Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-turn Agentic Tool-use on WebShop
Loading...
79.6
Success Rate
HERO
14.912
31.706
48.5
65.294
Jun 10, 2026
Success Rate
Turns
Updated 1mo ago
Evaluation Results
Method
Method
Links
Success Rate
Turns
HERO
Model=Qwen3-30B-A3B-In...
2026.06
79.6
7
GRPO
Model=Qwen3-30B-A3B-In...
2026.06
76.1
7.3
HERO
Model=Qwen3-4B-Instruct
2026.06
68.9
8.1
GRPO
Model=Qwen3-4B-Instruct
2026.06
65.2
8.1
Environment Feedback Only
Model=Qwen3-30B-A3B-In...
2026.06
44.3
9.3
Environment Feedback Only
Model=Qwen3-4B-Instruct
2026.06
37.5
10.4
Base
Model=Qwen3-30B-A3B-In...
2026.06
32.7
8.5
Full-Demo Privileged Teacher
Model=Qwen3-30B-A3B-In...
2026.06
25.9
9.9
Base
Model=Qwen3-4B-Instruct
2026.06
21.8
8.6
Full-Demo Privileged Teacher
Model=Qwen3-4B-Instruct
2026.06
17.4
10.6
Feedback
Search any
task
Search any
task