According to CoinMeta, Tencent's Hunyuan officially released a new generation of large language model Hy4 on November 4th. Compared to the previous Hy4preview, the new model further expands the scale of pre-training and post-training, introduces visual understanding capabilities for the first time, and focuses on enhancing command following, long-text comprehension, long-range agent, and performance in real-world productivity tasks. The Hy4 model has a total parameter size of 770 billion, adopts a hybrid expert architecture, and supports computer operations for the first time, enabling it to execute complex tasks using both command lines and graphical interfaces. Evaluation results show that Hy4 has joined the ranks of the world's top open-source models, and the model weights will also be made open-source in the near future.