Topaz Wonder 4 Is Out: Features, Pricing and Access

The short answer
Topaz Labs released Wonder 4 on Nov 3, 2026, an all-in-one image upscale, sharpen and denoise model built on a Mixture of Experts architecture. It is available in Topaz for Web and Mobile, and through API providers at roughly $0.0428 to $0.0856 per run in one listing.
On this page (12 sections)
What is Topaz Wonder 4?
Wonder 4 is an image enhancement model from Topaz Labs that upscales, sharpens and denoises photos in one pass. Topaz announced it on Nov 3, 2026, describing it as an all-in-one model for maximizing images in one click, now available in Topaz for Web and Mobile.
Magnifying lenses sharpening separate regions of a photo grid
According to Topaz Labs' launch post, the model brings an "industry-first model architecture" for universal photo restoration. That claim is Topaz's own and has not been independently checked.
Creator @maxescu called Topaz "still unmatched in this business" in a reaction post. That is opinion, not a benchmark.
How does Wonder 4's Mixture of Experts work?
Wonder 4 uses a Mixture of Experts design that splits an image into areas, analyzes each one, and applies several enhancements at once, tailored to that area. Topaz says this delivers better quality across all types of photos, rather than one setting for the whole frame.
In practice, that suggests a portrait with a soft background, sharp text and noisy shadows can get different treatment in each region. Topaz has not published a technical paper, expert counts or side-by-side comparisons, so how well it works on real ad assets is still to be seen.
Topaz Wonder 4 release date and availability
Wonder 4 was released on Nov 3, 2026 and is available now in Topaz for Web and Mobile. Topaz's November 2026 product updates page features it, and API providers also list it as topazlabs:wonder@4 for image to image and upscale tasks.
The desktop apps are a separate question. Topaz's launch post only mentions Web and Mobile, and the pricing page lists Wonder 3 in the Web image workflow list. Whether Wonder 4 reaches Topaz Photo or Gigapixel has not been announced.
Topaz Wonder 4 pricing and free access
Topaz for Web, Image is listed at $12/mo (regular $19) in a limited-time offer of up to 50% off. API provider pricing in one listing runs $0.0428 to $0.0856 per run. Topaz has not announced a free tier for Wonder 4.
| Option | Price | Notes |
|---|---|---|
| Topaz for Web, Image | $12/mo (was $19) | Unlimited image rendering, 32 MP export limit, 2 concurrent jobs |
| Topaz for Web, Video | $19/mo (was $29) | 200 monthly cloud video credits |
| Topaz Studio | $399/yr or $69/mo | All apps, unlimited cloud image rendering, 300 monthly video credits |
| API listing (runs) | $0.0428 or $0.0856 | About 23 or 11 runs per $1, varies by example |
The listing does not explain why the two examples cost different amounts. Likely factors such as input size or upscale factor are not confirmed.
Wonder 4 key features and specs
Wonder 4 handles upscaling, sharpening and denoising in a single model, with image to image input. The API listing shows an upscale factor parameter, output formats of JPG, PNG and WEBP, sync or async delivery, webhook support and zero data retention.
| Spec | Details |
|---|---|
| Developer | Topaz Labs |
| Release date | Nov 3, 2026 |
| Tasks | Upscale, image to image |
| Architecture | Mixture of Experts |
| Output formats | JPG, PNG, WEBP |
| Web export limit | 32 MP |
| Maximum upscale factor | Not announced |
| Commercial use | Limited: orgs under $1M USD annual revenue |
Wonder 4 vs Wonder 3 and other Topaz tools
Wonder 3 still appears in the Topaz for Web image workflow list alongside Standard Upscale, Creative Upscale, Face Enhancer, Denoise and Sharpen options. Topaz has not published a comparison between Wonder 3 and 4.
The main difference Topaz claims is the regional, multi-enhancement architecture. Until there are side-by-side results, treat the improvement as a claim. Run your own worst-case assets through both before switching.
How to use Wonder 4 for AI video and static ads
Wonder 4 is an upscale, sharpen and denoise model, so a natural spot for it is the last step before you export: sharpen soft product shots, enlarge small images and prepare thumbnails. Topaz describes it as an enhancement model, and nothing in its announcement mentions generating new images.
Designer desk comparing a sharp product photo against a soft one
- Upscale AI-generated product images before placing them in static layouts. Our guide to static ads with AI covers the layout side.
- Sharpen and enlarge low-resolution stills you plan to use in ads, such as stills pulled from UGC clips.
- Chain it after generation in Flow Builder, or compare upscalers in the AI Playground.
Check the 32 MP export cap on Web plans and the commercial use terms if your company is above $1M in annual revenue. Also inspect faces, packaging text and logos after any upscale, since enhancement models can alter fine detail.
What to watch next
Watch for Topaz's technical details, a Wonder 3 comparison, and whether Wonder 4 reaches the desktop apps. Per-run pricing may also shift as more API providers add it.
For related model news, see Krea 3 pricing and Recraft Logo Maker.
What people are posting
JUST RELEASED! 🪄 Meet Wonder 4—the revolutionary all-in-one image upscale, sharpen, and denoise model for maximizing images in one click. Wonder 4 brings an industry-first model architecture, unlocking the ability for universal, effortless photo restoration. The debut “Mixture https://t.co/qjDtguSEzP
Topaz is "Wonder"-ful. They keep on shipping incredible models. Still unmatched in this business.
Key takeaways
- Topaz Labs announced Wonder 4 on Nov 3, 2026 as an all-in-one image upscale, sharpen and denoise model.
- It uses a Mixture of Experts design that splits an image into areas and enhances each one separately.
- Wonder 4 is available in Topaz for Web and Mobile.
- An API provider listing shows $0.0428 and $0.0856 per run, with the model ID topazlabs:wonder@4.
- Topaz for Web, Image is listed at $12/mo (down from $19) in a limited-time offer, with a 32 MP export limit.
- Topaz has not published resolution limits or upscale factors for Wonder 4 in the material available so far.
Compare 200+ AI video and image models side by side
Run Kling, Veo, Sora, Seedance, Flux, Ideogram and Recraft in one Playground, or turn any product URL into ready-to-launch ads.
Frequently asked
When was Topaz Wonder 4 released?
Topaz Labs announced Wonder 4 on Nov 3, 2026 with a post saying it was just released. It is available now in Topaz for Web and Mobile.
How much does Topaz Wonder 4 cost?
Topaz for Web, Image is listed at $12/mo (regular $19) in a limited-time offer of up to 50% off. One API provider listing shows $0.0428 to $0.0856 per run depending on the example.
What is Mixture of Experts in Wonder 4?
Topaz says the technology splits an image into areas, analyzes them, then applies multiple enhancements at once for each specific area. The goal is better quality across different types of photos.
Can you use Wonder 4 for commercial ads?
Topaz's web plans list limited commercial use: personal and commercial use for organizations under $1M USD annual revenue. Larger companies should check Topaz's terms before using it.
Is there an API for Wonder 4?
Yes, it is listed with API providers under the ID topazlabs:wonder@4, with image to image and upscale tasks. Output formats include JPG, PNG and WEBP.
Sources
More AI model news & creative guides

Flux.3 Image Klein Is Out: Specs, Pricing and Access
Black Forest Labs released FLUX 3 Image Klein, a 5.6B open weights image editing model with 4K output, API pricing from $0.008 per image and local support.

P-H3-Image: Pruna's Open Image Model From MiniMax H3
Pruna AI open-sourced P-H3-Image, an image generation and editing model built from MiniMax-H3: 8 steps, up to 16 MP, 8-image edits, ComfyUI support.

Hy4 Preview: Tencent Hunyuan Release, Pricing and Access
Tencent Hy4 is teased as coming soon. See the Hy4 preview's Agent Arena ranking, listed pricing, 1,000K context window and what it means for ad teams.
AI UGC Ads: What They Are, What They Cost and How to Make One
AI UGC ads explained for D2C brands: how they compare with creator UGC on cost and speed, which formats work, the disclosure rules and how to make one.
Hook Rate and Hold Rate: Formulas and Benchmarks for Meta Ads
How to calculate hook rate and hold rate for Facebook and Instagram video ads, what good looks like, how to add them to Ads Manager and what to fix when they drop.
How to Make Static Ads With AI: A Guide for D2C Brands
How to make static ads with AI, step by step: the sizes Meta wants, eight formats worth testing, a brief template and the checks to run before you upload.
Your move
See your product as a winning ad
Turn a product URL into static ads and UGC-style videos, rebuild the competitor ads that already work, and read your Meta data to know what to make next.
AdGraam News DeskAI video and image news
The AdGraam News Desk tracks AI image and video model launches, leaks and rumors from labs, leakers and creators, and checks each claim against its sources. Unconfirmed stories say so, and each page is updated as its story develops.