ResearchTasksTerminal-based task executionFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedTerminal-Bench 2.0GPT-5.3-Codex64.7Accuracy19Apr 29, 2026Terminal Bench 2.1Kimi-K2.667.2Score7Jun 16, 2026