RLTasksby ulam.ai
Models / Anthropic

Anthropic / MODEL EVIDENCE

Claude Opus 4.6

85 benchmarks · 150 observations · 76 sources · released 2026-02-05

Scores at a glance

85 benchmarks · large dot: Claude Opus 4.6; small dots: closest models where reported

Terminal-Bench 2.0 ↗65.4
50.868.5

Claude Opus 4.6 (max) · Terminal-Bench 2.0(Terminus-2) ↗

12 more reported results

79.8% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

76.4% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

75.28% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

74.72% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

72.08% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

71.91% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

69.89% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

66.87% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

66.52% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

62.92% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

57.98% · Claude Opus 4.6 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

69.9% · Claude Opus 4.6 (no thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

GPQA Diamond ↗90.53%
90.5390.53

Claude Opus 4.6 (32k thinking) · GPQA diamond · Best score (across scorers) · % ↗

4 more reported results

88.76% · Claude Opus 4.6 (64k thinking) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

88.38% · Claude Opus 4.6 (max) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

89.65 % · Claude Opus 4.6 (Thinking) · GPQA Diamond · Accuracy · Vals AI · GPQA Diamond ↗

91.3 · Claude Opus 4.6 (max) · GPQA-Diamond · Kimi K2.6 technical model card ↗

V* ↗86.4*
86.498.4

Claude Opus 4.6 (max) · V* (w/ python) ↗

FrontierMath Tiers 1-3 (v1, old) ↗40.7%
40.740.7

Claude Opus 4.6 (max) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % ↗

3 more reported results

40% · Claude Opus 4.6 (32k thinking) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-2025-02-28-Private ↗

39.66% · Claude Opus 4.6 (64k thinking) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-2025-02-28-Private ↗

38.28% · Claude Opus 4.6 (no thinking) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-2025-02-28-Private ↗

FrontierMath Tier 4 (v1, old) ↗22.9%
22.922.9

Claude Opus 4.6 (max) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % ↗

3 more reported results

20.83% · Claude Opus 4.6 (64k thinking) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tier-4-2025-07-01-Private ↗

20.83% · Claude Opus 4.6 (32k thinking) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tier-4-2025-07-01-Private ↗

14.58% · Claude Opus 4.6 (no thinking) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tier-4-2025-07-01-Private ↗

Chess Puzzles ↗21%
076

Claude Opus 4.6 (max) · Chess Puzzles · Best score (across scorers) · % ↗

Cybench ↗93%
9393

Claude Opus 4.6 (unknown thinking) · Cybench · Unguided % Solved · % ↗

GSO ↗41.2%
41.241.2

Claude Opus 4.6 (high) · GSO-Bench · Score OPT@1 · % ↗

2 more reported results

33.33% · Claude Opus 4.6 (unknown thinking) · GSO-Bench · Score OPT@1 · % · Epoch AI · GSO-Bench ↗

33.33% · Claude Opus 4.6 (no thinking) · GSO-Bench · Score OPT@1 · % · Epoch AI · GSO-Bench ↗

SimpleBench ↗67.6%
45.880.6

Claude Opus 4.6 (no thinking) · SimpleBench · Score (AVG@5) · % ↗

GBAEval ↗44.12%
44.1244.12

Claude Opus 4.6 (unknown thinking) · GBAEval · Overall score · % ↗

CL-bench ↗20.7%
11.427.9

Claude Opus 4.6 (unknown thinking) · CL-bench · Overall · % ↗

LMCA ↗55.76%
2.7868.26

Claude Opus 4.6 (max) · LMCA · Score · % ↗

DTBench ↗91.2%
41.6598.93

Claude Opus 4.6 (max) · DTBench · Accuracy · % ↗

LegalBench ↗85.3 %
32.9788.56

Claude Opus 4.6 (Thinking) · LegalBench · Accuracy ↗

MedCode ↗49.13 %
49.1349.13

Claude Opus 4.6 (Thinking) · MedCode · Accuracy ↗

1 more reported result

48.24 % · Claude Opus 4.6 (Nonthinking) · MedCode · Accuracy · Vals AI · MedCode ↗

MedScribe ↗86.74 %
86.7486.74

Claude Opus 4.6 (Nonthinking) · MedScribe · Accuracy ↗

1 more reported result

86.13 % · Claude Opus 4.6 (Thinking) · MedScribe · Accuracy · Vals AI · MedScribe ↗

MMLU-Pro ↗89.11 %
30.2892.38

Claude Opus 4.6 (Thinking) · MMLU Pro · Accuracy ↗

Vibe Code Bench ↗57.57 %
57.5757.57

Claude Opus 4.6 (Nonthinking) (OpenHands) · Vibe Code Bench · original task set ↗

1 more reported result

53.5 % · Claude Opus 4.6 (Thinking) (OpenHands) · Vibe Code Bench · original task set · Vals AI · Vibe Code Bench ↗

Fortress ↗20.52%
5.7874.39

claude-opus-4-6 (Non-Thinking) · Scale Labs · Risk elicitation score ↗

TutorBench ↗53.55%
36.1268.55

claude-opus-4-6 (Non-Thinking) · Scale Labs · Score ↗

VISTA ↗45.48%
15.1854.65

claude-opus-4-6 (Non-Thinking) · Scale Labs · Score ↗

NYU CTF Bench ↗79%
11.179

Claude Opus 4.6 · NYU CTF Bench · Cotool 81-challenge subset · accuracy ↗

AlgoTune ↗1.47 ×
1.312.05

Claude Opus 4.6 (high) · Harmonic-mean speedup ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

17.1 percentile-point gap

21 of 33 shared benchmark comparisons

17.3 percentile-point gap

26 of 33 shared benchmark comparisons

19.0 percentile-point gap

22 of 33 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Reports featuring this model

Kimi K2.6 · earlier-2026 baseline · Kimi →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. ARC Prize - Leaderboard ↗

    ARC Prize Foundation

  2. Arena leaderboard — WebDev evaluation ↗

    Arena

  3. APEX-Agents-AA Benchmark | Artificial Analysis ↗

    Artificial Analysis

  4. Global-MMLU-Lite Benchmark | Artificial Analysis ↗

    Artificial Analysis

  5. Terminal-Bench Hard Benchmark | Artificial Analysis ↗

    Artificial Analysis

  6. OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning — published result table ↗

    Opsahl-Ong, Krista, Singhvi, Arnav, Collins, Jasmine, Zhou, Ivan, Wang, Cindy, Baheti, Ashutosh et al. · 2026

    All 13 authors

    Opsahl-Ong, Krista, Singhvi, Arnav, Collins, Jasmine, Zhou, Ivan, Wang, Cindy, Baheti, Ashutosh, Oertell, Owen, Portes, Jacob, Havens, Sam, Elsen, Erich, Bendersky, Michael, Zaharia, Matei, Chen, Xing

  7. TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts — published result table ↗

    Jaffe, Oliver, Sherburn, Dane · 2026

  8. FrontierCode Leaderboard ↗

    Cognition

  9. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  10. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  11. Algotune — methodology and published results ↗

    Epoch AI · 2026

  12. APEX-Agents ↗

    Epoch AI

  13. ARC-AGI-2 ↗

    Epoch AI

  14. Chess Puzzles ↗

    Epoch AI

  15. CL-bench ↗

    Epoch AI

  16. CL-bench Life ↗

    Epoch AI

  17. Earthborne Rangers (EBR-bench) ↗

    Epoch AI

  18. FrontierMath Tier 4 (v1, old) ↗

    Epoch AI

  19. FrontierMath Tier 4 (v2) ↗

    Epoch AI

  20. FrontierMath Tiers 1-3 (v1, old) ↗

    Epoch AI

  21. FrontierMath Tiers 1-3 (v2) ↗

    Epoch AI

  22. Furniture Assembly Benchmark ↗

    Epoch AI

  23. GPQA Diamond ↗

    Epoch AI

  24. Humanity's Last Exam ↗

    Epoch AI

  25. METR Time Horizons ↗

    Epoch AI

  26. Mystery Game Puzzles ↗

    Epoch AI

  27. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  28. Remote Labor Index ↗

    Epoch AI

  29. SimpleQA Verified ↗

    Epoch AI

  30. SWE-bench Verified ↗

    Epoch AI

  31. Terminal-Bench 2.0 ↗

    Epoch AI

  32. Capabilities & benchmarking — retained external evaluation records ↗

    Epoch AI · 2026

  33. Vending Bench 2 — methodology and published results ↗

    Epoch AI · 2026

  34. Text Arena (Coding) ↗

    Epoch AI

  35. WeirdML (v2) ↗

    Epoch AI

  36. AI benchmarks & capabilities ↗

    Epoch AI

  37. FutureSearch Evals ↗

    FutureSearch

  38. Claude Opus 4.6 (unknown thinking) — GBA Eval ↗

    GBA Eval

  39. WeirdML ↗

    Håvard Tveit Ihle

  40. moonshotai/Kimi-K2.6 · Hugging Face ↗

    Moonshot AI

  41. DrugDiscoveryBench — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  42. EnigmaEval ↗

    Scale Labs / SEAL

  43. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  44. HiL-Bench (Human-in-Loop Benchmark) — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  45. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  46. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  47. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  48. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  49. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  50. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  51. SWE Atlas - Refactoring — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  52. SWE Atlas - Test Writing — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  53. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  54. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  55. AI Model Benchmarks — LM Council ↗

    LM Council

  56. Metr Autonomy Task Horizon Suite — methodology and published results ↗

    METR · 2026

  57. System Card: Claude Opus 4.6 ↗

    Anthropic · 2026

  58. NYU CTF Bench — Cotool evaluation, February 2026 ↗

    Cotool · 2026

  59. Terminal-Bench 2.0 leaderboard ↗

    Terminal-Bench Team

  60. AIME — evaluation methodology and results ↗

    Vals AI · 2026

  61. CaseLaw — evaluation methodology and results ↗

    Vals AI · 2026

  62. CorpFin — evaluation methodology and results ↗

    Vals AI · 2026

  63. GPQA Diamond — evaluation methodology and results ↗

    Vals AI · 2026

  64. LiveCodeBench — evaluation methodology and results ↗

    Vals AI · 2026

  65. LegalBench — evaluation methodology and results ↗

    Vals AI · 2026

  66. MedCode — evaluation methodology and results ↗

    Vals AI · 2026

  67. MedQA — evaluation methodology and results ↗

    Vals AI · 2026

  68. MedScribe — evaluation methodology and results ↗

    Vals AI · 2026

  69. MMLU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  70. MMMU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  71. MortgageTax — evaluation methodology and results ↗

    Vals AI · 2026

  72. ProofBench — evaluation methodology and results ↗

    Vals AI · 2026

  73. SAGE — evaluation methodology and results ↗

    Vals AI · 2026

  74. SWE-bench Verified — evaluation methodology and results ↗

    Vals AI · 2026

  75. TaxEval — evaluation methodology and results ↗

    Vals AI · 2026

  76. Vibe Code Bench — evaluation methodology and results ↗

    Vals AI · 2026

Names used in source tables

Claude Opus 4.6 · Claude Opus 4.6 (120k thinking) · Claude Opus 4.6 (32k thinking) · Claude Opus 4.6 (64k thinking) · Claude Opus 4.6 (Claude Code) max · Claude Opus 4.6 (Max) · Claude Opus 4.6 (Non-reasoning, High) · Claude Opus 4.6 (Nonthinking) · Claude Opus 4.6 (Nonthinking) (OpenHands) · Claude Opus 4.6 (Thinking) · Claude Opus 4.6 (Thinking) (Mini-SWE-agent) · Claude Opus 4.6 (Thinking) (OpenHands) · Claude Opus 4.6 (high) · Claude Opus 4.6 (low) · Claude Opus 4.6 (max) · Claude Opus 4.6 (medium) · Claude Opus 4.6 (no thinking) · Claude Opus 4.6 (unknown thinking) · Opus 4.6 (Claude Code) · Opus 4.6 (Mini-SWE-Agent) · Opus 4.6 (max) · Opus-4.6 (Claude Code) · claude-opus-4-6 (Non-Thinking) · claude-opus-4-6 (max)