benchgap
Calibration

τ³-bench results → GDPval-AA

GDPval-AA is estimated from τ³-bench results with a offset logistic curve fitted on 14 models measured on both: y = 0.0207 + (0.2248 − 0.0207) / (1 + exp(−132.15·(x − 0.6484))), R² = 0.58, cross-validated error 9.6 pp. It is used for 0 estimates.