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

Human Correlation on AgentBench

0.77Pearson r

ADARUBRIC-DA

0.27080.40040.530.6596Mar 22, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.03
0.770.7670.147
2026.03
0.740.7370.117
2026.03
0.720.720.1
2026.03
0.620.62-
2026.03
0.590.590.05
2026.03
0.520.5170.123
2026.03
0.410.410.23
2026.03
0.290.2870.353