DeepSeek Officially Releases V4 Pro AI Model

Chinese AI startup DeepSeek has officially announced the release of its updated V4 Pro model, designed to help it regain position against fast-developing local rivals. The DeepSeek V4-Pro-0813 model significantly expands AI agent management capabilities and is accessible via API, mobile app, and web interface. The company also announced price increases for API access to V4 Pro and V4 Flash, with pricing now dependent on peak-hour load. The community is watching closely because the cheaper V4 Flash surprisingly outperformed the April preview of V4 Pro in independent tests, indicating rapid technical improvements between preview and official release. Earlier in 2025, DeepSeek gained attention with its reasoning model R1, which raised questions about building powerful AI systems much cheaper than American competitors. Since then, many rivals have challenged its leadership in China, including Moonshot AI, Zhipu AI, MiniMax, Alibaba, and ByteDance. DeepSeek has also faced the challenge of turning leadership into a sustainable business, having raised $7.4 billion in investments in June and preparing a new funding round at a valuation of about $71 billion in July. The company plans to double its workforce across all divisions, including data centers and AI agent departments, and has increased hiring of chip developers to create its own AI accelerator to reduce dependence on Nvidia and Huawei.