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

Policy Evaluation on PolicyBench Level 3 US

77Accuracy

o4-mini

65.986468.845771.70574.5643Apr 14, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.04
77
74.6
2026.04
69.9
2026.04
69.44
69.39
2026.04
68.47
68.37
2026.04
68.28
2026.04
68.13
2026.04
66.55
2026.04
66.41