Chinese AI company DeepSeek released a preview of its next-generation model, V4, on Friday, April 25, 2026, claiming the open-source system can compete with leading closed-source AI from Anthropic, Google, and OpenAI.
DeepSeek says V4 represents a significant improvement over its previous models, particularly in coding — a capability that has become central to AI agents and driven the success of tools like ChatGPT Codex and Claude Code. The company also highlighted V4’s compatibility with domestic Huawei hardware, marking a milestone for China’s chip industry.
The release comes roughly a year after DeepSeek disrupted the US AI industry with its R1 model, which the company claimed was trained at a fraction of the cost of comparable US systems. DeepSeek has not disclosed V4’s training costs or the hardware used to build it.
The announcement arrives amid ongoing controversy surrounding the company. US officials have accused DeepSeek of using banned Nvidia chips, and Anthropic has claimed DeepSeek misused Claude to improve its own products.
The V4 preview may signal that DeepSeek remains a competitive force in the global AI race, though the lack of transparency around training costs and hardware could draw further scrutiny from US regulators and rivals.
Source: The Verge