DeepSeekLaunchAI ModelsSeptember 18, 2026

DeepSeek releases V4.1-Flash with novel YOCO decoder-decoder architecture

DeepSeek-V4.1-Flash ships with a 552B backbone, native image and text input, and up to a 1M-token context window. The YOCO architecture targets GPU memory and prefill latency; DeepSeek says it will work with the open-source community on inference support.

How this story unfolded

8 days · 11 reports · 42 community posts · 53 of 55 shown

  1. Sep 10
  2. Sep 11
  3. Sep 12
  4. Sep 13
  5. Sep 15
  6. Sep 16
  7. Sep 17
  8. Sep 18

More stories today

Open the live feed