All models
DeepSeek LLM Big Open weights

DeepSeek V4 Flash

202 days since DeepSeek announced it ·

DeepSeek V4 Flash is an open-weights LLM from DeepSeek, announced 24 Apr 2026. Its context window is 1M tokens. It has 284B parameters.

Input
—
per 1M tokens
Cache write
—
per 1M tokens
Cache read
—
per 1M tokens
Output
—
per 1M tokens
Context window
1M
tokens
Parameters
284B
total

No price on the lab’s own pages.

Our coverage

We haven’t written about DeepSeek V4 Flash yet.

DeepSeek · Big · Flash

open weights

A new one every ~139 days — due in ~113d.

  1. DeepSeek-V4.1-Flash-1017 +37d
  2. DeepSeek V4.1 Flash +139d
  3. DeepSeek V4 Flash +144d
  4. DeepSeek V3.2 +63d
  5. DeepSeek V3.2-Exp +7d
  6. DeepSeek V3.1-Terminus +32d
  7. DeepSeek V3.1

More from DeepSeek

All models
  1. DeepSeek-V4-Pro-0813
  2. DeepSeek V4 Pro
  3. DeepSeek-Math-V2

Facts last updated , from DeepSeek’s own pages.