ResearchTasksReward Model TransferFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedUltraFeedback (UF)ANCHOR7.93AOG16May 26, 2026Anthropic Helpful (AH)SEAM9.95AOG16May 26, 2026HelpSteer3 (H3) v1 (test)ANCHOR8.96AOG16May 26, 2026RAILANCHOR4.62AOG8May 26, 2026PKU-SafeRLHFANCHOR2.46AOG8May 26, 2026Anthropic Harmless (AHar)ANCHOR4.62AOG8May 26, 2026