List your company

Multi-hop Question Answering on MuSiQue (SR)

31Success Rate (SR)Gemini-3-Flash
511.7518.525.25May 13, 2026

Evaluation results

MethodLinks
Gemini-3-FlashProtocol=Prompting2026.0531
Gemini-2.5-flashProtocol=Prompting2026.0526
MATPOBackbone=Qwen3-8B2026.0517
Qwen3-30B-A3BProtocol=Prompting2026.0516
MATPOBackbone=Qwen3-4B2026.0516
ICRLBackbone=Qwen3-4B2026.0516
ICRLBackbone=Qwen3-8B2026.0516
ScalingInterBackbone=Qwen3-4B2026.0515
Critique-GRPOBackbone=Qwen3-4B2026.0515
GSPOBackbone=Qwen3-8B2026.0515
GRPOBackbone=Qwen3-4B2026.0514
GRPOBackbone=Qwen3-8B2026.0514
ScalingInterBackbone=Qwen3-8B2026.0514
Critique-GRPOBackbone=Qwen3-8B2026.0514
GSPOBackbone=Qwen3-4B2026.0510
Qwen3-4BProtocol=Prompting2026.056
Qwen3-8BProtocol=Prompting2026.056