On June 27, Tencent expanded its Hunyuan large model family with the introduction of Hunyuan-A13B, which is now fully open-sourced. Boasting a total of 80 billion parameters, yet leveraging only 13 billion active parameters, this model delivers performance on par with top open-source models, while offering faster inference speeds and enhanced cost-effectiveness. Notably, Hunyuan-A13B can be deployed under extreme conditions using just a single mid-to-low-end GPU card. Seamlessly integrated into the open-source mainstream inference framework ecosystem, it promises to be highly accessible and user-friendly for individual developers.