Originally known as Gigapixel AI, Gigapixel is designed to enlarge digital images while preserving their sharpness by utilizing machine learning methods to fill in missing details. Its algorithms are trained using a dataset from "millions of photos of all subjects and styles" and can enlarge photos "up to 600%" before the image quality begins to noticeably degrade. Although the application is primarily intended for photographers, it can also be used for rendering, and in version 5.3, a special artificial intelligence mode for CG images was introduced. In addition to scaling low-resolution renders for faster high-resolution image creation, users can remove residual noise from renders by upscaling them and then downsizing them to the original size.
Gigapixel 7.0 contains a "huge number of changes," albeit mostly under the hood, including a rework of the artificial intelligence engine to make it "faster and compatible with new graphics processors." Topaz Labs also released the second version of two AI models in the program, Standard and High Fidelity, improving the consistency of deblurring and noise reduction operations. The update touched on the program's user interface, which now supports panning and zooming gestures in the preview, as well as side-by-side before-and-after comparison in split-screen mode.
Users can now drag the list of files up and down, as well as use keyboard shortcuts to select or deselect all images in the list. Additionally, the handling of previews for RAW camera images has been updated to be closer to Photo AI, Topaz Labs' specialized software for enhancing photos.
Gigapixel 7.0 is available as a standalone application for Windows 10 and above and macOS 11.0 and above, as well as a plugin for Photoshop and Lightroom Classic. The cost of new licenses is $99. The program is based on OpenGL and supports AMD, Apple Silicon, Intel, and NVIDIA graphics processors.