ResearchDatasetsJEEBenchJEEBenchFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsMath reasoningJEEBench74.4Accuracy82Scientific ReasoningJEEBench57.34Mean Accuracy22Downstream AccuracyJEEBench44.8Accuracy12