这才是 Agent 最终该有的形态。
用户不应该关心底层是 GPT、Claude、Grok 还是别的模型。
对一个任务自动拆解、路由到最合适的模型和工具,失败就 fallback,最后只对结果负责。
模型会不断换,真正长期积累的是 context、routing、eval 和 execution。
model is plumbing, outcome is the product.
Important note regarding Grok @Bot:
Going forward, @SpaceX will use the best back end model for any given task, including Claude Opus 5.5, MidJourney, Suno and other leading APIs.
Whatever is most likely to give you the best outcome.