Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Web Navigation and E-commerce on WebShop In-Distribution
Loading...
94.73
Average Score
SkillFlow
49.0428
60.9039
72.765
84.6261
May 13, 2026
May 20, 2026
May 27, 2026
Jun 3, 2026
Jun 10, 2026
Jun 17, 2026
Jun 25, 2026
Average Score
Success Rate
Updated 29d ago
Evaluation Results
Method
Method
Links
Average Score
Success Rate
SkillFlow
Method Category=Ours
2026.05
94.73
93.75
SkillRL
Method Category=Agent+RL
2026.05
87.96
82.81
FlowSteer
Method Category=Agent+RL
2026.05
85.43
79.2
v4-flash
Method Category=Baseline
2026.05
85.17
67.97
AgentFlow
Method Category=Agent+RL
2026.05
83.61
74.22
Task-Perturbed NLL Optimization
Backbone=Qwen3-8B
2026.06
81
-
Task-Perturbed NLL Optimization
Backbone=Qwen3-4B
2026.06
77
-
GRPO
Backbone=Qwen3-8B
2026.06
76
-
GRPO
Backbone=Qwen3-4B
2026.06
74
-
Qwen3.5-AFlow
Method Category=AFlow,...
2026.05
71.09
59.38
Qwen3.5-9B
Method Category=Baseline
2026.05
56.93
32.03
Qwen3.5-GRPO
Method Category=GRPO,...
2026.05
54.1
36.71
Qwen3.5-SFT
Method Category=SFT, B...
2026.05
50.8
35.16
Feedback
Search any
task
Search any
task