ResearchTasksEmbodied vision-language reasoningFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedHuman evaluations HQwen3-VL-235B-A22B-Thinking0.879Score13Feb 26, 2026Original benchmark BQwen3-VL-235B-A22B-Thinking61.26Score13Feb 26, 2026A2Eval B~Qwen3-VL-235B-A22B-Thinking65.97Score13Feb 26, 2026