Calibration

HealthBench (length-adjusted) → HealthBench Professional

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

HealthBench Professional is estimated from HealthBench (length-adjusted) with a Michaelis–Menten + offset curve fitted on 7 models measured on both: y = 0.0108 + 2.0000·x / (1.28142 + x), R² = 0.78, cross-validated error 1.8 pp. It is used for 0 estimates.