List your company

Question Answering on MuSiQue (Exact Match (%))

22.6Exact MatchStepSearch
3.5688.50913.4518.391May 13, 2026

Evaluation results

MethodLinks
StepSearchBackbone=Qwen2.5-7B-Instruct2026.0522.6
ReSearchBackbone=Qwen2.5-7B-Instruct2026.0522.3
PyRAG-RLBackbone=LLaMA-3.1-8B-Instruct2026.0522.1
PyRAG-RLBackbone=Qwen2.5-7B-Instruct2026.0520.7
PyRAG-RLBackbone=Qwen3-4B-Instruct2026.0518.6
ZEROSEARCHBackbone=Qwen2.5-7B-Instruct2026.0518.4
Search-R1Backbone=Qwen2.5-7B-Instruct2026.0514.6
RAG-RLBackbone=LLaMA-3.1-8B-Instruct2026.0511.4
RAG-RLBackbone=Qwen2.5-7B-Instruct2026.059.6
RAG-RLBackbone=Qwen3-4B-Instruct2026.058.8
RAG-SFTBackbone=LLaMA-3.1-8B-Instruct2026.058.5
RAG-SFTBackbone=Qwen2.5-7B-Instruct2026.056.8
Vanilla RAGBackbone=LLaMA-3.1-8B-Instruct2026.056.3
RAG-SFTBackbone=Qwen3-4B-Instruct2026.056.2
Vanilla RAGBackbone=Qwen2.5-7B-Instruct2026.054.7
Vanilla RAGBackbone=Qwen3-4B-Instruct2026.054.3