Unverified50% confidenceFactExact time
Tencent open-sourced the Hy-Memory-30B-A3B model, a 30B-parameter mixture-of-experts (MoE) model that activates about 3B parameters per token.
1
Sources
50%
Confidence
Long-term
Relevance
10/17/2026
First Seen
Sources
Related Entities
Related Claims
Unverified混元3是混合专家模型（MoE），总参数量为295B，激活参数约21B76% similarVerifiedHy3 adopts a Mixture of Experts (MoE) architecture with a total parameter count of 295 billion (295B), activating only 21 billion (21B) parameters per inference.74% similarVerified35B-A3B is a MoE architecture naming convention, where 35B refers to a total parameter count of approximately 35 billion, and the A in A3B stands for activated parameters of approximately 3 billion.72% similarUnverifiedCommunity testing (34 English + 34 Russian conversations) shows that all quantized versions of Hy-Memory-30B-A3B can output valid JSON (68/68), but only 23/34 fully succeed.72% similarUnverifiedgemma-4-26b-a4b 模型名称中的 'a4b' 表示该模型采用了混合专家（MoE）架构，激活参数量约为4B72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1054158API
curl https://kongchang.com/api/v1/knowledge/claims/1054158MCP
get_claim(id=1054158)