Clinical Benchmarks

OpenAI

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

Models
21
Results
97
Boards
14 of 14

openai.com

Models

ModelKindReleasedBoards
GPT-6.1 TerraModel26 Oct 20262
GPT-6.1 SolModel29 Sep 20267
GPT-6 LunaModel22 Sep 20267
GPT-6 SolModel22 Sep 20267
GPT-6 AstraModel3 Sep 20267
GPT-5.6 LunaModel9 Jul 20265
GPT-5.6 SolModel9 Jul 202611
GPT-5.6 TerraModel9 Jul 20265
GPT-5.5 InstantModel5 May 20262
GPT-5.5Model23 Apr 20267
GPT-5.4 miniModel17 Mar 20262
GPT-5.4 nanoModel17 Mar 20262
GPT-5.4Model5 Mar 20267
GPT-5.3-CodexModel5 Feb 20261
GPT-5.2Model11 Dec 20254
GPT-5.1Model12 Nov 20253
GPT-5Model7 Aug 20254
GPT-5 miniModel7 Aug 20253
GPT-5 nanoModel7 Aug 20252
o3Model16 Apr 20252
o4-miniModel16 Apr 20252

Results

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

Clinical reasoning and knowledge

  1. HealthBench Professional
    Anthropic public-API reproduction; max effort; no system prompt; Opus 4.8 grader; length-adjusted; raw 74.0%
    0.703
    Rank 1 of 43Leads this board
    Independent run
    Measured Sep 2026
  2. HealthBench Professional
    Run by Anthropic through OpenAI's public API; max effort; Opus 4.8 grader; length-adjusted; not OpenAI's published figure
    0.655
    Rank 5 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Independent run
    Measured Oct 2026
  3. HealthBench Professional
    length-adjusted, max reasoning effort (68.2 unadjusted, 3,185 mean response chars, as corrected by OpenAI on September 22; first printed as 69.5 and 4,097); GPT-6 Astra system card Table 6, column 'gpt-6 Astra'.
    0.647
    Rank 7 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  4. HealthBench Professional
    OpenAI; length-adjusted; 64.2 (67.2, 3038) (adjusted, raw, mean response characters); addendum Table 7 refers to the GPT-6 Astra card for method
    0.642
    Rank 8 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  5. HealthBench Professional
    OpenAI; length-adjusted; 62.7 (64.2, 2531) (adjusted, raw, mean response characters); addendum Table 5 refers to the GPT-6 Astra card for method
    0.627
    Rank 10 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Oct 2026
  6. HealthBench Professional
    OpenAI; length-adjusted; maximum reasoning effort; 60.8 (61.2, 2119) (adjusted, raw, mean response characters)
    0.608
    Rank 13 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  7. HealthBench Professional
    OpenAI; length-adjusted; maximum reasoning effort; 60.8 (59.5, 1573) (adjusted, raw, mean response characters)
    0.608
    Rank 13 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Sep 2026
  8. HealthBench Professional
    length-adjusted, max reasoning effort (64.1 unadjusted, 3,228 mean response chars); GPT-5.6 system card Table 6, column GPT-5.6-SOL. API max-effort setting, not the August ChatGPT production setting measured in row 492.
    0.605
    Rank 15 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  9. HealthBench Professional
    length-adjusted, max reasoning effort (62.4 unadjusted, 3,618 mean response chars); GPT-5.6 system card Table 6, column GPT-5.6-TERRA.
    0.577
    Rank 20 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  10. HealthBench Professional
    length-adjusted, max reasoning effort (59.8 unadjusted, 3,389 mean response chars); GPT-5.6 system card Table 6, column GPT-5.6-LUNA. API max-effort setting, not the August ChatGPT production setting measured in row 493.
    0.557
    Rank 24 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  11. HealthBench Professional
    ChatGPT production (Instant) deployment setting, length-adjusted, GPT-5.6 August Updates PDF column GPT-5.6 Sol (August) (56.6 unadjusted, 2,894 chars)
    0.540
    Rank 27 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Aug 2026
  12. HealthBench Professional
    length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.5 (57.2 unadjusted, 3818 chars)
    0.518
    Rank 31 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  13. HealthBench Professional
    length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.4 (51.9 unadjusted, 3308 chars)
    0.481
    Rank 35 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  14. HealthBench Professional
    length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5 (51.0 unadjusted, 3616 chars)
    0.462
    Rank 36 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  15. HealthBench Professional
    length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.2 (50.0 unadjusted, 3400 chars)
    0.459
    Rank 37 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  16. HealthBench Professional
    ChatGPT production (Instant) deployment setting, length-adjusted, GPT-5.6 August Updates PDF column GPT-5.6 Luna (August) (46.8 unadjusted, 2,920 chars)
    0.441
    Rank 39 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Aug 2026
  17. HealthBench Professional
    length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.1 (48.0 unadjusted, 4863 chars)
    0.396
    Rank 40 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured Jun 2026
  18. HealthBench Professional
    length-adjusted (40.7 unadjusted, 2,775 mean response chars); GPT-5.5 Instant system card Table 5, column GPT-5.5 INSTANT.
    0.384
    Rank 41 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Vendor-reported
    Measured May 2026
  19. MedXpertQA (MM)
    Qwen-run comparison in the Qwen3.8-Max launch post
    81.5
    Rank 1 of 22 here, 5 models on the boardLeads this board
    Independent run
  20. MedXpertQA (MM)
    Meta's Muse Spark launch table (Meta reports the better of vendor self-reports and its own reproduction)
    77.1
    Rank 6 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
    Measured Apr 2026
  21. MedXpertQA (MM)
    Qwen-run comparison in the Qwen3.5-397B-A17B model card
    73.3
    Rank 8 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
  22. MAST (Medical AI Superintelligence Test)
    MAST in preview; 'exact scores may change'
    60.2
    Rank 1 of 8 here, 11 models on the boardLeads this board
    Official leaderboard
    Measured Aug 2026
  23. 93.4
    Rank 3 of 15 here, 16 models on the boardLeader Claude Opus 5 94.6
    Official leaderboard
    Measured Oct 2026
  24. 91.1
    Rank 5 of 15 here, 16 models on the boardLeader Claude Opus 5 94.6
    Official leaderboard
    Measured Oct 2026
  25. 88.1
    Rank 8 of 15 here, 16 models on the boardLeader Claude Opus 5 94.6
    Official leaderboard
    Measured Oct 2026
  26. 87.2
    Rank 9 of 15 here, 16 models on the boardLeader Claude Opus 5 94.6
    Official leaderboard
    Measured Oct 2026
  27. APEX-1 GP
    Score (avg@4), reasoning max
    63.75
    Rank 5 of 12Leader Claude Opus 5.5 73.04
    Official leaderboard
    Measured Oct 2026
  28. APEX-1 GP
    Score (avg@4), reasoning max
    61.03
    Rank 7 of 12Leader Claude Opus 5.5 73.04
    Official leaderboard
    Measured Oct 2026
  29. APEX-1 GP
    Score (avg@4), reasoning max
    60.61
    Rank 8 of 12Leader Claude Opus 5.5 73.04
    Official leaderboard
    Measured Oct 2026

Documentation and coding

  1. 88.09
    Rank 12 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  2. 87.91
    Rank 14 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  3. 86.87
    Rank 18 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  4. MedScribe (Vals AI)
    model ID openai/gpt-6.1-sol; max_output_tokens=128000; reasoning_effort=max
    86.45
    Rank 21 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  5. MedScribe (Vals AI)
    model ID openai/gpt-5.6-sol; max_output_tokens=30000; reasoning_effort=max
    85.23
    Rank 29 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  6. MedScribe (Vals AI)
    model ID openai/gpt-5.6-luna; max_output_tokens=30000; reasoning_effort=max
    84.39
    Rank 34 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 openai/gpt-5.2-2025-12-11; max_output_tokens=30000; reasoning_effort=xhigh
    84.39
    Rank 34 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 openai/gpt-6-luna; max_output_tokens=128000; reasoning_effort=max
    83.71
    Rank 41 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 openai/gpt-5-2025-08-07; max_output_tokens=30000; reasoning_effort=high
    83.65
    Rank 42 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 openai/gpt-5.6-terra; max_output_tokens=30000; reasoning_effort=xhigh
    82.87
    Rank 48 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 openai/gpt-6-sol; max_output_tokens=128000; reasoning_effort=max
    82.03
    Rank 50 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 openai/gpt-5-mini-2025-08-07; max_output_tokens=30000; reasoning_effort=high
    80.58
    Rank 54 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 openai/gpt-5.4-2026-03-05; max_output_tokens=30000; reasoning_effort=xhigh
    77.55
    Rank 66 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 openai/gpt-5.4-nano-2026-03-17; max_output_tokens=30000; reasoning_effort=high
    77.09
    Rank 69 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 openai/o3-2025-04-16; max_output_tokens=30000; reasoning_effort=high
    76.65
    Rank 71 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 openai/gpt-5-nano-2025-08-07; max_output_tokens=30000; reasoning_effort=high
    72.86
    Rank 82 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 openai/o4-mini-2025-04-16; max_output_tokens=30000; reasoning_effort=high
    69.14
    Rank 94 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  18. 52.73
    Rank 13 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  19. MedCode (Vals AI)
    model ID openai/gpt-5.2-2025-12-11; reasoning_effort=xhigh; max_output_tokens=30000
    49.75
    Rank 18 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  20. MedCode (Vals AI)
    model ID openai/gpt-5-2025-08-07; reasoning_effort=high; max_output_tokens=30000
    49.63
    Rank 19 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  21. MedCode (Vals AI)
    model ID openai/gpt-5.5; reasoning_effort=xhigh; temperature=1; max_output_tokens=30000
    49.10
    Rank 24 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  22. MedCode (Vals AI)
    model ID openai/gpt-6.1-sol; reasoning_effort=max; max_output_tokens=128000
    48.84
    Rank 26 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  23. 48.49
    Rank 27 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  24. MedCode (Vals AI)
    model ID openai/o3-2025-04-16; reasoning_effort=high; max_output_tokens=30000
    47.29
    Rank 32 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  25. MedCode (Vals AI)
    model ID openai/gpt-6-sol; reasoning_effort=max; max_output_tokens=128000
    47.07
    Rank 34 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  26. MedCode (Vals AI)
    model ID openai/gpt-6-luna; reasoning_effort=max; max_output_tokens=128000
    44.69
    Rank 39 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  27. MedCode (Vals AI)
    model ID openai/gpt-5.6-sol; reasoning_effort=max; max_output_tokens=30000
    43.97
    Rank 41 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  28. MedCode (Vals AI)
    model ID openai/gpt-5.6-terra; reasoning_effort=xhigh; max_output_tokens=30000
    43.41
    Rank 43 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  29. MedCode (Vals AI)
    model ID openai/gpt-5-mini-2025-08-07; reasoning_effort=high; max_output_tokens=30000
    43.05
    Rank 46 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  30. MedCode (Vals AI)
    model ID openai/gpt-5.6-luna; reasoning_effort=max; max_output_tokens=30000
    42.39
    Rank 49 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  31. MedCode (Vals AI)
    model ID openai/gpt-5.4-2026-03-05; reasoning_effort=xhigh; max_output_tokens=30000
    41.29
    Rank 53 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  32. MedCode (Vals AI)
    model ID openai/gpt-5.4-nano-2026-03-17; reasoning_effort=high; max_output_tokens=30000
    41.03
    Rank 57 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  33. MedCode (Vals AI)
    model ID openai/o4-mini-2025-04-16; reasoning_effort=high; max_output_tokens=30000
    33.79
    Rank 81 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  34. MedCode (Vals AI)
    model ID openai/gpt-5-nano-2025-08-07; reasoning_effort=high; max_output_tokens=30000
    30.44
    Rank 93 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  35. MLCR-AA
    overall pass rate, Max
    41.7
    Rank 1 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  36. MLCR-AA
    overall pass rate, Max
    38.3
    Rank 3 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  37. MLCR-AA
    overall pass rate, Max
    33.9
    Rank 6 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  38. MLCR-AA
    overall pass rate, June 2026
    31.7
    Rank 7 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  39. MLCR-AA
    overall pass rate, Max
    30.6
    Rank 8 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026
  40. MLCR-AA
    overall pass rate, Max
    26.1
    Rank 11 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026

EHR and workflow agents

  1. PhysicianBench
    pass@1; Pass^3 28.0
    46.3
    Rank 7 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Official leaderboard
    Measured May 2026
  2. PhysicianBench
    Run by Anthropic through OpenAI's public API; Anthropic harness; 100 tasks; pass@1; max effort; Opus 5 rubric grader; not OpenAI's published figure
    33.6
    Rank 11 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Independent run
    Measured Oct 2026
  3. 27.7
    Rank 16 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Official leaderboard
    Measured May 2026
  4. 45
    Rank 2 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  5. 42
    Rank 3 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  6. 35
    Rank 5 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  7. 28
    Rank 7 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  8. HealthAgentBench
    $1.0/task; Codex harness; mean success across 3 attempts × 54 tasks
    22
    Rank 9 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  9. 16
    Rank 12 of 12Leader Claude Code (Opus 5) 55
    Official leaderboard
    Measured Jul 2026
  10. 25.3
    Rank 7 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  11. 20.9
    Rank 12 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Apr 2026
  12. CHI-Bench
    All Domains pass@1; codex harness
    16.0
    Rank 20 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Apr 2026
  13. CHI-Bench
    All Domains pass@1; codex harness
    13.3
    Rank 26 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  14. CHI-Bench
    All Domains pass@1; codex harness
    13.3
    Rank 26 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  15. CHI-Bench
    All Domains pass@1; codex harness
    8.4
    Rank 34 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  16. RCM-Bench
    Score (avg@3), medium reasoning
    91.21
    Rank 4 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  17. RCM-Bench
    Score (avg@3), medium reasoning
    87.37
    Rank 14 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  18. RCM-Bench
    Score (avg@3), medium reasoning
    86.96
    Rank 15 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  19. RCM-Bench
    Score (avg@3), medium reasoning
    85.81
    Rank 16 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  20. RCM-Bench
    Score (avg@3), provider-default reasoning
    84.28
    Rank 17 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  21. RCM-Bench
    Score (avg@3), medium reasoning
    81.53
    Rank 18 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  22. RCM-Bench
    Score (avg@3), reasoning disabled for tool compatibility
    71.48
    Rank 19 of 20Leader Claude Opus 5 92.68
    Official leaderboard
    Measured Oct 2026
  23. DAYJOB Health
    Score (pass@1 at 95%), reasoning max
    28.4
    Rank 5 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026
  24. DAYJOB Health
    Score (pass@1 at 95%), reasoning max
    25.2
    Rank 7 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026
  25. DAYJOB Health
    Score (pass@1 at 95%), reasoning max
    7.2
    Rank 14 of 15Leader Claude Opus 5.5 48.3
    Official leaderboard
    Measured Oct 2026

Safety

  1. 70.1
    Rank 8 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026
  2. 70.0
    Rank 9 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026
  3. First, Do NOHARM (v2)
    from the Model Leaderboard SAFETY column (NOHARM v2 F1 weighted, shown with CI); not in the Latest Flagships ranking
    68.6
    Rank 10 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026