Expired60% confidenceFactExact time
Liang Wenfeng stated that the largest model in the world currently has approximately 800B activated parameters, while domestic models are still at a scale of tens of billions of activated parameters, an order of magnitude behind.
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
7/25/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/23/2026(expired)
Sources
梁文锋深度自白：DeepSeek的克制哲学、开源策略与AGI路线图
Related Claims
UnverifiedMistral Large 4 adopts a Mixture of Experts (MoE) architecture with a total parameter count of approximately 1 trillion, activating approximately 50 billion parameters per token63% similarUnverified幻方量化管理规模曾超过千亿人民币60% similarUnverifiedThe parameter scale of large language models has leaped from the hundreds of millions to the trillions.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/613783API
curl https://kongchang.com/api/v1/knowledge/claims/613783MCP
get_claim(id=613783)