Calibration

LiveCodeBench → AA-SciCode

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

AA-SciCode is estimated from LiveCodeBench with a Michaelis–Menten + offset curve fitted on 6 models measured on both: y = 0.2644 + 2.0000·x / (8.90635 + x), R² = 0.56, cross-validated error 7.0 pp. It is used for 0 estimates.