The Tencent Text-to-Image model is a brand new two-stage image generation model with billions of parameters, achieving multiple industry-leading technological breakthroughs. This model is very competitive in real image generation, virtual image generation, and stylized image generation, and has extremely high controllability and scalability. Tencent's Text-to-Image model is widely used in commercial fields such as advertising, social media, gaming, and media, generating 5 million images per day. We'll introduce Tencent's breakthrough in algorithmic technology, as well as the commercial achievements of multiple industrial applications.