Bench#44 of 89
  54.7%independentAA run, telecom, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: GLM-5.2 at 99.1%

Reasoning and knowledge

• GPQA Diamond#105 of 144
  67.2%independentAA run, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: GPT-6 Astra at 96.1%

• Humanity’s Last Exam#109 of 153
  10.4%independentAA run, no tools, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Claude Sonnet 5.5 at 64.5%

• CritPt#98 of 144
  0%independentAA run, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Muse Wonder Full at 33%

• ARC-AGI-2#23 of 25
  4%independent32K thinking budget; ARC Prize verifiedARC Prize leaderboard data (semi-private) ↗

  Leader: GPT-6 Astra at 95%

Vision and long context

• AA Long Context Reasoning#66 of 144
  74.3%independentAA-LCR v1.1, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Kimi K3 at 88.7%

• MMMU-Pro#59 of 72
  58.6%independentAA run, extended thinkingArtificial Analysis: Claude 4.5 Haiku (Reasoning) ↗

  Leader: Claude Opus 5.5 at 87.7%

Human preference

• LMArena Text#49 of 57
  1415independentclaude-haiku-4-5-20251001; 129,278 votesLMArena text leaderboard ↗

  Leader: Gemini 4 Argon at 1525

• LMArena WebDev#56 of 58
  1329independentclaude-haiku-4-5-20251001; 27,776 votesLMArena Code Arena | WebDev leaderboard ↗

  Leader: Claude Opus 5.5 at 1818

Other published results