Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
DeepSearch Question Answering on Xbench
Loading...
34
Average Score
APPO
6.96
13.98
21
28.02
Jun 10, 2026
Average Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Average Score
APPO
Backbone=Qwen3-14B, Me...
2026.06
34
DeepSeek-R1-671B
Backbone=DeepSeek-R1-6...
2026.06
32.7
ARPO
Backbone=Qwen3-14B, Me...
2026.06
32
APPO
Backbone=Qwen3-8B, Met...
2026.06
28
GRPO
Backbone=Qwen3-14B, Me...
2026.06
27
ARPO
Backbone=Qwen3-8B, Met...
2026.06
25
WebThinker
Backbone=Qwen3-14B, Me...
2026.06
23
Search-o1
Backbone=Qwen3-14B, Me...
2026.06
21
GRPO
Backbone=Qwen3-8B, Met...
2026.06
20
ReAct
Backbone=Qwen3-14B, Me...
2026.06
20
GPT-4o
Backbone=GPT-4o, Metho...
2026.06
18
ReAct
Backbone=Qwen3-8B, Met...
2026.06
16
Vanilla RAG
Backbone=Qwen3-14B, Me...
2026.06
15
Qwen3-32B-thinking
Backbone=Qwen3-32B, Me...
2026.06
14
Qwen3-14B
Backbone=Qwen3-14B, Me...
2026.06
14
WebThinker
Backbone=Qwen3-8B, Met...
2026.06
13
QwQ-32B
Backbone=QwQ-32B, Meth...
2026.06
10.7
DeepSeek-R1-32B
Backbone=DeepSeek-R1-3...
2026.06
10
Search-o1
Backbone=Qwen3-8B, Met...
2026.06
10
Qwen3-8B
Backbone=Qwen3-8B, Met...
2026.06
9
Vanilla RAG
Backbone=Qwen3-8B, Met...
2026.06
8
Feedback
Search any
task
Search any
task
