First Sonnet to score 100% on AIME, closed the Opus gap entirely
| Benchmark | Score | Rank |
|---|---|---|
MATH Competition-level mathematics problems | 100% | #2 / 49 |
HumanEval Coding ability - generating correct Python functions | 95.8% | #6 / 49 |
ARC-C Grade-school science questions requiring reasoning | 98.2% | #7 / 40 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 87.4% | #8 / 30 |
SWE-bench Real-world GitHub issue resolution | 77.2% | #9 / 38 |
MMLU Tests knowledge across 57 subjects from STEM to humanities | 91.5% | #13 / 53 |
MMMUvals.ai College-level multimodal reasoning across 30+ disciplines | 79.3% | #13 / 33 |
HellaSwag Common sense reasoning about everyday situations | 95.1% | #14 / 36 |
ARC-AGIARC Prize Novel reasoning tasks requiring fluid intelligence | 14.7% | #15 / 21 |
Arena Elo Human preference ranking via blind comparisons | 1412 | #18 / 41 |
LiveCodeBenchvals.ai Contamination-free competitive programming (filtered by cutoff date) | 73% | #22 / 31 |
GPQA PhD-level science questions even experts struggle with | 83.4% | #23 / 54 |