Unverified50% confidenceFactExact time
MoE架构最早可追溯至1991年学术研究，GPT-4、Mixtral 8x7B、Gemini 1.5均采用类似设计
1
Sources
50%
Confidence
Long-term
Relevance
7/25/2026
First Seen
Sources
Redis之父如何将2840亿参数DeepSeek塞进普通笔记本
Related Claims
Unverified专家混合（MoE）架构最早由Jacobs等人于1991年提出，后被Mixtral、GPT-4等现代大模型借鉴80% similarVerifiedGPT-4和Mixtral等主流大模型也采用了MoE架构78% similarUnverifiedGPT-4 is reportedly built on an MoE design, and Mixtral 8x7B is a representative open-source implementation.78% similarVerifiedDeepSeek-V3, Google's Gemini, and GPT-4 (reportedly) all adopt a similar MoE architecture.73% similarUnverifiedTesters assessed that GPT-5.6 has reached an aesthetic standard in front-end design comparable to that of Kimi and Gemini.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/615580API
curl https://kongchang.com/api/v1/knowledge/claims/615580MCP
get_claim(id=615580)