phi-3-mini 3.8B
#209 of 274 scored.
MicrosoftApr 23, 2024Open weights
Capabilities Index
Range 107.9–120.7
117.4
Context
Tokens
—
Input
Per 1M tokens
—
Output
Per 1M tokens
—
Benchmarks
No independent scores yet.
Epoch AI hasn't run phi-3-mini 3.8B yet. Its scores show up here the morning after they do.
Everything else Epoch has for it
- OpenBookQA84.0%
- ARC AI279.9%
- HellaSwag68.9%
- TriviaQA64.0%
- BBH62.3%
- MMLU58.4%
- Winogrande41.6%
- ANLI29.2%
- Chess Puzzles0.0%
On the index
Around it.
Sources