ResearchDatasetsMJ-BenchFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsPairwise preference predictionMJ-Bench v1 (test)69.8Alignment Score7Safety JudgeMJ-Bench82.23Macro-F13Image-text AlignmentMJ-Bench60.59Macro F13