ResearchTasksSTEM & ReasoningFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedMathVerseGemini 3-Pro91.62Accuracy28Mar 12, 2026MMMU-ProGemini 3-Pro81Accuracy20Jul 7, 2026VisualPuzzleGemini 3-Pro71.48Accuracy6Feb 26, 2026VLMAreBlindERNIE 5.091.38Accuracy5Feb 26, 2026VisuLogicGemini 3-Pro37.6Accuracy5Feb 26, 2026