Unverified50% confidenceFactExact time
Cognition（Devin 的开发团队）提出了 Agentic MapReduce 架构，通过多 Agent 协同实现对完整代码库的规模化推理
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
Agentic MapReduce：AI Agent 实现全代码库分布式推理的关键架构
rss7/1/2026
Related Claims
UnverifiedAgentic MapReduce 架构中 Map 阶段的每个 worker 是具备自主决策能力的智能体，Reduce 阶段需要 Agent 对分散的推理结果进行语义层面的整合与再推理81% similarUnverifiedAgentic MapReduce 提供可横向扩展的路径，理论上只需增加 Agent 数量即可处理更大规模代码库，无需等待底层模型上下文窗口突破74% similarUnverifiedAccording to the Kunpeng team's analysis, a complete agent architecture consists of large model inference (the brain), the agent core (Gateway), memory system, action module, and tool calling.73% similarUnverifiedAgent的Planning模块通常基于思维链CoT或ReAct等推理框架实现72% similarUnverifiedAgent 能力经历了从基于 ReAct 范式的单步推理到以 LangGraph、AutoGen、Claude Code 为代表的多智能体协作框架的转变72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/459307API
curl https://kongchang.com/api/v1/knowledge/claims/459307MCP
get_claim(id=459307)