Diffusion Model

image

The dominant architecture for AI image and video generation. It learns by adding noise to training images until they are pure static, then learning to reverse the process. At generation time it starts from random noise and denoises it step by step into a picture matching your prompt. Stable Diffusion, Imagen 3, and FLUX are all diffusion models. Practically, this explains key settings: sampling steps (how many denoising passes) and seed (the starting noise). Same prompt plus same seed equals the same image — change either one and the output changes.