Shared benchmarks

28
Shared benchmarks for Claude Sonnet 5 and Claude Sonnet 5.5
BenchmarkClaude Sonnet 5Claude Sonnet 5.5
  • 54.5%
  • 44.6%medium
  • 75.5%max
Evaluation details for APEX-Agents Original

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on APEX-Agents Original

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on APEX-Agents Original
Record for Claude Sonnet 5.5 on APEX-Agents Original
  • 3.7%low
  • 5.8%medium
  • 9.9%high
  • 7.6%xhigh
  • 10.7%max
  • 44.7%
Evaluation details for AutomationBench 1.0.6

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on AutomationBench 1.0.6
Record for Claude Sonnet 5 on AutomationBench 1.0.6
Record for Claude Sonnet 5 on AutomationBench 1.0.6
Record for Claude Sonnet 5 on AutomationBench 1.0.6
Record for Claude Sonnet 5 on AutomationBench 1.0.6

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on AutomationBench 1.0.6
  • 68.7%
  • 86.6%
Evaluation details for BrowseComp 130-question subset — Mercor web research agent

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 88.8%
  • 96.1%
Evaluation details for CharXiv Descriptive and reasoning — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 24.1%low
  • 28.0%medium
  • 30.8%high
  • 32.0%xhigh
  • 34.1%max
  • 35.8%low
  • 39.2%medium
  • 47.8%high
  • 53.1%xhigh
  • 55.5%max
Evaluation details for CursorBench 4.0

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on CursorBench 4.0
Record for Claude Sonnet 5 on CursorBench 4.0
Record for Claude Sonnet 5 on CursorBench 4.0
Record for Claude Sonnet 5 on CursorBench 4.0
Record for Claude Sonnet 5 on CursorBench 4.0

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on CursorBench 4.0
Record for Claude Sonnet 5.5 on CursorBench 4.0
Record for Claude Sonnet 5.5 on CursorBench 4.0
Record for Claude Sonnet 5.5 on CursorBench 4.0
Record for Claude Sonnet 5.5 on CursorBench 4.0
DeepSWE 1.1Potentially non-equivalent: Evaluator sets differ.
  • 54.0%1.1 · DataCurve
  • 71.0%1.1 · Anthropic
Potentially non-equivalent · Evaluation details for DeepSWE 1.1

Evaluator sets differ. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on DeepSWE 1.1

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on DeepSWE 1.1
  • 49.6%
  • 70.5%
Evaluation details for DeepSWE 1.1 — Mercor mini-swe-agent, 500 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Claude Sonnet 5.5

  • 53.9%
  • 58.1%
Evaluation details for Finance Agent Benchmark 2 — October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 89.0%
  • 92.1%
Evaluation details for Global MMLU Original

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on Global MMLU Original

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Global MMLU Original
  • 90.5%xhigh
  • 80.3%max
  • 95.6%
Evaluation details for GPQA Diamond — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Claude Sonnet 5.5

  • 80.6%
  • 95.2%
Evaluation details for GPQA Diamond — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Claude Sonnet 5.5

  • 5.0%
  • 2.9%
Evaluation details for HLAB October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on Harvey's Legal Agent Benchmark October 2026 evaluation

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Harvey's Legal Agent Benchmark October 2026 evaluation
  • 43.2%
  • 41.4%low
  • 43.2%medium
  • 47.9%high
  • 53.0%xhigh
  • 56.9%max
Evaluation details for Humanity's Last Exam Full set — No tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Humanity's Last Exam Full set — No tools
  • 36.5%low
  • 47.2%medium
  • 52.8%high
  • 54.6%xhigh
  • 47.6%low
  • 50.5%medium
  • 56.7%high
  • 62.0%xhigh
  • 64.5%max
Evaluation details for Humanity's Last Exam Full set — With tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5.5

  • 63.0%
  • 80.7%
Evaluation details for MedXpertQA MM — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on MedXpertQA MM — Mercor single-shot

Claude Sonnet 5.5

  • 89.3%
  • 91.6%
Evaluation details for MILU Original

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on Multi-task Indic Language Understanding Benchmark Original

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Multi-task Indic Language Understanding Benchmark Original
  • 88.5%
  • 91.0%
Evaluation details for MMLU-Pro Original — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Claude Sonnet 5.5

  • 75.5%
  • 85.8%
Evaluation details for MMMU-Pro 1,730 questions — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on MMMU-Pro 1,730 questions — Mercor single-shot

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on MMMU-Pro 1,730 questions — Mercor single-shot
  • 32.9%xhigh
  • 33.7%max
  • 46.5%
Evaluation details for SimpleQA Verified Original — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 39.5%
  • 50.3%
Evaluation details for SWE Atlas Codebase QnA — Mercor mini-swe-agent, 250 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 78.3%
  • 90.3%
Evaluation details for SWE-bench Multilingual

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on SWE-bench Multilingual

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on SWE-bench Multilingual
  • 89.2%
  • 90.9%
Evaluation details for SWE-bench Multilingual — Mercor mini-swe-agent, 1000 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 82.8%
  • 96.7%
Evaluation details for SWE-bench Verified — Mercor mini-swe-agent, 1000 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 63.2%
  • 81.3%
Evaluation details for SWE-bench Pro Public

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on SWE-bench Pro Public

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on SWE-bench Pro Public
  • 74.5%
  • 83.1%
Evaluation details for Terminal-Bench 2.1 — Vals AI Terminus 2

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on Terminal-Bench 2.1 — Vals AI Terminus 2

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Terminal-Bench 2.1 — Vals AI Terminus 2
  • 12.4%
  • 61.8%
Evaluation details for Terminal-Bench 4.0 — Claude Code

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on Terminal-Bench 4.0 — Claude Code

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Terminal-Bench 4.0 — Claude Code
Toolathlon VerifiedPotentially non-equivalent: Evaluator sets differ.
  • 71.6%Verified · Toolathlon Authors
  • 77.8%Verified · Anthropic
Potentially non-equivalent · Evaluation details for Toolathlon Verified

Evaluator sets differ. Evaluation methodology is not recorded by the registry.

Claude Sonnet 5

Record for Claude Sonnet 5 on Toolathlon Verified

Claude Sonnet 5.5

Record for Claude Sonnet 5.5 on Toolathlon Verified
  • 81.3%
  • 92.4%
Evaluation details for Vibe Code Bench 1.1 — OpenHands, October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.