ResearchBenchmarksMulti-question Reasoning on 2Wiki-5QFollow18.5Exact MatchM2R3.947.7211.515.28Apr 10, 2026Evaluation ResultsMethodMethodLinksExact MatchM2R2026.0418.5ReSearch2026.0417SURE2026.0414.8COFT2026.0411.1Naive RAG2026.047.7Naive Generation2026.044.5