Clinical Benchmarks

Anthropic

Anthropic models with results here, each shown on its own board.

Models
16
Results
114
Boards
14 of 14

www.anthropic.com

Models

ModelKindReleasedBoards
Claude Haiku 5.5Model7 Oct 20264
Claude Sonnet 5.5Model28 Sep 20267
Claude Opus 5.5Model22 Sep 20268
Claude Fable 5.1Model1 Sep 20266
Claude Opus 5Model24 Jul 202612
Claude Sonnet 5Model30 Jun 20267
Claude Fable 5Model9 Jun 20267
Claude Opus 4.8Model28 May 20266
Claude Opus 4.7Model16 Apr 20266
Claude Sonnet 4.6Model17 Feb 20265
Claude Opus 4.6Model5 Feb 20267
Claude Opus 4.5Model24 Nov 20253
Claude Haiku 4.5Model15 Oct 20256
Claude Sonnet 4.5Model29 Sep 20252
Claude Opus 4.1Model5 Aug 20252
Claude Sonnet 4Model22 May 20252

Results

One line per result. Tall tick: the board leader.

Clinical reasoning and knowledge

  1. HealthBench Professional
    Anthropic; max effort; Opus 4.8 grader; safety classifiers enabled; length-adjusted; raw 77.1%
    0.692
    Rank 2 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  2. HealthBench Professional
    length-adjusted, Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt (raw 70.3%). Measured as Claude Mythos 5; Anthropic itself prints 66.0 in the Fable 5 column of the Opus 5 card with footnote Mythos 5.
    0.660
    Rank 3 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  3. HealthBench Professional
    Anthropic; adaptive max effort; Opus 4.8 grader; five trials; no tools or custom system prompt; safety classifiers and refusal fallback to Opus 5; length-adjusted; raw 77.1%
    0.656
    Rank 4 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  4. HealthBench Professional
    Anthropic; max effort; Opus 4.8 grader; safety classifiers enabled; length-adjusted; raw 71.0%; five runs (October 1 and 2); thirteen repeat max runs on October 4 averaged 64.2%
    0.648
    Rank 6 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Oct 2026
  5. HealthBench Professional
    Measured as Claude Fable 5; Anthropic September card; length-adjusted; adaptive max effort; Opus 4.8 grader; five trials; no tools or custom system prompt; raw 68.9%
    0.633
    Rank 9 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  6. HealthBench Professional
    length-adjusted (method published in the HealthBench Professional paper); Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over five trials, no tools or customized system prompt; Fable 5.1 run with safety classifiers active and a refusal-fallback to Claude Opus 5 (raw 74.2%).
    0.621
    Rank 12 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  7. HealthBench Professional
    length-adjusted, Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt (raw 73.4%).
    0.598
    Rank 17 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jul 2026
  8. HealthBench Professional
    length-adjusted, Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt (raw 62.4%).
    0.578
    Rank 19 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  9. HealthBench Professional
    Anthropic June evaluation; length-adjusted; adaptive max effort; Opus 4.8 grader; five trials; no tools or custom system prompt
    0.574
    Rank 21 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  10. HealthBench Professional
    length-adjusted, adaptive thinking at max effort, Claude Sonnet 4.6 grader (the grader used by the Opus 4.8 card itself); the other Claude rows on this board use the Claude Opus 4.8 grader, under which Anthropic later prints 57.4 for this model.
    0.558
    Rank 23 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured May 2026
  11. HealthBench Professional
    length-adjusted; adaptive thinking at max effort; Claude Sonnet 4.6 grader; 5 trials; comparison model in the Opus 4.8 card
    0.519
    Rank 30 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured May 2026
  12. HealthBench Professional
    length-adjusted; Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt; comparison column in the Sonnet 5 card. Other Anthropic prints: 44.4% (Fable card Figure 8.18.2.A), 41.7% (Opus 4.8 card, Sonnet 4.6 grader)
    0.442
    Rank 38 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  13. HealthBench Professional
    Anthropic; extended thinking (Haiku 4.5 has no effort setting); Opus 4.8 grader; safety classifiers enabled; length-adjusted; measured for the Haiku 5.5 card
    0.322
    Rank 43 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Oct 2026
  14. MedXpertQA (MM)
    Qwen-run comparison in the Qwen3.8-Max launch post
    80.0
    Rank 4 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
  15. MedXpertQA (MM)
    Qwen-run comparison in the Qwen3.8-Max launch post
    71.7
    Rank 9 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
  16. MedXpertQA (MM)
    Meta's Muse Spark launch table (Meta reports the better of vendor self-reports and its own reproduction)
    64.8
    Rank 15 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
    Measured Apr 2026
  17. MedXpertQA (MM)
    Qwen3.5 model card comparison; source-specific evaluation, not harmonized across vendors
    63.6
    Rank 16 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
    Measured Feb 2026
  18. 57.1
    Rank 6 of 8 here, 11 models on the boardLeader GPT-5.6 Sol 60.2
    Official leaderboard
    Measured Aug 2026
  19. 56.6
    Rank 7 of 8 here, 11 models on the boardLeader GPT-5.6 Sol 60.2
    Official leaderboard
    Measured Aug 2026
  20. 94.6
    Rank 1 of 15 here, 16 models on the boardLeads this board
    Official leaderboard
    Measured Oct 2026
  21. APEX-1 GP
    Score (avg@4), reasoning max
    73.04
    Rank 1 of 12Leads this board
    Official leaderboard
    Measured Oct 2026
  22. APEX-1 GP
    Score (avg@4), reasoning max
    72.12
    Rank 2 of 12Leader Claude Opus 5.5 73.04
    Official leaderboard
    Measured Oct 2026

Documentation and coding

  1. MedScribe (Vals AI)
    model ID anthropic/claude-opus-5-5; temperature=1; max_output_tokens=128000; compute_effort=max
    91.43
    Rank 1 of 106 here, 105 models on the boardLeads this board
    Official leaderboard
    Measured Sep 2026
  2. 91.29
    Rank 2 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  3. MedScribe (Vals AI)
    model ID anthropic/claude-sonnet-5-5; temperature=1; max_output_tokens=128000; compute_effort=max
    91.10
    Rank 3 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  4. 90.98
    Rank 4 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  5. 88.52
    Rank 10 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  6. 86.74
    Rank 19 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  7. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-6-thinking; temperature=1; max_output_tokens=30000; compute_effort=max
    86.13
    Rank 22 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  8. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-8; temperature=1; max_output_tokens=30000; compute_effort=max
    85.75
    Rank 24 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  9. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-5-20251101-thinking; temperature=1; max_output_tokens=30000; compute_effort=high
    85.32
    Rank 27 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  10. MedScribe (Vals AI)
    model ID anthropic/claude-haiku-4-5-20251001-thinking; temperature=1; max_output_tokens=30000
    85.23
    Rank 29 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  11. MedScribe (Vals AI)
    model ID anthropic/claude-sonnet-4-5-20250929; temperature=1; max_output_tokens=30000
    84.52
    Rank 32 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  12. MedScribe (Vals AI)
    model ID anthropic/claude-sonnet-4-5-20250929-thinking; temperature=1; max_output_tokens=30000
    84.10
    Rank 37 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  13. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-5-20251101; temperature=1; max_output_tokens=30000; compute_effort=high
    83.25
    Rank 45 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  14. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-7; temperature=1; max_output_tokens=30000; compute_effort=max
    82.95
    Rank 47 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  15. MedScribe (Vals AI)
    model ID anthropic/claude-sonnet-5; temperature=1; max_output_tokens=30000; compute_effort=max
    76.05
    Rank 75 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  16. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-1-20250805-thinking; temperature=1; max_output_tokens=30000
    73.90
    Rank 80 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  17. MedScribe (Vals AI)
    model ID anthropic/claude-sonnet-4-20250514; temperature=1; max_output_tokens=30000
    72.41
    Rank 85 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  18. MedScribe (Vals AI)
    model ID anthropic/claude-opus-4-1-20250805; temperature=1; max_output_tokens=30000
    71.75
    Rank 89 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  19. MedScribe (Vals AI)
    model ID anthropic/claude-sonnet-4-20250514-thinking; max_output_tokens=30000
    69.35
    Rank 93 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  20. 63.57
    Rank 1 of 104 here, 103 models on the boardLeads this board
    Official leaderboard
    Measured Sep 2026
  21. 56.07
    Rank 4 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  22. 54.86
    Rank 7 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  23. 53.51
    Rank 8 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  24. 53.22
    Rank 10 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  25. MedCode (Vals AI)
    model ID anthropic/claude-sonnet-5-5; compute_effort=max; temperature=1; max_output_tokens=128000
    52.92
    Rank 12 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  26. MedCode (Vals AI)
    model ID anthropic/claude-opus-5-5; compute_effort=max; temperature=1; max_output_tokens=128000
    49.80
    Rank 17 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  27. MedCode (Vals AI)
    model ID anthropic/claude-opus-4-5-20251101-thinking; compute_effort=high; temperature=1; max_output_tokens=30000
    49.16
    Rank 22 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  28. MedCode (Vals AI)
    model ID anthropic/claude-opus-4-6-thinking; compute_effort=max; temperature=1; max_output_tokens=30000
    49.13
    Rank 23 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  29. MedCode (Vals AI)
    model ID anthropic/claude-opus-4-6; compute_effort=max; temperature=1; max_output_tokens=30000
    48.24
    Rank 28 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  30. MedCode (Vals AI)
    model ID anthropic/claude-sonnet-5; compute_effort=max; temperature=1; max_output_tokens=30000
    47.54
    Rank 31 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  31. MedCode (Vals AI)
    model ID anthropic/claude-opus-4-1-20250805-thinking; temperature=1; max_output_tokens=30000
    47.23
    Rank 33 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  32. MedCode (Vals AI)
    model ID anthropic/claude-opus-4-5-20251101; compute_effort=high; temperature=1; max_output_tokens=30000
    45.17
    Rank 36 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  33. MedCode (Vals AI)
    model ID anthropic/claude-sonnet-4-5-20250929-thinking; temperature=1; max_output_tokens=30000
    44.13
    Rank 40 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  34. MedCode (Vals AI)
    model ID anthropic/claude-opus-4-1-20250805; temperature=1; max_output_tokens=30000
    41.37
    Rank 52 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  35. MedCode (Vals AI)
    model ID anthropic/claude-sonnet-4-5-20250929; temperature=1; max_output_tokens=30000
    40.57
    Rank 60 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  36. MedCode (Vals AI)
    model ID anthropic/claude-sonnet-4-20250514-thinking; max_output_tokens=30000
    34.96
    Rank 76 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  37. MedCode (Vals AI)
    model ID anthropic/claude-sonnet-4-20250514; temperature=1; max_output_tokens=30000
    33.94
    Rank 80 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  38. MedCode (Vals AI)
    model ID anthropic/claude-haiku-4-5-20251001-thinking; temperature=1; max_output_tokens=30000
    32.68
    Rank 85 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  39. MLCR-AA
    overall pass rate, Max, Default Fallback
    41.7
    Rank 1 of 27Leads this board
    Independent run
    Measured Nov 2026
  40. MLCR-AA
    overall pass rate, Max
    35.0
    Rank 5 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  41. MLCR-AA
    overall pass rate
    27.8
    Rank 9 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  42. MLCR-AA
    overall pass rate, Max, Default Fallback
    15.6
    Rank 20 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026

EHR and workflow agents

  1. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; safety classifiers enabled
    68.4
    Rank 1 of 28 here, 12 models on the boardLeads this board
    Vendor-reported
    Measured Sep 2026
  2. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; safety classifiers enabled
    63.2
    Rank 2 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  3. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; safety classifiers enabled
    61.0
    Rank 3 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  4. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; safety classifiers enabled
    57.6
    Rank 4 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  5. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; xhigh effort; Opus 5 rubric grader; safety classifiers enabled
    56.4
    Rank 5 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  6. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; high effort; Opus 5 rubric grader; safety classifiers enabled
    47.6
    Rank 6 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  7. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; safety classifiers enabled; five attempts per task; run October 1 and 2
    43.0
    Rank 8 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Oct 2026
  8. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; safety classifiers enabled
    37.4
    Rank 9 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  9. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; xhigh effort; Opus 5 rubric grader; safety classifiers enabled; five attempts per task; run October 5
    35.8
    Rank 10 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Oct 2026
  10. 31.7
    Rank 12 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Official leaderboard
    Measured May 2026
  11. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; high effort; Opus 5 rubric grader; safety classifiers enabled; five attempts per task; run October 5
    31.6
    Rank 13 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Oct 2026
  12. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; medium effort; Opus 5 rubric grader; safety classifiers enabled
    30.0
    Rank 14 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  13. 29.3
    Rank 15 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Official leaderboard
    Measured May 2026
  14. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; low effort; Opus 5 rubric grader; safety classifiers enabled
    27.2
    Rank 17 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Sep 2026
  15. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; medium effort (API default); Opus 5 rubric grader; safety classifiers enabled; five attempts per task; run October 5
    25.2
    Rank 18 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Oct 2026
  16. 23.0
    Rank 19 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Official leaderboard
    Measured May 2026
  17. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; low effort; Opus 5 rubric grader; safety classifiers enabled; five attempts per task; run October 5
    17.8
    Rank 21 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Oct 2026
  18. PhysicianBench
    Anthropic harness; 100 tasks; pass@1; extended thinking (no effort setting); Opus 5 rubric grader; safety classifiers enabled; measured for the Haiku 5.5 card
    1.4
    Rank 28 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Vendor-reported
    Measured Oct 2026
  19. HealthAgentBench
    $3.3/task; harness+model evaluated jointly
    55
    Rank 1 of 12Leads this board
    Official leaderboard
    Measured Jul 2026
  20. 36
    Rank 4 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  21. 32
    Rank 6 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  22. 27
    Rank 8 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  23. 19
    Rank 10 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  24. 17
    Rank 11 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  25. CHI-Bench
    community-submitted harness config validated by automated workspace judge
    54.7
    Rank 1 of 43 here, 45 models on the boardLeads this board
    Independent run
    Measured Aug 2026
  26. 37.3
    Rank 2 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  27. 37.3
    Rank 2 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Independent run
    Measured Aug 2026
  28. 33.3
    Rank 4 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  29. 28.0
    Rank 5 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  30. 26.2
    Rank 6 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Apr 2026
  31. 24.4
    Rank 9 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Apr 2026
  32. 24.0
    Rank 10 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  33. 20.0
    Rank 13 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  34. CHI-Bench
    All Domains pass@1; openclaw harness
    17.3
    Rank 17 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  35. CHI-Bench
    All Domains pass@1; claude-code harness
    6.2
    Rank 36 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Apr 2026
  36. RCM-Bench
    Score (avg@3), provider-default reasoning
    92.68
    Rank 1 of 20Leads this board
    Official leaderboard
    Measured Oct 2026
  37. RCM-Bench
    Score (avg@3), provider-default reasoning
    91.24
    Rank 3 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  38. RCM-Bench
    Score (avg@3), provider-default reasoning
    91.05
    Rank 5 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  39. RCM-Bench
    Score (avg@3), provider-default reasoning
    90.14
    Rank 7 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  40. RCM-Bench
    Score (avg@3), provider-default reasoning
    89.82
    Rank 8 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  41. RCM-Bench
    Score (avg@3), provider-default reasoning
    88.38
    Rank 10 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  42. RCM-Bench
    Score (avg@3), provider-default reasoning
    87.85
    Rank 13 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  43. RCM-Bench
    Score (avg@3), provider-default reasoning
    70.23
    Rank 20 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  44. DAYJOB Health
    Score (pass@1 at 95%), reasoning adaptive/max
    48.3
    Rank 1 of 15Leads this board
    Official leaderboard
    Measured Oct 2026
  45. DAYJOB Health
    Score (pass@1 at 95%), reasoning adaptive/max
    34
    Rank 2 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026
  46. DAYJOB Health
    Score (pass@1 at 95%), reasoning adaptive/max
    31.2
    Rank 3 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026
  47. DAYJOB Health
    Score (pass@1 at 95%), reasoning adaptive/max
    22
    Rank 9 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026
  48. DAYJOB Health
    Score (pass@1 at 95%), reasoning adaptive/max
    11.2
    Rank 12 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026

Safety

  1. First, Do NOHARM (v2)
    v2 run on ARISE; 19 models on the board
    74.6
    Rank 6 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026
  2. 65.0
    Rank 11 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026