What is the DeepSeek-V4-Flash Official API?

Official API for DeepSeek-V4-Flash in public beta as of 31 July 2026. Model id deepseek-v4-flash with Responses API support and upgraded Agent claims per docs.

Are $0.14 input and $0.28 output permanent?

They are the published Flash rates now. Peak 2× pricing is announced for Beijing peak hours but not yet active.

Does this beat ChatGPT or Claude for writing quality?

No writing-quality leaderboard ships in this release. Benchmarks are agent/coding. Self-test required.

How is Flash different from Pro in this announcement?

Flash entered public beta today. Pro is higher priced. Codex support for Pro expected early August 2026.

Do I need Codex to use it?

No. Call the API directly or via any client. Codex needs version ≥ 0.144.0 and currently supports Flash only.

Is it safe for brand or customer data?

Attractive for draft cost; review DeepSeek terms and your compliance rules before sending private corpora.

DeepSeek AI writing models API pricing

DeepSeek-V4-Flash API Public Beta: What Content Creators Should Know

DeepSeek-V4-Flash Official API entered public beta on 31 July 2026: API pricing, ~1M context, Agent upgrades — plus honest limits for content creators.

Mahmoud Hussein