ResearchTasksTool-Need PredictionFor AI assistantsMarkdownContent indexTool-Need PredictionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedNemotron held-out core (test)GPT-OSS 20B75.3Tool-Need Accuracy2May 11, 2026