Collections

Upscale images

These models increase image resolution and quality.

Read our guide to upscaling images with AI to learn about various upscaling models.

Key capabilities of upscaling models:

  • Super-resolution - Upscale images by inferring high frequency details. Allows increasing resolution without losing quality.
  • Noise reduction - Remove artifacts and imperfections during upscaling. Results in cleaner image.
  • Face restoration - Improve facial features and textures for more natural faces.
  • Control over style - Adjust degree of realism vs hallucination during upscaling.

Our Pick: batouresearch/magic-image-refiner

For most upscaling needs, we recommend the batouresearch/magic-image-refiner model. This very flexible model can be used for upscaling, refining an image, or inpainting. The model can upscale images to either 1024x1024px or 2048x2048px, producing stunning results with significant detail.

Increase the resemblance parameter to get a more precise recreation of your original input image. Or, if you’re looking for something new and interesting, crank up the creativity parameter to encourage hallucination and create a new image inspired by your original input.

You may also be interested in the sister model, batouresearch/high-resolution-controlnet-tile, which upscales to a larger resolution of 2560x2560. However, it runs slower and produces less realistic-looking results.

Budget Pick: nightmareai/real-esrgan

If you need to upscale a large volume of images, we suggest using nightmareai/real-esrgan.

It runs fast on cheaper GPUs, like the Nvidia T4 (~1.8s for a 2x upscale), and produces reasonably good upscaled images without too many image scaling artifacts. Real-ESRGAN also includes an optional face_enhance option, which can help improve the quality and realism of AI-generated faces.

You can run Real-ESRGAN on an Nvidia A100 for faster upscaling speed (~0.7s for a 2x upscale) and 2.5x the amount of GPU RAM, allowing for significantly larger images. Just keep in mind that A100 GPUs cost more than T4 GPUs.

Recommended models

nightmareai/real-esrgan

Real-ESRGAN with optional face correction and adjustable upscale

46M runs

jingyunliang/swinir

Image Restoration Using Swin Transformer

5.6M runs

mv-lab/swin2sr

3 Million Runs! AI Photorealistic Image Super-Resolution and Restoration

3.5M runs

philz1337x/clarity-upscaler

High resolution image Upscaler and Enhancer. Use at ClarityAI.co. A free Magnific alternative. Twitter/X: @philz1337x

1.7M runs

cjwbw/real-esrgan

Real-ESRGAN: Real-World Blind Super-Resolution

1.5M runs

batouresearch/magic-image-refiner

A better alternative to SDXL refiners, providing a lot of quality and detail. Can also be used for inpainting or upscaling.

620.1K runs

cjwbw/rudalle-sr

Real-ESRGAN super-resolution model from ruDALL-E

471.2K runs

batouresearch/high-resolution-controlnet-tile

Fermat.app open-source implementation of an efficient ControlNet 1.1 tile for high-quality upscales. Increase the creativity to encourage hallucination.

348.6K runs

jingyunliang/hcflow-sr

Image Super-Resolution

220.1K runs

zsxkib/diffbir

✨DiffBIR: Towards Blind Image Restoration with Generative Diffusion Prior

124.7K runs

nightmareai/latent-sr

Upscale images with the latent diffusion superresolution model

112.7K runs

fewjative/ultimate-sd-upscale

Ultimate SD Upscale with ControlNet Tile

107.5K runs

cjwbw/supir

Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild. This version uses LLaVA-13b for captioning.

91.9K runs

xinntao/esrgan

Image 4x super-resolution

74.3K runs

lucataco/pasd-magnify

(Academic and Non-commercial use only) Pixel-Aware Stable Diffusion for Realistic Image Super-resolution and Personalized Stylization

28.2K runs

cswry/seesr

SeeSR: Towards Semantics-Aware Real-World Image Super-Resolution

19.2K runs

juergengunz/ultimate-portrait-upscale

Upscale Portrait Images with ControlNet Tile

19.1K runs

lucataco/demofusion-enhance

Image to Image enhancer using DemoFusion

9.1K runs

lucataco/stable-diffusion-x4-upscaler

Stable Diffusion x4 upscaler model

5.8K runs

lucataco/controlnet-tile

Controlnet v1.1 - Tile Version

3.6K runs