Calibration
HLE → GPQA
GPQA is estimated from HLE with a Michaelis–Menten curve fitted on 42 models measured on both: y = 0.9671·x / (0.02838 + x), R² = 0.54, cross-validated error 2.2 pp. It is used for 2 estimates.
| Estimated model | HLE | GPQA | Source |
|---|---|---|---|
| Agents-A1 | 47.6% | 91.3% | estimated ± 2.2 pp, high confidence |
| Apodex 1.1 | 56.1% | 92.1% | estimated ± 2.2 pp, high confidence |