Tencent has released Hy4, a new generation of its flagship large language model, positioning it as a “top-tier” open-weight option in China’s fast-moving AI race. Multiple outlets say Hy4 expands beyond text to include vision and “computer use,” framing the update as a broad capability jump from earlier versions.
According to TechNode and the South China Morning Post, Tencent highlights that Hy4 can interpret images and visual materials such as charts and documents, and it can reason across multiple images. Both also describe enhanced agent-like functionality, including the ability to operate tools through command-line interfaces and graphical workflows.
Nikkei Asia and Pandaily focus more generally on Hy4’s placement within the open-weight landscape, while TechNode provides additional detail about targeted training applications and multi-program tasking. Across reports, Tencent links Hy4’s capabilities to use in enterprise software ecosystems, though the specific examples and emphasis differ by outlet.
