Stable Diffusion XL SDXL
More info
- Creator
- Stability AI
- Released
- Type
- Open text-to-image model
- License
- OpenRAIL++ (variants)
Second generation of Stability AI's open image model — 1024×1024 native, more detailed than SD 1.5.
Two U-Nets (base + refiner). Standard fare on CivitAI, run locally with ComfyUI / Automatic1111 / Forge. Huge ecosystem of community finetunes (Pony, Juggernaut, RealVisXL). Stability AI has had corporate turbulence (CEO changes, lost funding), making SD 3 / SD 3.5 (2024) land lukewarmly. Flux (Black Forest Labs, ex-Stability founders) has taken over as "the new SD".