DeepSeek-V3
#151 of 274 scored.
DeepSeekDec 26, 2024Open weights
Capabilities Index
Range 127.5–135.7
132.3
Context
Takes text
164K
Input
Per 1M tokens
$0.32
Output
Per 1M tokens
$0.89
Benchmarks
How it scores.
- GPQA Diamond42.0%
122nd of 186
- Mock AIME15.8%
132nd of 176
Everything else Epoch has for it
- ARC AI293.7%
- HellaSwag85.2%
- BBH83.3%
- MMLU82.9%
- TriviaQA82.9%
- Winogrande70.4%
- PIQA69.4%
- MATH level 564.8%
- Aider polyglot48.4%
- METR Time Horizons47.4%
- FrontierMath-2025-02-28-Private3.0%
- SimpleBench2.7%
On the index
Around it.
Sources