ResearchTasksVisual consistency assessmentFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedEditReward-Bench 2025 (test)Gemini 3 Pro80.88Subject Addition Accuracy6Mar 31, 2026VCReward-BenchGemini 3 Pro93.59Subject Addition Consistency6Mar 31, 2026