Topaz Gigapixel AI

The upscaler that spent years as the internet's favorite "pay $99 once" software — until late 2025, when that era quietly ended.

Images · Image Upscaling · 4.4 ★

What is Topaz Gigapixel AI?

Gigapixel AI built its reputation on doing one specific thing extremely well: enlarging an image without the mushy, blurred result you get from a basic resize. It uses AI models trained specifically to reconstruct plausible fine detail as it scales an image up — rescuing a soft 12-megapixel shot for a poster print, or bringing an old 640×480 family photo up to something usable today. For years, it was also known for something almost as notable as its output quality: a one-time $99 purchase, no subscription, a genuine rarity in creative software.

That changed in September and October 2025, when Topaz Labs ended perpetual licensing entirely across its product line. If you bought Gigapixel before that cutoff, you keep what you paid for, but it's now formally a "legacy" product receiving only bug fixes — every new AI model and major feature going forward is subscription-exclusive, folded into a reorganized lineup called Topaz Studio. If you're reading an older review that quotes the famous $99 figure as current, it no longer is; that's worth knowing before you go shopping based on outdated information. What hasn't changed is the core technical strength: up to 6x upscaling in a single pass using nine specialized models — Standard, High Fidelity, Recover v2, Redefine Creative, and others — running entirely on your own GPU, with no image ever uploaded to a server.

🔄
Sep/Oct 2025
Perpetual licenses discontinued industry-wide at Topaz
📈
Up to 6x
Upscaling in a single pass
🧩
9
Specialized AI upscaling models included
🔒
100% local
Processing stays on your own machine, no cloud upload

The perpetual-license discontinuation is well-documented across multiple independent sources (CG Channel and others) dated late 2025. Pricing reflects Topaz Labs' published subscription rates as of mid-2026 and has already changed more than once in the past year.

Key features

📈

High-quality upscaling

Reconstructs plausible fine detail while enlarging, rather than just stretching and blurring pixels.

🧩

Nine specialized models

Different models suited to faces, low-light shots, creative reinterpretation, and more.

🔒

Fully local processing

Images are never uploaded anywhere — everything runs on your own GPU, offline-capable.

📦

Batch upscaling

Process large folders of images in one run, useful for archives or bulk photo restoration jobs.

🎨

Creative re-imagining mode

Redefine Creative can add plausible new detail for heavily degraded or very low-resolution sources.

🖥️

Native desktop app

Dedicated Windows and macOS applications, optimized for each platform.

Available models

Standard & High Fidelity General-purpose upscaling models for everyday photos
Recover v2 Tuned for rescuing detail from very low-resolution or degraded sources
Redefine Creative Adds plausible new detail for heavily degraded images

Integrations & platforms

Windows desktop app macOS desktop app Adobe Lightroom & Photoshop plugin

Pros, cons & best for

👍

Pros

  • Still widely regarded as the benchmark for pure AI upscaling quality
  • Fully local processing — genuine privacy advantage over cloud tools
  • Nine specialized models cover a wide range of source image problems
👎

Cons

  • No longer available as a one-time purchase — subscription is now mandatory for new features
  • Only enlarges existing images; it's not a generator or full photo editor
  • Output resolution is limited by your own GPU's VRAM
🎯

Best for

  • Photographers rescuing detail from soft, noisy, or low-resolution shots
  • Print work requiring a much larger file than the original source
  • Archives and agencies batch-processing large legacy image libraries

Take a look inside

Our verdict

4.4 / 5

Gigapixel's upscaling quality genuinely still sets the bar in 2026 — the nine specialized models handle a wider range of source problems than most competitors bother to address, and running entirely on your own hardware remains a real advantage for anyone who'd rather not upload sensitive images to a cloud service. What's changed is the value math: the beloved $99-once era is over, and evaluating it now means comparing an ongoing subscription against your actual upscaling volume, not a single small purchase you make and forget. For photographers, archives, and print work where output quality genuinely matters, it remains a strong pick; for occasional or lighter use, the subscription-only shift is worth weighing against cheaper browser-based alternatives.

FAQ

Can I still buy Gigapixel AI as a one-time purchase?

No, Topaz Labs ended perpetual licensing entirely in September and October 2025. If you already own a perpetual license, you keep it, but it's now a "legacy" product that only gets bug fixes, not new AI models.

Why do some reviews still quote a $99 price?

They're describing the old perpetual-license model, which is no longer available for new purchases. Current pricing is subscription-based, starting around $149/year for Gigapixel alone.

Does Gigapixel upload my images to the cloud?

No, processing happens entirely on your own machine using your GPU, which is a genuine privacy advantage compared to cloud-based upscaling tools.

Is Gigapixel a full photo editor?

No, it's specifically an upscaler. For denoising, sharpening, or color grading beyond the upscale itself, you'd need Topaz Photo AI, Lightroom, or another dedicated editing tool.

How large an image can Gigapixel produce?

Up to 6x the original resolution in a single pass, though very high output resolutions are constrained by your own GPU's available VRAM.

Is there a cheaper way to get Topaz's upscaling?

Bundling it into Topaz Studio, which includes Photo AI and Video AI alongside Gigapixel, can offer better value than paying for Gigapixel alone if you need more than one of Topaz's tools.