benchgap
Calibration

VulcanBench v3 → CursorBench 4.0

CursorBench 4.0 is estimated from VulcanBench v3 with a offset logistic curve fitted on 5 models measured on both: y = 0.3507 + (0.4866 − 0.3507) / (1 + exp(−199.55·(x − 0.8687))), R² = 0.65, cross-validated error 3.6 pp. It is used for 0 estimates.

Fitted on

ModelVulcanBench v3CursorBench 4.0
Claude Opus 587.0%46.6%
GPT-5.6 Luna85.5%35.9%
GPT-5.6 Sol87.0%41.7%
GPT-5.6 Terra87.0%41.3%
Grok 4.687.0%41.4%