Calibration

FrontierCode 1.1 Main → SWE-bench Pro

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

SWE-bench Pro is estimated from FrontierCode 1.1 Main with a Michaelis–Menten curve fitted on 8 models measured on both: y = 2.0000·x / (0.80633 + x), R² = 0.70, cross-validated error 7.5 pp. It is used for 2 estimates.

Estimated model	FrontierCode 1.1 Main	SWE-bench Pro	Source
SWE-1.7	42.3%	68.8%	estimated ± 7.5 pp, medium confidence
SWE-2	50.0%	76.6%	estimated ± 7.5 pp, medium confidence