Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Long-form factuality evaluation on FactBench

84.4Accuracy

CURE

73.16876.0847981.916Apr 13, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.04
84.40.1360.1950.667
2026.04
82.50.1590.2140.648
2026.04
820.1370.1840.563
2026.04
78.10.1570.20.605
2026.04
77.1---
2026.04
73.60.1620.2190.541
2026.04
73.60.2120.2760.575