@_akhaliq· @_akhaliq · X·2026-09-01 03:43· 2026-09-01AI 评分25阅读原文 AI 导读LoopArena 将模型作为循环工程的运行时控制器进行基准测试 paper: https://t.co/GEzWnPfqjr https://t.co/1jqniijCDn正文LoopArenaBenchmarking Models as Runtime Controllers for Loop Engineeringpaper: https://t.co/GEzWnPfqjr https://t.co/1jqniijCDn来源：@_akhaliq · x.com#论文/研究#评测/基准#Agent