DeepSeek-v4.1 Flash: Pushing the Limits of KV Cache Compression

(zartbot.github.io)

37 points by mfiguiere1 hour ago

2 comments

  • arikrahman18 minutes ago
    I am very impressed with the KV Cache Compression work as well as the prefix cacheing making queries converge on practically free.
  • smy200115 minutes ago
    Should we flag this since It's AI generated?