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

Computer-Using Agent Task on WebArena shopping_admin 12-task subset (Cold) (Run 1)

7.25Success Rate

PreAct gate-ON + cache-miss CUA fallback

5.696.0956.56.905Jun 16, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
7.25
2026.06
7
2026.06
7
2026.06
6
2026.06
5.75