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

Reward Verification on AgentRewardBench VisualWebArena

100Precision

Rule-based (VWA Oracle)

71.9279.2186.593.79Jul 15, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.07
100
2025.07
86
2025.07
85
2025.07
80
2025.07
80
2025.07
75
2025.07
73