Claude Sonnet 5
#26 of 274 scored.
AnthropicJun 30, 2026Closed
Capabilities Index
Range 153.5–158.8
156.2
Context
Takes text, image, file
1M
Input
Per 1M tokens
$2.00
Output
Per 1M tokens
$10
Benchmarks
How it scores.
- GPQA Diamond87.4%
36th of 186
- FrontierMath65.6%
29th of 81
- SimpleQA Verified33.7%
54th of 77
- Mock AIME94.7%
36th of 176
Everything else Epoch has for it
- DTBench87.5%
- ProofBench77.0%
- WeirdML68.8%
- Surface Evolver Bench60.0%
- LMCA58.9%
- APEX-Agents54.5%
- DeepSWE53.8%
- SimpleBench52.7%
- FrontierCode42.7%
- GSO-Bench37.3%
- Chess Puzzles31.6%
- FrontierMath-Tier-4-v2-Private29.3%
- Mystery Game Puzzles28.4%
On the index
Around it.
Sources