ResearchTasksCoding AgentFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSWE-bench Verified single-run, single-patch evaluation protocol (test)Qwen3-Coder51.6Resolution Rate11May 19, 2026Aggregated (RebenchT, CodeCI, Bird)TDScaling34.99Overall Average Score5Feb 26, 2026BirdTDScaling43.83Pass@15Feb 26, 2026CodeCITDScaling39.43Avg@25Feb 26, 2026RebenchTTDScaling33.13OH-p@15Feb 26, 2026
SWE-bench Verified single-run, single-patch evaluation protocol (test)Qwen3-Coder51.6Resolution Rate11May 19, 2026