ResearchTasksReward AggregationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedAverage HealthBench, WritingBench, PLawBenchRuscaRL + GEAR75.7Average Score18Jun 4, 2026PLawBenchRuscaRL + GEAR78.4Score18Jun 4, 2026WritingBenchRuscaRL + GEAR85Score18Jun 4, 2026HealthBench 500RuscaRL + GEAR65.7Score18Jun 4, 2026