ResearchTasksCorrectness EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedPPE Correctness (test)Skywork-Reward-V2-Llama-3.1-8B-40M87.2Score24Mar 4, 2026