AI Models

AI Models are the program filters Topaz Labs has created to determine how to adjust or enhance your images, graphics, digital scans, or artwork.


The Standard AI Model is the overall recommended model to use across a variety of images. This model has been trained to see a variety of files but works best for photos, graphics, and generated images.

Standard v2, first introduced in version 7.0.0, is the default version in 7.1.0


Standard v2 reduces blurry patches that occurred in Standard v1.


High fidelity

The High Fidelity AI Model is trained to study high quality images with high resolution. Images from high end cameras are best. This filter can maintain the original look of images, without creating distortions.

This model intends to keep the original look of the image. It can be the most common upscaling filter after other denoise/sharpen filters.

Preserve the textures and authentic capture capture details with High Fidelity and minimally change details using the Denoise slider.


Standard v2 addresses more natural textures preserving them while maintaining sharpness.


This particular model has a slower processing time in order to produce better output results.

To keep the look of grain or noise in the image, tune down the Denoise and Fix Compression sliders for more natural results.

Photo by Taryn Elliott

Low resolution

Low resolution files are commonly identified as having fewer pixels. This model works best with screenshots, images pulled from the web or images with 72 pixels per inch (ppi).


Low resolution v1 (Legacy Model as of v.7.1.0) v1 is trained to remove more blur by default.


Photo by Esther Huynh

Text & shapes

Text & shapes enables more distinct patterns and shapes during upscaling with man-made objects, textures, and even written text and fonts.

This AI model is suitable for graphic design, typography, images with street signs, or text on architecture, automobiles, etc.


Photo by Marlon Soto


Photo courtesy of Fondo Antiguo de la Biblioteca de la Universidad de Sevilla

Art & CG

The Art/CG AI Model is best for artwork, drawings, illustrations, and computer generated graphics like Midjourney


Midjourney Illustration created by legendary_fox_47504

Very Compressed

The Very Compressed AI Model is best used for images with a lot of compression artifacts. Use this for images saved at a small size, scanned images or old digital images.


Photo by Rompalli Harish


The Lines AI Model emphasizes and heavily contrasts lines for architecture, cityscapes, and any image with thick lines.


Photo by Joshua


Recovery AI Model recognizes pixels and gives creative results. Designed with graphic designers, photographers, and web developers in mind, this model helps improve low resolution image quality from extremely small sized dimensions.

This Ai Model is currently in Beta version. There are a list of common conflicts and known issues that comes with attempting to use Recovery. Please review the Recovery FAQs ( before proceeding to run this AI model.

How to run Recovery

Import your image and select an upscale size, then click on Recovery.

For smaller images below 1000px, a 6x scale multiplier is recommended.


In the Settings section, click on a thumbnail to sample a preview of what the Ai model can output. Move the cursor to the Preview section to find a white box showing the parameter of the area you want to sample. Click anywhere in the Preview to begin rendering.

A new pop up panel appears and begins rendering your sample.


After the sample rendering is complete, use the Detail slider bar to adjust the generative results to be more creative.


Zoom into the Preview sample to further inspect the details of the results or select the dropdown area to choose a fixed Zoom option

To view the original input file, select "Original" at the top of the Preview pop up box. Flip back to your enhanced sample by clicking "Recovered".

Click Select Again to adjust and change your preview sample area.

If you are satisfied with the rendering, click Recover Full Image to begin the full processing.

Note that depending on the size of the image input, the processing can range between minutes and hours. Please review our Recovery FAQs ( to understand multiple common issues or conflicts when running this AI model.


Once processing is completed, use the Detail slider to fine tune the details. You do not need to resample and preview the image again. Running Recover Full Image multiple times on the same image generates varied results every time.


Check out Recovery FAQs for related troubleshooting information

Photo by Joi Ito

Legacy Models

Gigapixel's Generation v1 AI Models are the original models used in previous versions that are now considered under our Legacy models.

How to find Legacy Models:

Select the three dots next to the Auto Symbol > Show Legacy Models


Resize mode