Calibration

HLE w/o tools → ARC-AGI-3

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

ARC-AGI-3 is estimated from HLE w/o tools with a offset logistic curve fitted on 7 models measured on both: y = 0.0027 + (0.3016 − 0.0027) / (1 + exp(−200.00·(x − 0.5136))), R² = 1.00, cross-validated error 11.6 pp. It is used for 1 estimate.

Estimated model	HLE w/o tools	ARC-AGI-3	Source
Pareto 26.9	49.0%	0.5%	estimated ± 11.6 pp, low confidence