PIXART-α is a Transformer-based text-to-image generation model. Its image generation quality rivals that of state-of-the-art image generators. It supports high-resolution image synthesis, features significantly faster training than existing large-scale T2I models, and boasts low training costs, saving nearly $300,000 and reducing CO2 emissions by 90%. PIXART-α excels in image quality, artistry, and semantic control, providing new insights for the AIGC community and startups to accelerate the development of high-quality, cost-effective generative models from scratch.