GPT-5.1
#57 of 274 scored.
OpenAINov 13, 2025Closed
Capabilities Index
Range 148.2–151.1
149.6
Context
Takes image, text, file
400K
Input
Per 1M tokens
$1.25
Output
Per 1M tokens
$10
Benchmarks
How it scores.
- GPQA Diamond83.5%
52nd of 186
- SWE-bench Verified68.0%
25th of 32
- ARC-AGI-217.6%
40th of 79
- Humanity's Last Exam19.8%
16th of 40
- SimpleQA Verified48.0%
30th of 77
- Mock AIME88.6%
58th of 176
- Terminal-Bench47.6%
15th of 35
Everything else Epoch has for it
- DTBench83.5%
- ARC-AGI72.8%
- WeirdML60.8%
- FrontierMath-2025-02-28-Private54.4%
- LMCA51.6%
- SimpleBench43.8%
- DeepResearch Bench42.8%
- VPCT38.0%
- Chess Puzzles28.4%
- CL-bench23.7%
- FrontierMath-Tier-4-2025-07-01-Private20.8%
- CL-bench Life17.3%
- GSO-Bench13.7%
- Mystery Game Puzzles10.8%
On the index
Around it.
Sources