Chinese artificial intelligence company DeepSeek unveiled a preview of its highly anticipated next-generation AI model, V4, on Friday, positioning the open-source system as a direct competitor to leading closed-source models from prominent US rivals, including Anthropic, Google, and OpenAI.

According to DeepSeek, V4 represents a major advancement over its predecessor models, particularly in coding—a capability that has become pivotal in the development of AI agents. This improvement aligns with industry trends exemplified by tools such as ChatGPT Codex and Claude Code.

The release of V4 also serves as a significant milestone for China’s semiconductor industry. DeepSeek explicitly emphasized the model’s compatibility with domestic hardware, including technology from Huawei.

The announcement arrives nearly a year after DeepSeek first disrupted the global AI landscape with its previous model, which gained widespread attention for its performance and cost-efficiency.

Source: The Verge