List your company

Multi-hop Question Answering on Bamboogle (SR)

72SRGemini-3-Flash
24.1636.584961.42May 13, 2026

Evaluation results

MethodLinks
Gemini-3-FlashProtocol=Prompting2026.0572
Gemini-2.5-flashProtocol=Prompting2026.0559.5
Qwen3-30B-A3BProtocol=Prompting2026.0544
GSPOBackbone=Qwen3-8B2026.0542
MATPOBackbone=Qwen3-8B2026.0541
MATPOBackbone=Qwen3-4B2026.0540
ICRLBackbone=Qwen3-8B2026.0540
Critique-GRPOBackbone=Qwen3-4B2026.0538.5
ICRLBackbone=Qwen3-4B2026.0538
GRPOBackbone=Qwen3-8B2026.0538
ScalingInterBackbone=Qwen3-8B2026.0538
Critique-GRPOBackbone=Qwen3-8B2026.0538
GSPOBackbone=Qwen3-4B2026.0534
ScalingInterBackbone=Qwen3-4B2026.0532
Qwen3-8BProtocol=Prompting2026.0530
GRPOBackbone=Qwen3-4B2026.0530
Qwen3-4BProtocol=Prompting2026.0526