Stable Diffusion 3 medium demo

Try Stable Diffusion 3 Medium for free in this page, without an account: a text-to-image model that understands full sentences, places several subjects correctly and can render short text in quotes.

Our service is free. If you like our work and want to support us, we accept donations (Paypal).

Try Stable Diffusion 3 Medium for free in this page, without an account: a text-to-image model that understands full sentences, places several subjects correctly and can render short text in quotes. Write what you want to see and get a 1024×1024 image in under a minute.

How to use Stable Diffusion 3 Medium

Describe the scene as a sentence: "a red cube on a blue sphere, to the left of a green cone" works better than a keyword list.
Put any text you want in the image between quotes, for example a poster title.
Click Run; the image appears in the frame after the queue.
Adjust the prompt rather than the negative prompt: SD 3 reacts less to negatives than older models.

Limits and questions

What is the difference with SDXL?

SD 3 Medium uses a diffusion transformer with three text encoders: better prompt understanding and readable text, but weaker reaction to artist names and fewer community fine-tunes. See the comparison guide.

Is Stable Diffusion 3.5 available here?

This demo runs SD 3 Medium. SD 3.5 (October 2024) improved anatomy and quality; you can run it locally with ComfyUI, see the installation guide.

Why is the demo slow or unavailable?

This demo is hosted by a third party (Hugging Face Space) and embedded here; it can be slow, queued or temporarily unavailable when the host is busy. Nothing to install, no account needed.

Read the guide: SDXL vs Stable Diffusion 3 vs SD 1.5