4x-UltraSharp & RealESRGAN x4plus in ComfyUI
Pixel-space super-resolution models paired with tiled KSampler passes for crisp 4K output. Requires 1 GB–2 GB (tiled) VRAM and loads from models/upscale_models/.
Updated Oct 17, 2026 · 1 min read
- Model Family
- ESRGAN Upscalers
- VRAM Target
- 1 GB–2 GB (tiled)
- ComfyUI Folder
- models/upscale_models/
- Precision
- FP16 / FP32 pth or safetensors
- Default Sampler
- ImageUpscaleWithModel or UltimateSDUpscale
Where to place 4x-UltraSharp & RealESRGAN x4plus weights on Windows
Pixel-space super-resolution models paired with tiled KSampler passes for crisp 4K output.
Place the primary `.safetensors` or `.gguf` weight file inside `models/upscale_models/` inside your ComfyUI directory, or point extra_model_paths.yaml to your existing shared models drive.
After moving the file, press **R** on the ComfyUI Studio canvas (or click Refresh) so the loader dropdown lists the new checkpoint immediately without restarting the server.
Recommended loader nodes and sampler settings
- **Precision & Memory:** FP16 / FP32 pth or safetensors. Cards at the lower end of 1 GB–2 GB (tiled) should pair this model with FP8 or GGUF quantization and VAE Decode (Tiled).
- **Sampler & Scheduler:** ImageUpscaleWithModel or UltimateSDUpscale. Wire the model output into KSampler and connect the matching CLIP and VAE streams.
- **Installer:** If you have not set up the desktop environment yet, get the Windows build at /download.
4x-UltraSharp & RealESRGAN x4plus Quick Reference for ComfyUI Windows
| Parameter | Recommended Setting | Notes |
|---|---|---|
| Target Directory | models/upscale_models/ | Refresh canvas with R key after copying |
| VRAM Tier | 1 GB–2 GB (tiled) | Use --lowvram or GGUF Q4 on 8 GB cards |
| Weight Format | FP16 / FP32 pth or safetensors | Never load untrusted .ckpt pickle files |
| Sampler Recipe | ImageUpscaleWithModel or UltimateSDUpscale | Wire through core KSampler node |
FAQs
Which folder does 4x-UltraSharp & RealESRGAN x4plus go into in ComfyUI?
Put 4x-UltraSharp & RealESRGAN x4plus inside `models/upscale_models/` in your ComfyUI installation, then press R on the canvas to refresh node dropdowns.
Can I run 4x-UltraSharp & RealESRGAN x4plus on an 8 GB NVIDIA GPU?
Yes — use FP8 or GGUF Q4/Q5 weights via ComfyUI-GGUF and keep tiled VAE decoding enabled so peak VRAM stays under 8 GB.
Run on your hardware. Free forever.
The full ComfyUI engine: open source, fast, extensible, and yours to run however you want. Download the desktop app and open this workflow on your own machine.
