ResearchTasksSpatial Reasoning (Multi-Image)FollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedMMSI-BenchHuman97.2Accuracy44May 27, 2026SPAR-BenchHuman67.3Accuracy31Jun 12, 2026ERQAGemini2.5-Pro51.02Accuracy23May 25, 2026