Tencent announced Hy4, its latest-generation large language model, on November 4, 2026. The company presented it as a new addition to the global open-source AI model ecosystem, with improvements aimed at practical applications. The release introduces capabilities that extend beyond conventional text-based interactions, including visual understanding and computer use.
According to Tencent, Hy4 supports image processing, complex chart analysis and mathematical visual reasoning. It also introduces the ability to interact with command-line and graphical user interfaces. The company said the model received specialised training involving professional software such as Blender and CAD, helping it plan and execute sophisticated tasks. Tencent reported that evaluations on its WorkBuddy platform produced an average performance score of 83.18 and a pass rate of 96.15%. Hy4 ranked first in nine of twelve evaluated scenarios within the mainland Chinese market. Tencent also said it planned to release the model weights and system cards in the coming days, with attention to cybersecurity and biological safety.
For learners and developers, Hy4 is significant because it combines language processing with visual analysis and direct computer interaction. These capabilities are relevant to software development, office productivity, design and automated workflows.
Tencent's announcement indicated that model weights and system cards would follow the initial release, allowing the open-source community to examine the model and its documented safety information.