ResearchTasksMulti-task Model Merging EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedCross-task Utility Aggregate 8 Vision DatasetsUnprotected Baseline67.23Average Accuracy9Jun 30, 2026Multi-task Category-level Evaluation Suite Qwen-3-4BMETIS1.18Average Score9Jun 16, 2026