AI 导读
Artificial Analysis 对比了 Claude Opus 5.5 的智能指数与单任务成本，五个 effort 档位中有四个（max、xhigh、high、medium）位于帕累托前沿。这些档位在得分 50 以上的模型中成本更低或表现更优。
正文
Four of Claude Opus 5.5’s five effort levels sit on the Intelligence vs Cost per Task frontier: Opus 5.5 max, xhigh, high, and medium all sit on the Pareto frontier, costing less or outperforming other models scoring 50+ https://t.co/Ojhw7VyeUF
来源：@ArtificialAnlys · x.com