TECH

DeepSeek Releases Official V4-Flash Model as China’s AI Race Intensifies

Published: Aug. 1, 2026 1:37 a.m. GMT+8

00:00

00:00/00:00

Listen to this article 1x

DeepSeek released the official version of its DeepSeek-V4-Flash model on July 31. Photo: VCG

Chinese artificial intelligence startup DeepSeek has released the official version of its DeepSeek-V4-Flash model, introducing significantly enhanced autonomous agent capabilities and further reducing API costs.

The update, published Friday in the company’s developer documentation, arrived slightly behind its mid-July target and without the highly anticipated V4-Pro iteration. DeepSeek stated that the Flash model’s architecture remains identical to its April preview, with the performance gains driven entirely by extensive post-training.

You've accessed an article available only to subscribers

Subscribe today for just $.99.

VIEW OPTIONS

Unlock exclusive discounts with a Caixin group subscription — ideal for teams and organizations.

Save an extra $50. Introductory offer for new readers. Subscribe now.

Share this article

Open WeChat and scan the QR code

DIGEST HUB

Digest Hub

Back

Explore the story in 30 seconds

Chinese AI startup DeepSeek released V4-Flash with enhanced agent abilities and up to 50% cheaper API costs.
It outperformed its preview but still trails Moonshot's Kimi K3 (2.8 trillion parameters).
It raised nearly $7.4B, valuing it at 350 billion yuan, to double headcount and focus on AI agents.
It raised nearly $7.4B, valuing it at 350 billion yuan, to double headcount and focus on AI agents.

AI generated, for reference only

Who’s Who

AI generated, for reference only

What Happened When

April 2026：: DeepSeek released the preview version of DeepSeek-V4-Flash, which had an identical architecture to the official version.

June 2026：: DeepSeek announced a dynamic pricing mechanism that would double costs during peak hours, though it has not yet been implemented.

Before mid-July 2026：: DeepSeek set a target to release the official version of its DeepSeek-V4-Flash model by mid-July 2026.

Early summer 2026：: The V4-Flash preview ranked as the most-used model on OpenRouter for seven consecutive weeks.

Friday in 2026：: DeepSeek released the official version of DeepSeek-V4-Flash, slightly behind its mid-July 2026 target, without the V4-Pro iteration. The article was published on this date.

In 2026 (specific timing not given but during the product rollout period)：: DeepSeek conducted a massive private funding round, raising nearly 50 billion yuan ($7.4 billion), propelling its valuation to 350 billion yuan with backing from Tencent Holdings Ltd. and NetEase Inc.

AI generated, for reference only