ResearchDatasetsAIME, MATH, GSM8K, HumanEval, MMLUFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsGeneral-domain capability retentionAIME-2024, MATH-500, GSM8K, HumanEval, MMLU62.9AIME-2024 Accuracy21
General-domain capability retentionAIME-2024, MATH-500, GSM8K, HumanEval, MMLU62.9AIME-2024 Accuracy21