ResearchTasksMaze TraversalMaze TraversalFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedMaze Traversal 10x10 mazes (10 mazes x 10 runs)gpt-4.110Correct Count (out of 10)12Apr 3, 2026