Calibration
AA AutomationBench → GDPval-AA
GDPval-AA is estimated from AA AutomationBench with a linear curve fitted on 23 models measured on both: y = 0.4999·x + 0.2414, R² = 0.76, cross-validated error 6.9 pp. It is used for 1 estimate.
| Estimated model | AA AutomationBench | GDPval-AA | Source |
|---|---|---|---|
| Claude Haiku 5.5 | 35.4% | 41.8% | estimated ± 6.9 pp, medium confidence |