A focused specialist that lives and breathes mathematics — it handles algebraic reasoning, equation solving, and multi-step numerical problems with a precision that generalist models of similar size rarely match. Its instruction-tuned nature means it follows mathematical problem statements cleanly and shows working in structured steps. Outside of math-heavy tasks, though, it's a narrow tool; don't expect broad general knowledge or strong prose.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| IFEval | 26.4 | accuracy | 4d ago |