ResearchTasksLong-horizon Research Task ReproductionLong-horizon Research Task ReproductionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedPaperBench Code (dev)Claude 4.5 Sonnet72.22FRE Score7Feb 26, 2026