ResearchTasksScientific CodingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSciCodegpt-oss-120B0.428Accuracy18Jun 23, 2026ScienceAgentBench deterministic-eval 2024Deep negative knowledge retry47.4Pass Rate5Jun 23, 2026