in the loop

Claude Opus 5
#6 of 274 scored.

AnthropicJul 24, 2026Closed

Capabilities Index

Range 160.1–166.5

162.8

Context

Takes text, image, file

1M

Input

Per 1M tokens

$5.00

Output

Per 1M tokens

$25

Benchmarks

How it scores.

  • GPQA Diamond91.8%

    13th of 186

  • FrontierMath85.6%

    11th of 81

  • ARC-AGI-290.4%

    5th of 79

  • SimpleQA Verified59.9%

    16th of 77

  • Mock AIME98.9%

    16th of 176

Everything else Epoch has for it

  • ProofBench99.0%
  • ARC-AGI97.5%
  • DTBench96.5%
  • WeirdML91.8%
  • SimpleBench76.7%
  • LMCA75.8%
  • DeepSWE73.7%
  • FrontierMath-Tier-4-v2-Private73.2%
  • APEX-Agents65.8%
  • Balrog63.4%
  • Mystery Game Puzzles54.8%
  • FrontierCode53.4%
  • FrontierSWE52.0%
  • EBR-bench45.7%
  • Furniture Assembly44.0%
  • Chess Puzzles39.0%
  • PostTrainBench35.0%
  • OSWorld 2.031.4%

On the index

Around it.

  1. #3GPT-6.1 SolOpenAI · $2.00 in166.1
  2. #4Claude Sonnet 5.5Anthropic · $2.00 in165.0
  3. #5Claude Fable 5.1Anthropic · $10 in164.7
  4. #6Claude Opus 5Anthropic · $5.00 in162.8
  5. #7GPT-6 SolOpenAI · $2.00 in162.7
  6. #8GPT-5.5 ProOpenAI · $30 in162.1
  7. #9Claude Fable 5Anthropic · $10 in162.1

Sources