[{"data":1,"prerenderedAt":96},["ShallowReactive",2],{"optimization":3},{"headers":4,"rows":35},[5,16],[6,9,12],{"en":7,"zh":7,"rowspan":8},"Agent",2,{"en":10,"zh":11},"Correctness","正确性",{"en":13,"zh":14,"colspan":15},"Resource use","资源使用",5,[17,20,23,26,29,32],{"en":18,"zh":19},"B1 · Success rate","B1 · 成功率",{"en":21,"zh":22},"B3 · Token usage\u003Cbr>(k\u002Ftask)","B3 · Token 用量\u003Cbr>（千\u002F任务）",{"en":24,"zh":25},"B4 · Execution time\u003Cbr>(min\u002Ftask)","B4 · 执行时间\u003Cbr>（分钟\u002F任务）",{"en":27,"zh":28},"B5 · Interaction turns\u003Cbr>(\u002Ftask)","B5 · 交互轮次\u003Cbr>（\u002F任务）",{"en":30,"zh":31},"B6 · Cost\u003Cbr>(USD\u002Ftask)","B6 · 成本\u003Cbr>（美元\u002F任务）",{"en":33,"zh":34},"B7 · Cost\u003Cbr>(USD\u002Fpass)","B7 · 成本\u003Cbr>（美元\u002F成功任务）",[36,51,66,81],[37,39,41,43,45,47,49],{"en":38,"zh":38},"AutoBA",{"en":40,"zh":40},"0.34",{"en":42,"zh":42},"72.38",{"en":44,"zh":44},"34.41",{"en":46,"zh":46},"27.10",{"en":48,"zh":48},"0.247",{"en":50,"zh":50},"0.726",[52,54,56,58,60,62,64],{"en":53,"zh":53},"AutoBA*",{"en":55,"zh":55},"0.55",{"en":57,"zh":57},"188.13",{"en":59,"zh":59},"29.04",{"en":61,"zh":61},"47.32",{"en":63,"zh":63},"0.680",{"en":65,"zh":65},"1.236",[67,69,71,73,75,77,79],{"en":68,"zh":68},"BioMaster",{"en":70,"zh":70},"0.05",{"en":72,"zh":72},"68.92",{"en":74,"zh":74},"14.81",{"en":76,"zh":76},"7.77",{"en":78,"zh":78},"0.267",{"en":80,"zh":80},"5.340",[82,84,86,88,90,92,94],{"en":83,"zh":83},"BioMaster*",{"en":85,"zh":85},"0.32",{"en":87,"zh":87},"160.53",{"en":89,"zh":89},"20.46",{"en":91,"zh":91},"19.23",{"en":93,"zh":93},"0.573",{"en":95,"zh":95},"1.790",1791659693485]