ResearchTasksOperational safety evaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedOFFTOPICEVAL Hindi 1.0 (test)GPT-OSS (120B)98.86AR_ID12Mar 16, 2026OFFTOPICEVAL Chinese 1.0 (test)GPT-OSS (20B)99.32Acceptance Rate (ID)12Mar 16, 2026OFFTOPICEVAL English 1.0 (test)Llama-3.3 (70B)99.62AR_ID12Mar 16, 2026