Unverified70% confidenceFactExact time
Mixtral 8×22B has 141B total parameters, but only approximately 39B are activated per inference.
2
Sources
70%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
VerifiedMixtral 8x7B模型拥有8个专家，每个Token只激活2个，总参数量达到47B，但单次推理的计算量仅相当于一个13B参数的稠密模型79% similarUnverifiedMixtral-8x7B achieves performance close to that of a 70B dense model with a total of 47B parameters, while offering faster inference speed75% similarUnverifiedMixtral 8x7B的实际激活参数量仅为总参数量的约1/473% similarUnverifiedThe actual total parameter count of Mixtral-8x7B is 46.7B, while the naming multiplication of 8x7=56B would mislead hardware planning by about 20%.70% similarVerifiedDeepSeek-V3 has 671B total parameters, but only approximately 37B parameters are activated per inference.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/527879API
curl https://kongchang.com/api/v1/knowledge/claims/527879MCP
get_claim(id=527879)