List your company

Multi-Hop Question Answering on Bamboogle (EM, F1)

78.4EMSynPlanResearch-R1
2.646422.313241.9861.6468Jan 23, 2026Jan 30, 2026Feb 7, 2026Feb 14, 2026Feb 22, 2026Mar 1, 2026Mar 9, 2026

Evaluation results

MethodLinks
SynPlanResearch-R1Model Backbone=Qwen3-8B2026.0378.4—
SimpleDeepSearcherModel Backbone=Qwen3-8B2026.0376.5—
SynPlanResearch-R1Model Backbone=Qwen3-4B2026.0375.3—
Search-R1Model Backbone=Qwen3-8B2026.0374.5—
SimpleDeepSearcherModel Backbone=Qwen3-4B2026.0373.8—
Search-R1Model Backbone=Qwen3-4B2026.0369.8—
Rejection SamplingModel Backbone=Qwen3-8B2026.0365.2—
Rejection SamplingModel Backbone=Qwen3-4B2026.0363.7—
Search-o1Model Backbone=Qwen3-8B2026.0354.6—
Search-o1Model Backbone=Qwen3-4B2026.0345.3—
RAGModel Backbone=Qwen3-8B2026.0332.3—
GraphAnchorBackbone=Qwen3-32B2026.0132.2342.73
HierSearchBackbone=Qwen2.5-7B-Instruct2026.0132—
O2-SearcherBackbone=Qwen2.5-7B-Instruct2026.0130.89—
Dep-SearchBackbone=Qwen2.5-7B-Instruct2026.0130.56—
Search-O1Backbone=Qwen2.5-7B-Instruct2026.0130.4—
RA-ISFBackbone=Qwen2.5-7B-Instruct2026.0128.9—
R1-SearcherBackbone=Qwen2.5-7B-Instruct2026.0128.89—
GraphAnchorBackbone=Llama3.1-8B-Instruct2026.0128.6936.55
IRCoTBackbone=Qwen2.5-7B-Instruct2026.0127.78—
DeepNoteBackbone=Llama3.1-8B-Instruct2026.0127.237.46
GraphAnchorBackbone=Qwen2.5-14B-Instruct2026.0127.235.22
RAGModel Backbone=Qwen3-4B2026.0327.2—
Search-R1Backbone=Qwen2.5-7B-Instruct2026.0126.39—
ZeroSearchBackbone=Qwen2.5-7B-Instruct2026.0125.56—
GraphAnchorBackbone=Qwen2.5-7B-Instruct2026.0124.834.25
Dep-SearchBackbone=Qwen2.5-3B-Instruct2026.0124—
DeepNoteBackbone=Qwen2.5-7B-Instruct2026.012432.66
IRCoTBackbone=Llama3.1-8B-Instruct2026.0123.232.46
HierSearchBackbone=Qwen2.5-3B-Instruct2026.0122.56—
Search-O1Backbone=Qwen2.5-3B-Instruct2026.0122.4—
O2-SearcherBackbone=Qwen2.5-3B-Instruct2026.0121.44—
Search-R1Backbone=Qwen2.5-7B-Instruct2026.012025.13
Iter-RetGenBackbone=Llama3.1-8B-Instruct2026.012026.09
R1-SearcherBackbone=Qwen2.5-3B-Instruct2026.0118.5—
Direct InferenceModel Backbone=Qwen3-8B2026.0317.5—
Directly InferenceBackbone=Qwen2.5-3B-Instruct2026.0116.8—
Iter-RetGenBackbone=Qwen2.5-7B-Instruct2026.0116.824.82
Vanilla RAGBackbone=Qwen3-32B2026.0116.825.91
ZeroSearchBackbone=Qwen2.5-3B-Instruct2026.0116.67—
Vanilla RAGBackbone=Llama3.1-8B-Instruct2026.011624.4
Vanilla RAGBackbone=Qwen2.5-7B-Instruct2026.0115.2—
Vanilla RAGBackbone=Qwen2.5-14B-Instruct2026.0115.226.74
Vanilla RAGBackbone=Qwen2.5-3B-Instruct2026.0114.4—
Directly InferenceBackbone=Qwen2.5-7B-Instruct2026.0114.4—
Vanilla RAGBackbone=Qwen2.5-7B-Instruct2026.0113.621.62
IRCoTBackbone=Qwen2.5-7B-Instruct2026.0112.821.15
Search-R1Backbone=Qwen2.5-3B-Instruct2026.0112.5—
Direct InferenceModel Backbone=Qwen3-4B2026.0311.1—
RA-ISFBackbone=Qwen2.5-3B-Instruct2026.016.2—
IRCoTBackbone=Qwen2.5-3B-Instruct2026.015.56—