Text models · Open weights

DeepSeek releases V4.1-Flash and begins phasing out V4-Pro

DeepSeek's V4.1-Flash is a 552B-parameter model with native multimodal support and lower prices, and V4-Pro requests will route to it.

Published September 10, 2026Source DeepSeek API docs ↗

DeepSeek announced V4.1-Flash on September 10, 2026. It is the smallest model in a new architecture family and has native visual understanding.

What to know

It is a 552 billion parameter mixture-of-experts model with a new encoder-decoder design: 8 billion active parameters for input and 16 billion for output.
DeepSeek says it needs much less memory for its cache than the previous generation, which it says lowers agent costs.
In the API, set the model to deepseek-flash. The older V4-Flash names temporarily route to V4.1-Flash.
DeepSeek says tests by multiple parties put it ahead of V4-Pro, and it is phasing V4-Pro out. From September 14, deepseek-v4-pro requests route to V4.1-Flash at V4.1-Flash prices until V4.1-Pro launches.

If you built on V4-Pro, review your model settings and pricing.

Source: DeepSeek API docs. We summarise in our own words; check the source for the full details.