Unverified50% confidenceFactExact time
The experiment uses the same Qwen 3.6-35B-A3B model, and the two versions with and without MTP have exactly the same number of parameters (35 billion parameters).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/28/2026
First Seen
Valid until: 1/26/2027
Sources
Related Entities
Related Claims
UnverifiedThe model has a core of 125B parameters, plus an n-gram lookup table with 51B parameters and a multi-token prediction (MTP) draft layer with 4B parameters, totaling approximately 180B.74% similarUnverified研究表明当模型参数规模达到3B以上时，1.58-bit模型在多项基准测试中表现可接近甚至匹配同等规模的全精度模型67% 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.65% similarUnverified2.4万亿参数模型在BF16精度下考虑KV Cache和激活值后实际约需1.4-1.5TB显存65% similarUnverifiedOn the 64G version, a test run successfully ran Qwen 3.8's 27-billion-parameter model.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1106415API
curl https://kongchang.com/api/v1/knowledge/claims/1106415MCP
get_claim(id=1106415)