- dataset: id: harborframework/terminal-bench-science task_id: terminalbenchscience_0_1 value: 3.3 source: url: https://artificialanalysis.ai/evaluations/terminal-bench-science name: "Artificial Analysis Terminal-Bench-Science leaderboard" user: ArtificialAnalysis notes: "MiniMax-M3.1-Flash; mini-swe-agent harness, pass@1 averaged over three repeats."