ResearchTasksDisallowed Content EvaluationDisallowed Content EvaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedXSTesto1-preview97.6Non-Overrefusal Rate4May 1, 2026WildChato198Not Unsafe Rate4May 1, 2026Challenging Refusal Evaluationo1-preview93.4Non-unsafe Rate4May 1, 2026Standard Disallowed Content Evaluationgpt-5-thinking-mini99.6Hate (Aggregate)4May 1, 2026