Back to news
launchTencent2026-11-04

Tencent Hunyuan releases the official Hy4 on November 4, adding vision understanding and computer use

Tencent Hunyuan's official Hy4 has 770B total and 49B activated parameters with a 1M context, adds vision and computer use, and keeps API pricing unchanged from preview.

On November 4 Tencent Hunyuan released the official version of its latest large language model, Hy4. The company said that with another leap in capability, Hy has entered the ranks of the strongest open-source models for the first time. Hy4 model weights will be open-sourced shortly, a model system card will be published, and the company will push for more responsible open sourcing with particular attention to cyber and biosecurity risks. Compared with the Hy4 preview released on August 28, the official version significantly expands both pretraining and post-training scale, introduces vision capability for the first time, and substantially improves instruction following, long-document understanding, long-horizon tasks and deep research.

The newly added vision capability can combine perception, reasoning and tool calling to perform fine-grained image processing, complex chart analysis and mathematical visual reasoning. Hy4 also supports computer operation for the first time, flexibly combining command lines and graphical interfaces. Tencent carried out targeted training for professional software such as Blender and CAD as well as cross-application long-horizon tasks. The model targets visual and computer tasks, front-end and game development, financial and office scenarios, research and model development, and long-term personal assistants.

As previously reported, the Hy4 preview has 770B total parameters, 49B activated parameters and a 1M context length. The official version keeps preview pricing unchanged, with standard input at 6 yuan and output at 18 yuan per million tokens. Hy4 is live on Tencent Cloud and OpenRouter, and available in CodeBuddy, WorkBuddy, Yuanbao and ima.

混元Hy4视觉理解电脑操作开源