Unverified50% confidenceFactExact time
MoA 相比 AutoGen、LangGraph 等框架更轻量，不追求Agent自主规划与工具调用，专注于多视角答案融合
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
Unverified主流Agent框架如LangChain、AutoGen、CrewAI在架构设计上均将能力封装与工具接入视为两个独立关注点72% similarPartially Verified主流Agent框架如LangChain、AutoGen、CrewAI均围绕感知-决策-执行循环构建68% similarUnverified主流Agent框架如LangGraph、AutoGen和CrewAI都在底层实现了类似的循环控制逻辑67% similarUnverifiedMoA 的工作机制是指定几个模型作为顾问（reference）各自独立作答，再由一个更强的模型作为决策者（aggregator）综合判断给出最终结果66% similarUnverifiedAutoGen与LangChain等框架侧重工具链编排不同，更专注于Agent之间的角色分工与消息传递机制66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/524094API
curl https://kongchang.com/api/v1/knowledge/claims/524094MCP
get_claim(id=524094)