RLTasksby ulam.ai
Models / Anthropic

Anthropic / MODEL EVIDENCE

Claude Fable 5

82 benchmarks · 128 observations · 77 sources · released 2026-06-09

Scores at a glance

82 benchmarks · large dot: Claude Fable 5; small dots: closest models where reported

AA-AnalystAgent ↗48.75%
23.7560

Claude Fable 5 (Max, Opus 4.8 Fallback) · AA-AnalystAgent · strict score ↗

GPQA Diamond ↗93.18 %
27.0295.45

Claude Fable 5 · GPQA Diamond · Accuracy ↗

5 more reported results

92.63% · Claude Fable 5 (Max, Opus 4.8 Fallback) · GPQA Diamond · AA implementation · Artificial Analysis ↗

85.86% · Claude Fable 5 (max) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

83.33% · Claude Fable 5 (high) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

78.79% · Claude Fable 5 (low) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

92.6% · Claude Fable 5 · GPQA Diamond · GPT-6 Astra launch report ↗

DeepSWE v1.1 ↗69.9%
67.474.1

Claude Fable 5 · DeepSWE v1.1 ↗

5 more reported results

69.91% · Claude Fable 5 (xhigh) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

69.72% · Claude Fable 5 (max) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

68.6% · Claude Fable 5 (high) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

65.37% · Claude Fable 5 (medium) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

59.58% · Claude Fable 5 (low) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

HealthBench ↗60.9%†
52.163.4

Claude Fable 5 · HealthBench · HealthBench Professional · length-adjusted ↗

ARC-AGI-2 ↗89.2%
89.295

Claude Fable 5 · ARC-AGI-2 ↗

5 more reported results

89.17% · Claude Fable 5 (max) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

88.33% · Claude Fable 5 (xhigh) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

87.5% · Claude Fable 5 (high) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

82.5% · Claude Fable 5 (medium) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

76.81% · Claude Fable 5 (low) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

MirrorCode ↗63.89%
8.8977.4

Claude Fable 5 (high) · MirrorCode · Best score (across scorers) · % ↗

CritPt ↗28.57%
032.29

Claude Fable 5 (max) · CritPt · Accuracy · % ↗

GDP.pdf ↗29.8%
235

Claude Fable 5 (max) · GDP.pdf · GDP.pdf score · % ↗

SciCode ↗61%
1.2766.9

Claude Fable 5 (max) · SciCode · Score · % ↗

EnigmaEval ↗39.28%
0.3839.28

Claude Fable 5 (high) · EnigmaEval · Accuracy · % ↗

WeirdML (v3) ↗27.86%
4.3854.05

Claude Fable 5 (xhigh) · WeirdML v3 · Max performance · % ↗

Vals RSI Index ↗0.38 reference score
00.46

Claude Fable 5 (max) · Compression · reference-anchored score ↗

3 more reported results

0.41 reference score · Claude Fable 5 (max) · LM training · reference-anchored score · Vals RSI Index · LM training ↗

0.18 reference score · Claude Fable 5 (max) · Harness engineering · reference-anchored score · Vals RSI Index · Harness engineering ↗

0.00 reference score · Claude Fable 5 (max) · Post-training · reference-anchored score · Vals RSI Index · Post-training ↗

Kimi Code Bench 2.0 ↗76.9%
64.276.9

Claude Fable 5 (max, w/ fallback) · Kimi Code Bench 2.0 · score · Kimi K3 · official model-card comparison ↗

FrontierMath: Open Problems ↗1 credited problems
12

Claude Fable 5 · FrontierMath: Open Problems · FM Open Problems · first credited AI solution records · Epoch public solution registry · October 27, 2026 ↗

BTF-3 ↗0.13 Brier scale
0.120.15

Claude Fable 5 (high) · Pooled forecasting score ↗

Closest models by score

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

11.3 percentile-point gap

35 of 47 shared benchmark comparisons

16.3 percentile-point gap

30 of 47 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

GPT-6 Astra · OpenAI →

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. AA-AnalystAgent Benchmark | Artificial Analysis ↗

    Artificial Analysis

  4. EnterpriseOps-Gym-AA Benchmark | Artificial Analysis ↗

    Artificial Analysis

  5. GPQA Diamond Benchmark | Artificial Analysis ↗

    Artificial Analysis

  6. SciCode Benchmark | Artificial Analysis ↗

    Artificial Analysis

  7. 𝜏²-Bench Telecom Benchmark | Artificial Analysis ↗

    Artificial Analysis

  8. 𝜏³-Banking Benchmark | Artificial Analysis ↗

    Artificial Analysis

  9. Terminal-Bench 2.1 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  10. Terminal-Bench Hard Benchmark | Artificial Analysis ↗

    Artificial Analysis

  11. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  12. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  13. DeepSWE ↗

    Datacurve

  14. APEX-Agents ↗

    Epoch AI

  15. ARC-AGI-2 ↗

    Epoch AI

  16. Chess Puzzles ↗

    Epoch AI

  17. CritPt v1 ↗

    Epoch AI

  18. Earthborne Rangers (EBR-bench) ↗

    Epoch AI

  19. FrontierMath Erdős ↗

    Epoch AI

  20. FrontierMath Tier 4 (v2) ↗

    Epoch AI

  21. FrontierMath Tiers 1-3 (v2) ↗

    Epoch AI

  22. Furniture Assembly Benchmark ↗

    Epoch AI

  23. GPQA Diamond ↗

    Epoch AI

  24. MirrorCode ↗

    Epoch AI

  25. Mystery Game Puzzles ↗

    Epoch AI

  26. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  27. PostTrainBench v1.1 ↗

    Epoch AI

  28. Remote Labor Index ↗

    Epoch AI

  29. SimpleQA Verified ↗

    Epoch AI

  30. SimpleBench ↗

    Epoch AI

  31. Vending Bench 2 — methodology and published results ↗

    Epoch AI · 2026

  32. WeirdML (v2) ↗

    Epoch AI

  33. AI benchmarks & capabilities ↗

    Epoch AI

  34. FrontierMath: Open Problems — public problem and solution registry ↗

    Epoch AI · 2026

  35. FutureSearch Evals ↗

    FutureSearch

  36. WeirdML ↗

    Håvard Tveit Ihle

  37. Kimi K3: Open Frontier Intelligence — official model card ↗

    Moonshot AI · 2026

  38. DistressBench — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  39. EnigmaEval ↗

    Scale Labs / SEAL

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

    Scale Labs / SEAL · 2026

  41. Humanity's Sixth Sense — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  42. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  43. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  44. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  45. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

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

    Scale Labs / SEAL · 2026

  47. GPT-6 Astra: A new generation of intelligence ↗

    OpenAI

  48. GDP.pdf Benchmark | Surge AI ↗

    Surge AI

  49. FrontierSWE — software engineering benchmark ↗

    FrontierSWE

  50. Code Migration v1 — evaluation methodology and results ↗

    Vals AI · 2026

  51. CorpFin — evaluation methodology and results ↗

    Vals AI · 2026

  52. Excel Modeling Benchmark v1 — evaluation methodology and results ↗

    Vals AI · 2026

  53. Finance Agent v2 — evaluation methodology and results ↗

    Vals AI · 2026

  54. GPQA Diamond — evaluation methodology and results ↗

    Vals AI · 2026

  55. Harvey's Legal Agent Benchmark v1 — evaluation methodology and results ↗

    Vals AI · 2026

  56. LiveCodeBench — evaluation methodology and results ↗

    Vals AI · 2026

  57. LegalBench — evaluation methodology and results ↗

    Vals AI · 2026

  58. Legal Research Bench — evaluation methodology and results ↗

    Vals AI · 2026

  59. MedCode — evaluation methodology and results ↗

    Vals AI · 2026

  60. MedScribe — evaluation methodology and results ↗

    Vals AI · 2026

  61. MMLU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  62. MMMU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  63. MortgageTax — evaluation methodology and results ↗

    Vals AI · 2026

  64. ProgramBench — evaluation methodology and results ↗

    Vals AI · 2026

  65. ProofBench — evaluation methodology and results ↗

    Vals AI · 2026

  66. Public Benefits Bench — evaluation methodology and results ↗

    Vals AI · 2026

  67. Public Benefits Bench v1 — evaluation methodology and results ↗

    Vals AI · 2026

  68. Vals RSI Index — evaluation methodology and results ↗

    Vals AI · 2026

  69. SAGE — evaluation methodology and results ↗

    Vals AI · 2026

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

    Vals AI · 2026

  71. Tax Agent Bench — evaluation methodology and results ↗

    Vals AI · 2026

  72. TaxEval — evaluation methodology and results ↗

    Vals AI · 2026

  73. Terminal-Bench 4.0 — evaluation methodology and results ↗

    Vals AI · 2026

  74. Terminal-Bench-Science — evaluation methodology and results ↗

    Vals AI · 2026

  75. Vals Multimodal Index — evaluation methodology and results ↗

    Vals AI · 2026

  76. Vibe Code Bench — evaluation methodology and results ↗

    Vals AI · 2026

  77. Surface Evolver Bench ↗

    yhenon

Names used in source tables

Claude Fable 5 · Claude Fable 5 (Max, Opus 4.8 Fallback) · Claude Fable 5 (Mini-SWE-agent) · Claude Fable 5 (OpenHands) · Claude Fable 5 (high) · Claude Fable 5 (low) · Claude Fable 5 (max) · Claude Fable 5 (max, w/ fallback) · Claude Fable 5 (medium) · Claude Fable 5 (xhigh) · Claude Fable 5 - High · Claude Fable 5 - Low · Claude Fable 5 - Max · Claude Fable 5 - Medium · Claude-Fable-5 (max) · Fable 5 · Fable-5 (Claude Code) xHigh · claude-fable 5 · claude-fable-5