GPT-5 mini
#84 of 274 scored.
OpenAIAug 7, 2025Closed
Capabilities Index
Range 143.6–147.2
145.5
Context
Takes text, image, file
400K
Input
Per 1M tokens
$0.25
Output
Per 1M tokens
$2.00
Benchmarks
How it scores.
- GPQA Diamond66.7%
95th of 186
- SWE-bench Verified64.7%
26th of 32
- FrontierMath46.7%
47th of 81
- ARC-AGI-24.4%
56th of 79
- Humanity's Last Exam15.4%
19th of 40
- SimpleQA Verified21.6%
64th of 77
- Mock AIME86.7%
61st of 176
- Terminal-Bench34.8%
24th of 35
Everything else Epoch has for it
- MATH level 597.9%
- Lech Mazur Writing83.1%
- Fiction.LiveBench69.4%
- DTBench67.5%
- ARC-AGI54.3%
- WeirdML52.7%
- LMCA40.2%
- Chess Puzzles26.4%
- FrontierMath-Tier-4-v2-Private12.2%
- VPCT10.3%
- ProofBench9.0%
- Mystery Game Puzzles0.9%
On the index
Around it.
Sources