Calibration
We-Math → DynaMath
DynaMath is estimated from We-Math with a Michaelis–Menten curve fitted on 6 models measured on both: y = 1.2163·x / (0.35951 + x), R² = 0.76, cross-validated error 1.9 pp. It is used for 0 estimates.
Fitted on
| Model | We-Math | DynaMath |
|---|---|---|
| Claude Opus 4.5 | 70.0% | 79.7% |
| Gemini 3 Pro | 86.9% | 85.1% |
| GPT-5.2 | 79.0% | 86.8% |
| Kimi K2.5 | 84.7% | 84.4% |
| LFM3-VL-2B | 68.6% | 79.2% |
| Qwen3.5 397B | 87.9% | 86.3% |