in the loop

Claude Fable 5
#9 of 274 scored.

AnthropicJun 9, 2026Closed

Capabilities Index

Range 159.3–166.2

162.1

Context

Takes text, image, file

1M

Input

Per 1M tokens

$10

Output

Per 1M tokens

$50

Benchmarks

How it scores.

  • GPQA Diamond81.1%

    61st of 186

  • FrontierMath87.0%

    9th of 81

  • ARC-AGI-289.2%

    8th of 79

  • SimpleQA Verified70.7%

    6th of 77

  • Mock AIME100%

    1st of 176

Everything else Epoch has for it

  • ARC-AGI98.5%
  • DTBench97.3%
  • Surface Evolver Bench95.0%
  • ProofBench95.0%
  • WeirdML91.9%
  • FrontierMath-Tier-4-v2-Private90.2%
  • SimpleBench78.3%
  • LMCA71.9%
  • DeepSWE69.9%
  • MirrorCode63.9%
  • APEX-Agents63.6%
  • FrontierCode53.5%
  • Mystery Game Puzzles47.1%
  • FrontierSWE47.0%
  • PostTrainBench41.8%
  • EBR-bench39.5%
  • Chess Puzzles37.9%
  • Remote Labor Index16.1%
  • Furniture Assembly8.3%

On the index

Around it.

  1. #6Claude Opus 5Anthropic · $5.00 in162.8
  2. #7GPT-6 SolOpenAI · $2.00 in162.7
  3. #8GPT-5.5 ProOpenAI · $30 in162.1
  4. #9Claude Fable 5Anthropic · $10 in162.1
  5. #10GPT-5.6 SolOpenAI · $2.00 in161.7
  6. #11GPT-5.6 TerraOpenAI · $2.00 in159.6
  7. #12GPT-5.5OpenAI · $5.00 in159.1

Sources