DeepSeek-V4-Flash changelog

DeepSeek-V4.1-Flash replaces V4-Flash, V4-Flash retired

DeepSeek released V4.1-Flash with native vision under the model name deepseek-flash, and retired V4-Flash and V4-Flash-Vision-Exp. The old names temporarily route to V4.1-Flash.

Who is affected:

Apps calling deepseek-v4-flash or deepseek-v4-flash-vision-exp.

Action needed:

Set the model name to deepseek-flash.

DeepSeek introduces peak and off-peak API pricing

DeepSeek updated API pricing with peak and off-peak rates, with off-peak 50% lower than peak, effective 16:00 UTC on August 16, 2026.

Who is affected:

DeepSeek API users, especially those running heavy workloads during peak hours.

What to do:

Schedule flexible batch workloads in off-peak hours.

DeepSeek-V4-Flash official API release enters public beta

DeepSeek re-post-trained V4-Flash as DeepSeek-V4-Flash-0731 with much stronger agent benchmarks than V4-Pro-Preview and native Responses API support for Codex. The deepseek-v4-flash model name now serves it. The V4-Pro API and the app and web models were unchanged.

Who is affected:

Apps calling deepseek-v4-flash.

What to do:

Retest prompts on deepseek-v4-flash, since the model behind the name changed.

DeepSeek V4 preview released with open weights

DeepSeek released DeepSeek-V4-Pro (1.6T total, 49B active parameters) and DeepSeek-V4-Flash (284B total, 13B active) as open weights and on its API, with a 1M context window as the default.