Major leap for Mistral, closed the gap with US frontier labs
| Benchmark | Score | Rank |
|---|---|---|
HellaSwag Common sense reasoning about everyday situations | 94.8% | #15 / 36 |
ARC-C Grade-school science questions requiring reasoning | 97.2% | #17 / 40 |
HumanEval Coding ability - generating correct Python functions | 93.6% | #18 / 49 |
MMLU Tests knowledge across 57 subjects from STEM to humanities | 89.8% | #20 / 53 |
Arena Elo Human preference ranking via blind comparisons | 1352 | #26 / 41 |
MATH Competition-level mathematics problems | 85.4% | #28 / 49 |
SWE-bench Real-world GitHub issue resolution | 58.6% | #31 / 38 |
GPQA PhD-level science questions even experts struggle with | 72.1% | #37 / 54 |