Unverified50% confidenceFactExact time
Ling-3.1-flash activates only approximately 25 billion (~25B) parameters per token.
1
Sources
50%
Confidence
Long-term
Relevance
9/30/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe approximately 20:1 ratio between Qwen3 Next Flash's 125 billion total parameters and 6 billion activated parameters represents a fairly aggressive sparsity design.65% similarUnverifiedOn a task with one million input tokens, thirty thousand output tokens, and a 99% cache hit rate, MiMo-V2.6 Flash costs less than half of Qwen 3.8 Flash.64% similarUnverifiedWhisper-Medium 拥有约 3 亿参数的编码器部分63% similarUnverifiedV4-Flash has a total parameter count of 284B and only 13B active parameters.59% similarUnverifiedQwen 3.8's 27-billion-parameter model is approximately 15.3 GB after 4-bit quantization, but actual memory usage can surge to over 17 GB after a conversation of several thousand characters, meaning 16 GB of VRAM is insufficient and 24 GB is needed for comfortable use.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/962913API
curl https://kongchang.com/api/v1/knowledge/claims/962913MCP
get_claim(id=962913)