ResearchDatasetsBALROGFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsLong-horizon game playingBALROG NetHack6.8Progression Rate12Long-horizon game playingBALROG MiniHack40Progression Rate12Agentic reasoningBALROG31.5Accuracy8