ResearchTasksMulti-task Model MergingMulti-task Model MergingFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updated14 vision tasksFine-tuned94.3Average Absolute Accuracy42Jun 18, 202620 vision tasksFine-tuned94.7Average Absolute Accuracy40Jun 18, 2026Multi-task Category-level Evaluation Suite (Instruction Following, Math, Multilingual, Safety)METIS93.5Average Score9Jun 16, 2026MMLU, SciQA, and GSM8K (test)NeuroLoRA74.3Average (Individual)4Mar 16, 2026
Multi-task Category-level Evaluation Suite (Instruction Following, Math, Multilingual, Safety)METIS93.5Average Score9Jun 16, 2026