by Stability AI
Adversarial distillation of SDXL for near real-time image generation. 2.6B UNet, only 1-4 steps needed. Quality is lower than SDXL base but generation is almost instant. Great for real-time previewing.
VRAM requirements, GPU fit, and setup notes for SDXL Turbo, including 8GB/12GB fit guidance where relevant. Recommended runtimes: ComfyUI and Diffusers support. Best download size: ~6.9 GB at FP16.
Your hardware
Detecting...
Measured quality metrics for SDXL Turbo outputs.
How often humans prefer this model's output (0-100%)
Visual quality and composition rating (5-9 scale)
Compare which GPUs can run SDXL Turbo at different precisions. FP8 uses less memory than FP16 when available, and the grade shows how comfortably each GPU handles the workload.
| Resolution | VRAM Required | RTX 4090 24GB | RTX 3060 12GB | RTX 4060 8GB | MacBook Pro M4 Pro 24GB |
|---|---|---|---|---|---|
| 512×512 | 6.7 GB | S● | S● | A● | S● |
| 768×768 | 6.8 GB | S● | S● | A● | S● |
| 1024×1024 | 7.0 GB | S● | S● | A● | S● |
from diffusers import StableDiffusionXLPipeline
import torch
pipe = StableDiffusionXLPipeline.from_pretrained(
"stabilityai/sdxl-turbo",
torch_dtype=torch.float16
)
pipe.to("cuda")
image = pipe(
prompt="your prompt here",
num_inference_steps=1,
height=512,
width=512,
).images[0]
image.save("output.png")Get started
Setup instructions for running SDXL Turbo locally
1. Download the model
Get the checkpoint from HuggingFace
2. Place in:
ComfyUI/models/checkpoints/3. Launch ComfyUI
python main.pyComfyUI Workflow
Basic txt2img workflow for SDXL Turbo
Drag & drop into ComfyUI or use File → Import
VRAM allocation at 1024×1024 on RTX 4090 24GB (24 GB)
Time per image at 1024×1024, 28 steps, FP16.
Download SDXL Turbo in the precision that matches your GPU. Lower precision usually means less VRAM pressure, while higher precision keeps more quality.
| Format | Precision | Size | Provider | |
|---|---|---|---|---|
| safetensorsRecommended | FP16 | 6.9 GB | official | Download |
Few LoRAs available specifically for Turbo. Some SDXL LoRAs work but may degrade quality.
Frequently asked questions
SDXL Turbo (2.6B parameters) requires approximately 7.0 GB of VRAM at FP16 precision for standard 1024×1024 image generation. If you want a lighter setup, lower precisions like FP8 can reduce memory use when available.
SDXL Turbo usually needs more than 8GB for comfortable local use. Check the VRAM table above for the exact resolution and precision trade-off.
SDXL Turbo is marked for ComfyUI and Diffusers support in our catalog, so those are the runtimes we recommend first for local setup. If your workflow uses another front end, check the model's available formats and workflow notes above before downloading.
Yes, the RTX 4090 (24 GB VRAM) can run SDXL Turbo comfortably at FP16. Expected generation time is around 500ms per image at 1024×1024.
There are currently no known ControlNet adapters for SDXL Turbo. Check Hugging Face and Civitai for community-contributed adapters.
Few LoRAs available specifically for Turbo. Some SDXL LoRAs work but may degrade quality. The LoRA ecosystem for SDXL Turbo is rated as "minimal". Each LoRA adds roughly 0.2 GB of extra VRAM.
On a reference GPU (RTX 4090 24GB), SDXL Turbo generates a 1024×1024 image in approximately 500ms at FP16 with 28 inference steps. Faster GPUs with higher memory bandwidth will produce images more quickly.
See also