Hybrid reasoning MoE trained on 36T tokens across 119 languages
| Benchmark | Score | Rank |
|---|---|---|
MATH Competition-level mathematics problems | 85.7% | #28 / 50 |
MMLU Tests knowledge across 57 subjects from STEM to humanities | 88.4% | #33 / 54 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 85% | #36 / 52 |
LiveCodeBenchvals.ai Contamination-free competitive programming (filtered by cutoff date) | 78.2% | #41 / 54 |
TerminalArtificial Analysis Agentic terminal coding tasks requiring multi-step execution | 13.6% | #51 / 56 |
hleArtificial Analysis | 15% | #51 / 69 |
GPQA PhD-level science questions even experts struggle with | 81.1% | #55 / 81 |