Topaz Gigapixel AI 1.3.2

Topaz Gigapixel AI

A desktop AI image upscaler for macOS that enlarges photos while trying to preserve detail—aimed at photographers, retouchers, and print prep who need more resolution than the original file provides.

How upscaling works on Mac

Topaz Labs positions Gigapixel as a dedicated upscaling tool rather than a general photo editor. You load single images or batches, pick a model suited to the content—Standard or High Fidelity for general photos, Wonder for generative detail, Recover or Face Recovery for portraits—and choose a scale factor up to 16x. Processing can run locally on Apple Silicon or Intel Mac GPUs for most core models, or through Topaz cloud rendering where required (generative Wonder models on Intel Macs rely on cloud). Batch queues, before/after preview, and export to common formats fit a workflow of archiving scans, preparing large prints, or rescuing low-res client assets.

Plugins extend the same engine into Adobe Photoshop (via Automate) and Lightroom Classic as an external editor, so you can upscale without rebuilding a separate standalone pipeline—though cloud rendering is not available from the Lightroom plugin path according to Topaz documentation.

Models, hardware, and licensing

The current Gigapixel app supersedes the older Gigapixel AI branding and ships a wide model set: Wonder 2/3, Standard, Standard Max, High Fidelity, Low Res, Text & Shapes, Art & CG, Recover, Redefine, and Face Recovery, with unlimited local and cloud rendering on paid plans. macOS 12 Monterey through macOS 15 Sequoia are listed as supported; Apple Silicon Macs get the best local performance, while Intel Macs may need cloud for generative models. Licensing is subscription-based—standalone Gigapixel or bundled Topaz Studio—with Personal and Pro tiers differing mainly in cloud concurrency and commercial terms.

Version 1.3.2 in practice

Build 1.3.2 (June 17, 2026) is a stability and format patch: JXL import, HEIC crash fixes via an OpenImageIO update, cleaner Face Recovery v3 rendering, and a Mac installer fix when unpacking Wonder models. Topaz also documented batch-processing stalls affecting some users shortly after release and advised reverting to 1.3.1 until a fix ships—worth weighing if your workflow is queue-heavy. For occasional upscales and new format support, 1.3.2 is the current vendor release; for large unattended batches, check community status or stay on 1.3.1 if you hit freezes.