ResearchTasksDeep Research EvaluationDeep Research EvaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedDeep Research Bench first training epoch (step 600)Tournament-GRPO52.09Readability17May 27, 2026Deep Research Bench (step 1100)Tournament-GRPO53.81Readability16May 27, 2026AggregateSmolagents Open DR63.97WQ3Feb 26, 2026ResearchRubricsSmolagents Open DR66.6WQ Score3Feb 26, 2026LiveResearchBenchSmolagents Open DR61.71WQ3Feb 26, 2026DeepResearchBenchTongyi Deep Research63.95WQ3Feb 26, 2026