Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-Hop Question Answering on Bamboogle (test) (Pass@3)
Loading...
59.2
Pass@3
PivoARL
40.064
45.032
50
54.968
Jul 4, 2026
Pass@3
Updated 18d ago
Evaluation Results
Method
Method
Links
Pass@3
PivoARL
Base Model=Qwen3-4B, S...
2026.07
59.2
MetaRL
Base Model=Qwen3-4B, S...
2026.07
53.6
StepSearch
Base Model=Qwen3-4B, S...
2026.07
51.2
GiGPO
Base Model=Qwen3-4B, S...
2026.07
48
GRPO
Base Model=Qwen3-4B, S...
2026.07
47.2
Search-R1
Base Model=Qwen3-4B, S...
2026.07
46.4
ZeroSearch
Base Model=Qwen3-4B, S...
2026.07
46.4
GSPO
Base Model=Qwen3-4B, S...
2026.07
40.8
Feedback
Search any
task
Search any
task