StorySoftware
DeepSeek-V4.1-Flash goes live on the API; V4-Pro requests move to it from 14 September
DeepSeek put its V4.1-Flash model live on its API on 10 September 2026, with native multimodal support. From 14 September, requests for the older V4-Pro model are routed to V4.1-Flash at Flash rates until a V4.1-Pro launches.
- V4.1-Flash is a 552-billion-parameter mixture-of-experts model that uses 8 billion parameters per input token and 16 billion per output token.
- DeepSeek says its KV cache needs about a quarter of the high-bandwidth memory and an eighth of the SSD storage of the previous generation.
- New API pricing took effect at 04:00 UTC on 10 September 2026; off-peak rates are half of peak rates.
- The names deepseek-v4-flash and deepseek-v4-flash-vision-exp are retired and temporarily route to V4.1-Flash.
- Open weights are linked on Hugging Face; the announcement does not state the licence.