ResearchTasksScientific Method Comparative Reasoning ForecastingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedIndependently Constructed 2025 (test)Reason-SFT-DrGRPO67.49Accuracy11May 22, 2026