ResearchDatasetsControlled DomainsControlled DomainsFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsVerification behaviorSix controlled domains (Knowledge-work, Research, Multi-tool, Long-memory, Planning, Coding)5.6CBS Rate1Structural verificationControlled Domains Coding (20 held-out tasks)10Final Quality Score1
Verification behaviorSix controlled domains (Knowledge-work, Research, Multi-tool, Long-memory, Planning, Coding)5.6CBS Rate1