Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Web Agent Task Success on WebArena
Loading...
27.8
Task Success Rate (TSR)
DPO – ADARUBRIC-DA
-0.488
6.856
14.2
21.544
Oct 3, 2025
Oct 31, 2025
Nov 28, 2025
Dec 27, 2025
Jan 24, 2026
Feb 21, 2026
Mar 22, 2026
Task Success Rate (TSR)
Agent Success Rate (ASR)
Updated 25d ago
Evaluation Results
Method
Method
Links
Task Success Rate (TSR)
Agent Success Rate (ASR)
DPO – ADARUBRIC-DA
Backbone=Qwen2.5-7B, E...
2026.03
27.8
-
DPO – ADARUBRIC-GM
Backbone=Qwen2.5-7B, E...
2026.03
25.1
-
DPO – ADARUBRIC-WM
Backbone=Qwen2.5-7B, E...
2026.03
24.3
-
DPO – Prometheus
Backbone=Qwen2.5-7B, E...
2026.03
21
-
DPO – G-Eval
Backbone=Qwen2.5-7B, E...
2026.03
20.1
-
DPO – Random pairs
Backbone=Qwen2.5-7B, E...
2026.03
17.4
-
Qwen 2.5 7B Instruct (Environment Poisoning)
Defense=None, Stage=SF...
2025.10
17
100
SFT – Success only
Backbone=Qwen2.5-7B
2026.03
16.7
-
Qwen 2.5 7B Instruct (Environment Poisoning)
Defense=Granite Guardi...
2025.10
16.27
95.15
Qwen 2.5 7B Instruct (Environment Poisoning)
Defense=Llama-firewall...
2025.10
15.76
93.94
Base
Backbone=Qwen2.5-7B, E...
2026.03
12.3
-
Qwen 2.5 7B Instruct
Defense=None, Stage=Ba...
2025.10
0.6
-
Feedback
Search any
task
Search any
task