ResearchTasksJailbreak Success RateFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast Updatedmed-safetyGPT-4o7.6Success Rate (Raw)3Jun 4, 2026flamesDeepSeek-R138.6Success Rate (Raw)3Jun 4, 2026advbenchGPT-4o3.8Raw Jailbreak Success Rate3Jun 4, 2026