Claude Opus 4.6
#30 of 274 scored.
AnthropicFeb 5, 2026Closed
Capabilities Index
Range 153.4–157.3
155.2
Context
Takes text, image, file
1M
Input
Per 1M tokens
$5.00
Output
Per 1M tokens
$25
Benchmarks
How it scores.
- GPQA Diamond87.4%
36th of 186
- SWE-bench Verified78.7%
4th of 32
- FrontierMath66.0%
27th of 81
- ARC-AGI-269.2%
19th of 79
- Humanity's Last Exam31.1%
10th of 40
- SimpleQA Verified47.0%
32nd of 77
- Mock AIME94.4%
37th of 176
- Terminal-Bench79.8%
5th of 35
Everything else Epoch has for it
- ARC-AGI94.0%
- Cybench93.0%
- DTBench85.3%
- METR Time Horizons78.9%
- WeirdML78.0%
- LMCA65.6%
- SimpleBench61.1%
- DeepResearch Bench55.3%
- ProofBench50.0%
- APEX-Agents46.3%
- GSO-Bench41.2%
- FrontierMath-Tier-4-v2-Private26.8%
- FrontierCode26.6%
- CL-bench20.7%
- Mystery Game Puzzles17.4%
- CL-bench Life17.0%
- EBR-bench12.7%
- Chess Puzzles12.7%
- Remote Labor Index4.2%
- Furniture Assembly0.0%
On the index
Around it.
Sources