Stable Diffusion XL

The workhorse open-weight image model.

Open Source 6–12 GB VRAMRuns locally

Actually FreeNo SignupOpen SourceWatermark-FreeHobbyist-FriendlyAPIPlugin

Requires: ComfyUI, AUTOMATIC1111 (stable-diffusion-webui), InvokeAI, Stable Diffusion WebUI Forge

Visit Stable Diffusion XLUpdated 2026-04-10 · Direct link

Hardware requirements

Runs locally · Mid GPU (12 GB)

6–12 GB VRAM

Min VRAM

6 GB

Rec. VRAM

12 GB

Min RAM

16 GB

Rec. RAM

32 GB

Disk

15 GB

GPU class

Mid GPU

CUDAApple Silicon ✓GPU RequiredQuant: FP16, FP8

Comfortable on RTX 3060 12GB or M-series 16 GB Macs.

Screenshot placeholder · Stable Diffusion XL

What is Stable Diffusion XL?

SDXL 1.0 + refiner. Still the most LoRA- and ControlNet-supported open model. Runs on 6–8 GB with optimization; 12 GB+ is comfortable. Pairs with thousands of community fine-tunes (Pony, Juggernaut, etc).

Pros & cons

Pros

Largest ecosystem of LoRAs/ControlNets
Runs on modest hardware
Battle-tested

Cons

Beaten by Flux on quality
Hands & text are still rough without LoRAs

What's actually free?

Open weights, CreativeML license — commercial use allowed.

✓ Actually FreeNo SignupOpen SourceWatermark-Free

Alternatives

FLUX.1 [dev]

12B parameter open-weight diffusion model.

Open Source 8–24 GB VRAM

Min VRAM: 8 GB
GPU class: High-end GPU
Quant: FP16

Actually FreeNo SignupOpen SourceWatermark-Free