ResearchTasksMLLM Safety EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedOOD-MMSafe Constitution Mode 1.0 (test)Gemma-3-12B1.84RA13Mar 11, 2026OOD-MMSafe Malicious Mode 1.0 (test)o4-Mini1.97RA Score13Mar 11, 2026OOD-MMSafe Standard Mode 1.0 (test)Gemini-3-Pro1.35RA13Mar 11, 2026