ResearchBenchmarksMulti-question Reasoning on 2Wiki-3QFollow35.8Exact Match (EM)M2R16.14421.24726.3531.453Apr 10, 2026Evaluation ResultsMethodMethodLinksExact Match (EM)M2R2026.0435.8ReSearch2026.0433.9SURE2026.0431.5COFT2026.0427.2Naive Generation2026.0417.5Naive RAG2026.0416.9