ResearchTasksAlignment & Instruction FollowingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedScale AI Multi-ChallengeQwen3.5-122B-A10B61.5Pass@17Apr 15, 2026ArenaHard Hard Prompt v2Nemotron-Cascade-2 30B-A3B88.2Pass@14Mar 20, 2026ArenaHard Creative Writing v2Nemotron-Cascade-2 30B-A3B78.7Pass@13Mar 20, 2026ArenaHard Avg. v2Nemotron-Cascade-2 30B-A3B83.5Pass@13Mar 20, 2026