HunyuanImage-3.0
Tencent’s HunyuanImage-3.0 is an open-source, 80B-parameter MoE model that unifies text and image generation via an autoregressive framework, enabling superior prompt alignment, world-knowledge reasoning, and sharp text rendering