Calibration

FrontierSWE v2 → FrontierCode 1.1 Main

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

FrontierCode 1.1 Main is estimated from FrontierSWE v2 with a Michaelis–Menten curve fitted on 6 models measured on both: y = 0.5757·x / (0.06070 + x), R² = 0.50, cross-validated error 5.2 pp. It is used for 0 estimates.