Calibration
CursorBench 4.0 → AA Coding Index
AA Coding Index is estimated from CursorBench 4.0 with a Michaelis–Menten + offset curve fitted on 9 models measured on both: y = 0.0000 + 1.1192·x / (0.19326 + x), R² = 0.93, cross-validated error 1.0 pp. It is used for 4 estimates.
| Estimated model | CursorBench 4.0 | AA Coding Index | Source |
|---|---|---|---|
| Claude Opus 5.5 | 57.8% | 83.9% | estimated ± 1.0 pp, medium confidence |
| Claude Sonnet 5.5 | 55.5% | 83.0% | estimated ± 1.0 pp, medium confidence |
| Composer 2.5 | 27.7% | 65.9% | estimated ± 1.0 pp, medium confidence |
| Grok 4.7 | 46.3% | 79.0% | estimated ± 1.0 pp, high confidence |