Calibration

AA ITBench → ExploitGym

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

ExploitGym is estimated from AA ITBench with a Hill curve fitted on 7 models measured on both: y = 0.0000 + (0.5609 − 0.0000)·x^6.00 / (0.51713^6.00 + x^6.00), R² = 0.35, cross-validated error 10.3 pp. It is used for 1 estimate.

Estimated model	AA ITBench	ExploitGym	Source
GLM-5.2	42.7%	13.5%	estimated ± 10.3 pp, low confidence