[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag--ai":3},{"tag":4,"articles":11,"peer_article_count":12},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":10},"b74ac659-ec6b-4561-b982-a6fd85bac1d8","多代理AI","-ai",20,"多代理AI指的是由多個模型、代理與工具協作完成任務的系統，常見於程式開發、自動化工作流、檢索與長上下文推理。這類架構的重點不只在單次生成，而是規劃、分工、執行與回饋，特別影響企業內部部署、工具調用與成本控制。","Multi-agent AI refers to systems where several models, agents, and tools coordinate to complete work such as coding, workflow automation, retrieval, and long-context reasoning. The key issues are planning, delegation, execution, and feedback, especially for enterprise deployment, tool use, and cost control.",[],19]