Calibration

FrontierCode 1.1 Extended → SWE-bench Pro

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

SWE-bench Pro is estimated from FrontierCode 1.1 Extended with a linear curve fitted on 6 models measured on both: y = 2.3932·x + -0.6920, R² = 0.59, cross-validated error 9.0 pp. It is used for 0 estimates.