Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Reward Prediction on UI-TARS-7B trajectories
Loading...
96.5
Accuracy
PRORE
83.708
87.029
90.35
93.671
Sep 26, 2025
Accuracy
F1 Score
Updated 3d ago
Evaluation Results
Method
Method
Links
Accuracy
F1 Score
PRORE
Last N States=Proactive
2025.09
96.5
66.7
WebRL
Last N States=Full
2025.09
93.9
22.2
WebRL
Last N States=1
2025.09
93
20
Step-Critic
Last N States=1
2025.09
93
20
Step-Critic
Last N States=Full
2025.09
93
20
DistRL
Last N States=Full
2025.09
89.5
14.3
DigiRL
Last N States=Full
2025.09
86.8
11.7
DistRL
Last N States=1
2025.09
84.2
25
DigiRL
Last N States=1
2025.09
84.2
10
Feedback
Search any
task
Search any
task