What is Diffusion model?
The main technology behind AI images and video: the model starts from random noise and gradually refines it into a picture that matches your prompt.
During training, a diffusion model learns to remove noise from millions of images. To create something new, it starts from pure static and removes noise step by step, steered at each step by your prompt, until a clear image appears.
This is why the same prompt can give different results each time (different starting noise), and why settings like the seed and the number of steps change the output.
Why it matters to you
If you love one result and want variations, reuse the same seed (where the tool allows) and change only one word of the prompt.
Related terms
The number that sets the random starting point of an AI image. Same prompt + same seed gives (nearly) the same image.
Creating a picture from a written description. The core feature of ChatGPT images, Gemini, Midjourney and Flux.
Generating a short video clip from a written description — the technology behind Veo, Sora and Kling.