Calibration
CursorBench 4.0 → Vals LiveCodeBench
Vals LiveCodeBench is estimated from CursorBench 4.0 with a Michaelis–Menten curve fitted on 7 models measured on both: y = 1.0889·x / (0.10610 + x), R² = 0.48, cross-validated error 2.7 pp. It is used for 4 estimates.
| Estimated model | CursorBench 4.0 | Vals LiveCodeBench | Source |
|---|---|---|---|
| Composer 2.5 | 27.7% | 78.7% | estimated ± 2.7 pp, low confidence |
| GPT-5.6 Luna | 35.9% | 84.0% | estimated ± 2.7 pp, low confidence |
| Grok 4.7 | 46.3% | 88.6% | estimated ± 2.7 pp, low confidence |
| Muse Spark 1.3 | 41.6% | 86.8% | estimated ± 2.7 pp, low confidence |