Verified65% confidenceFactExact time
Mistral Large 4 is a hybrid instruction-and-reasoning MoE model with a total of 1 trillion parameters, activating only approximately 49 billion parameters during inference.
3
Sources
65%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLe Chonk is a hybrid instruction and reasoning MoE model with total parameters at the trillion level and only 49 billion activated parameters.74% similarUnverifiedMoE架构下模型总参数量远大于单次推理实际激活的参数量，Qwen3.8每次推理激活参数可能仅数百亿级别71% similarUnverifiedMixture-of-Experts (MoE) architectures such as Google's Switch Transformer and Mistral AI's Mixtral have enormous total parameter counts but activate only a small fraction of them during each inference.70% similarVerifiedModels with extremely large parameter scales typically adopt the MoE (Mixture of Experts) architecture, which activates only a portion of parameters per inference rather than computing the full set.67% similarVerified1.6万亿参数级别的巨型模型通常采用混合专家架构（MoE），每次推理只激活一小部分专家子网络67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/982897API
curl https://kongchang.com/api/v1/knowledge/claims/982897MCP
get_claim(id=982897)