On March 7, 2026, the Tencent Hunyuan team released a technical report proposing the 'HY-WU' (Hunyuan Wuxiang) paradigm, a novel functional memory paradigm designed to address the issues of 'catastrophic forgetting' and 'performance seesaw' in traditional AI models. This paradigm achieves dynamic routing by training a parameter generator to produce personalized parameters in real-time during model inference, thereby avoiding parameter conflicts in shared weights. In text-guided image editing tasks, HY-WU demonstrated exceptional performance, achieving a comprehensive score that ranks among the leading levels of open-source models.
