ResearchTasksPrivacy-conscious DelegationPrivacy-conscious DelegationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedPAPILLONBetterTogether(PO, mmGRPO)86.5Composite Score8May 12, 2026PUPA (test)GEPA+Merge96.46Score7Apr 30, 2026