ResearchTasksVision-Language CompositionalityVision-Language CompositionalityFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedSugarCrepeDINOv288.06Accuracy20Feb 27, 2026VisMin Hard (test)Qwen2.5-VL-7B + FOCUS84.47Text Score (T)16Jun 23, 2026ColorSwapTTM85.17Accuracy3Apr 27, 2026WhatsUp B (1 × 4 groups)TTM49.94GroupScore2Apr 27, 2026WhatsUp A 1 × 4 groupsTTM56.8GroupScore2Apr 27, 2026