Calibration

VITA-Bench → Claw-Eval

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

Claw-Eval is estimated from VITA-Bench with a Hill curve fitted on 7 models measured on both: y = 0.0000 + (0.6308 − 0.0000)·x^6.00 / (0.16500^6.00 + x^6.00), R² = 0.77, cross-validated error 7.1 pp. It is used for 2 estimates.

Estimated model	VITA-Bench	Claw-Eval	Source
Agents-A1	38.8%	62.7%	estimated ± 7.1 pp, low confidence
Agents-A1-4B	40.3%	62.8%	estimated ± 7.1 pp, low confidence