List your company
ResearchFormerly WizWand

Multi-hop QA on Bamboogle (Accuracy %)

74.9Accuracy (%)IGPO
20.09234.32148.5562.779Mar 18, 2026Mar 20, 2026Mar 23, 2026Mar 25, 2026Mar 28, 2026Mar 30, 2026Apr 2, 2026
Updated 7mo ago

Evaluation results

MethodLinks
IGPOBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0374.9
ProCeedRLBackbone=Qwen3-8B2026.0473.87
DeepResearcher2026.0472.8
DAPO/Search-R1Backbone=Qwen3-8B2026.0470.83
GiGPOBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0368.9
GiGPO2026.0468.9
Rewinding MoreBackbone=Qwen3-8B, Type=ablation2026.0465.87
RFTBackbone=Qwen3-8B2026.0464.27
Qwen3-8B-v3-SFTBackbone=Qwen3-8B2026.0462.13
ReAct PromptingBackbone=Qwen3-8B2026.0450.93
EvolveRBase Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0342
SkillRLBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0340.3
Search-R1Base Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0340.1
ZeroSearchBase Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0335.2
SLEA-RLBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0333.2
R1-InstructBase Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0327.5
Search-o1Base Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0327
PPOBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0326.2
GSPOBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0325.4
GRPOBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0325
RLOOBase Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0324.8
Reinforce++Base Model=Qwen2.5-7B-Instruct, Method Category=RL Methods2026.0323.7
RAGBase Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0323.2
CoTBase Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0322.6
Qwen2.5Base Model=Qwen2.5-7B-Instruct, Method Category=Prompting2026.0322.2