ResearchDatasetsOckBenchFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsMultimodal ReasoningOckBench 20-eval suite (test)71.9Accuracy14Vision-Language EvaluationOckBench47Tokens per Correct Answer14