Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Knowledge-Intensive Reasoning on WebWalker (accuracy)
Loading...
33.5
WebWalker Accuracy
APPO
0.74
9.245
17.75
26.255
Jun 10, 2026
WebWalker Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
WebWalker Accuracy
APPO
Backbone=Qwen2.5-7B-In...
2026.06
33.5
APPO
Backbone=Llama3.1-8B-I...
2026.06
32
GIGPO
Backbone=Llama3.1-8B-I...
2026.06
31.5
GPPO
Backbone=Qwen2.5-7B-In...
2026.06
31
ARPO
Backbone=Llama3.1-8B-I...
2026.06
30.5
GIGPO
Backbone=Qwen2.5-7B-In...
2026.06
30.5
CISPO
Backbone=Qwen2.5-7B-In...
2026.06
29
Reinforce++
Backbone=Llama3.1-8B-I...
2026.06
27.5
GPPO
Backbone=Llama3.1-8B-I...
2026.06
27.5
GRPO
Backbone=Llama3.1-8B-I...
2026.06
26.5
CISPO
Backbone=Llama3.1-8B-I...
2026.06
26
Reinforce++
Backbone=Qwen2.5-7B-In...
2026.06
26
ARPO
Backbone=Qwen2.5-7B-In...
2026.06
26
DAPO
Backbone=Llama3.1-8B-I...
2026.06
25.5
DAPO
Backbone=Qwen2.5-7B-In...
2026.06
24
GRPO
Backbone=Qwen2.5-7B-In...
2026.06
22
TIR Prompting
Backbone=Qwen2.5-7B-In...
2026.06
15.5
TIR Prompting
Backbone=Llama3.1-8B-I...
2026.06
15
Zero-shot
Backbone=Llama3.1-8B-I...
2026.06
3
Zero-shot
Backbone=Qwen2.5-7B-In...
2026.06
2
Feedback
Search any
task
Search any
task