GPT-5 nano
#122 of 274 scored.
OpenAIAug 7, 2025Closed
Capabilities Index
Range 134.1–141.4
139.4
Context
Takes text, image, file
400K
Input
Per 1M tokens
$0.05
Output
Per 1M tokens
$0.40
Benchmarks
How it scores.
- GPQA Diamond59.3%
102nd of 186
- FrontierMath20.0%
69th of 81
- ARC-AGI-22.6%
60th of 79
- SimpleQA Verified11.7%
73rd of 77
- Mock AIME81.1%
79th of 176
- Terminal-Bench21.8%
30th of 35
Everything else Epoch has for it
- MATH level 595.2%
- Fiction.LiveBench44.4%
- WeirdML38.1%
- DTBench37.8%
- Chess Puzzles23.2%
- ARC-AGI20.7%
- ProofBench12.0%
- LMCA9.3%
- VPCT5.8%
- FrontierMath-Tier-4-v2-Private2.4%
- Mystery Game Puzzles0.0%
On the index
Around it.
Sources