benchgap
Calibration

HLE w/o tools → AA-Omniscience Accuracy

AA-Omniscience Accuracy is estimated from HLE w/o tools with a linear curve fitted on 27 models measured on both: y = 0.9467·x + 0.0388, R² = 0.75, cross-validated error 8.6 pp. It is used for 3 estimates.

Estimated modelHLE w/o toolsAA-Omniscience AccuracySource
Pareto 26.949.0%50.3%estimated ± 8.6 pp, medium confidence
Beam36.2%38.2%estimated ± 8.6 pp, medium confidence
Solar Open 228.8%31.1%estimated ± 8.6 pp, medium confidence