Claude Opus 4.1
#92 of 274 scored.
AnthropicAug 5, 2025Closed
Capabilities Index
Range 141.7–146.0
144.1
Context
Takes image, text, file
200K
Input
Per 1M tokens
$15
Output
Per 1M tokens
$75
Benchmarks
How it scores.
- GPQA Diamond69.7%
86th of 186
- SWE-bench Verified73.4%
20th of 32
- FrontierMath12.6%
75th of 81
- Humanity's Last Exam7.1%
25th of 40
- Mock AIME68.9%
95th of 176
- Terminal-Bench38.0%
21st of 35
Everything else Epoch has for it
- Lech Mazur Writing84.7%
- METR Time Horizons66.8%
- DTBench66.7%
- SimpleBench52.0%
- DeepResearch Bench48.3%
- WeirdML45.9%
- LMCA43.6%
- GDPval43.6%
- Cybench42.0%
- Mystery Game Puzzles13.0%
- EBR-bench7.9%
- VPCT2.5%
- FrontierMath-Tier-4-v2-Private2.4%
- Chess Puzzles2.1%
On the index
Around it.
Sources