Chinese AI startup DeepSeek on Thursday officially released the V4‑Pro‑0813 model and announced changes to API pricing. According to the company, the new Pro release offers substantially improved agent capabilities and is available via API, mobile app and web interface.
DeepSeek said it will raise API prices for the V4 Pro and V4 Flash models and introduce peak and off‑peak pricing. The company framed the change as a means to improve capacity management and cost efficiency, though the new pricing structure may prompt reactions from customers.
Why the release drew attention
The launch attracted particular attention because DeepSeek’s cheaper V4 Flash unexpectedly outperformed a pre‑release April version of the V4 Pro in several independent tests. Since the Pro is intended as the company’s flagship model, that gap suggested DeepSeek made significant technical advances between the preview and the official release.
Hiring, capacity expansion and chip plans
DeepSeek said it is expanding its team, compute capacity and fundraising efforts to try to regain an edge over domestic rivals. Reuters reported in July that the company planned another funding round at about a $74 billion valuation, after raising roughly $7.4 billion in June as its first external financing.
The company said it intends to at least double its headcount, with a focus on data‑center and AI‑agent development teams. Reuters also reported that DeepSeek is recruiting chip‑design engineers to develop in‑house AI chips, a move aimed at reducing longer‑term dependence on suppliers such as Nvidia or Huawei.
Competitive context and challenges
DeepSeek came to prominence in early 2025 when its R1 model spread rapidly and raised questions about whether high‑performance AI systems can be built at substantially lower cost than U.S. rivals. Since then, a number of Chinese competitors — including Moonshot AI, Zhipu AI, MiniMax, Alibaba and ByteDance — have released models that have eroded DeepSeek’s early market advantage.
The company’s immediate challenge is turning rapid fame and technical progress into a sustainable business model. The pricing adjustments, hiring push and in‑house chip efforts all aim to stabilize operations and reduce external risks over time.
(An AI assistant contributed to preparing this article; the final content was edited and verified by our journalist.)
