Calibration

Artificial Analysis Intelligence Index → HealthBench (raw)

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

HealthBench (raw) is estimated from Artificial Analysis Intelligence Index with a Michaelis–Menten + offset curve fitted on 7 models measured on both: y = 0.0000 + 2.0000·x / (1.19480 + x), R² = 0.53, cross-validated error 6.7 pp. It is used for 4 estimates.

Estimated model	Artificial Analysis Intelligence Index	HealthBench (raw)	Source
Claude 4.1 Opus	18.6%	26.9%	estimated ± 6.7 pp, low confidence
o1-preview	11.4%	17.4%	estimated ± 6.7 pp, low confidence
o1-pro	12.4%	18.8%	estimated ± 6.7 pp, low confidence
o3-pro	21.9%	30.9%	estimated ± 6.7 pp, low confidence