ResearchBenchmarksDeep Research Report Generation and Action Adherence on MYSQA datasetFollow91.4Analysis CoverageMYSQA71.22476.46281.786.938Mar 17, 2026Evaluation ResultsMethodMethodLinksAnalysis CoverageAnalysis PrecisionCitation PrecisionCitation RecallAction AdherenceMYSQA2026.0391.489.991.881.483.2o3 DRevaluation samples=10...evaluation samples=10 examples2026.0389.190.279.256.793.8SQA2026.0388.989.190.576.981.3Sonar DR2026.038182.964.346.375OPENSC.2026.0377.297.482.560.482.5STORM2026.037292.273.364.774.4