Unverified50% confidenceFactExact time
千问三235B MoE总参数量2350亿，但每次推理实际激活的参数约22B
1
Sources
50%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
LangChain+MCP实战：大模型选型与本地部署避坑指南
bilibili码士集团-马小琪7/5/2026
Related Claims
VerifiedQwen3-235B has a total parameter count of 235 billion, adopts a MoE architecture, and activates approximately 22B parameters during each forward inference pass.84% similarUnverifiedMoE模型中'35B'指总参数量，每个token只激活少量Expert（通常2-4个），实际计算量可能仅相当于6-7B的Dense模型77% similarUnverifiedHy3 adopts a Mixture of Experts (MoE) architecture with a total parameter count of 295 billion (295B), activating only 21 billion (21B) parameters per inference.77% similarUnverified千问3.6的35B MOE模型每次只激活其中3B的参数77% similarUnverifiedBERT的参数规模为3.4亿，GPT-3的参数规模为1750亿74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/131916API
curl https://kongchang.com/api/v1/knowledge/claims/131916MCP
get_claim(id=131916)