benchgap
Calibration

DAYJOB Healthcare → EBR-bench

EBR-bench is estimated from DAYJOB Healthcare with a Hill curve fitted on 11 models measured on both: y = 0.2408 + (1.1662 − 0.2408)·x^0.64 / (0.15878^0.64 + x^0.64), R² = 0.75, cross-validated error 14.8 pp. It is used for 0 estimates.

Fitted on

ModelDAYJOB HealthcareEBR-bench
Claude Opus 4.80.8%28.6%
Claude Opus 59.2%45.7%
Claude Opus 5.525.2%71.4%
Gemini 3.1 Pro0.0%14.3%
Gemini 4 Argon11.6%64.8%
Gemini 4 Carbon15.6%87.6%
GPT-5.50.0%34.3%
GPT-5.6 Sol1.2%44.8%
GPT-6.1 Sol10.8%54.3%
GPT-6 Astra11.2%76.2%
GPT-6 Sol2.8%53.3%