ResearchTasksLong-Context Stability & RetrievalFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedLRA-L2Halo42.3Stability Rate (SR)6Feb 26, 2026InfBenchHalo55.8Stability Rate (SR)6Feb 26, 2026LongBenchHalo62.5SR (Stability)6Feb 26, 2026RULER Tier 2Halo76.4SR6Feb 26, 2026