ResearchDatasetsScience DomainFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsScientific ReasoningScience Domain In-Domain: SampleQA, GPQA(ALL), HLE3.26SampleQA Score18ReasoningScience Domain 20 tasks (test)0.11Total Cost (USD)3Multi-agent task routingScience Domain 1.0 (test)0.59Total Cost2