
Qwen 3 Pro Uncensored
Qwen 3 Pro Uncensored is Alibaba's image model, hosted on Vellria with nothing to install: text to image and image to image, output up to 1440×1440. Run it in Studio or over the REST API, 10 credits (100 credits = $1.00).
vellria-qwen-3-proQwen 3 Pro is developed by Alibaba. Vellria provides access to the model; it does not own it. All trademarks belong to their respective owners.
Output
No generation yet
Cover art
10 credits
Your request will cost 10 credits.
About Qwen Image 3.0 Pro
Also known as: Qwen-Image 3.0 Pro · Qwen 3 Pro
Qwen Image 3.0 Pro is the higher tier of Alibaba's Qwen-Image 3.0, the image model from the Qwen team that also makes the Qwen language models. Searches for Qwen Image usually mean this family: the earlier Qwen-Image and Qwen-Image-Edit were open releases, and 3.0 is the current hosted generation. The Pro tier keeps the family's literal reading of prompts and adds two controls that matter once you are refining rather than exploring: a fixed seed and a negative prompt.
A fixed seed means you can rerun the same prompt and get the same frame, then change one word and see exactly what that word did. That is how you finish an image instead of rolling the dice: lock the seed when a composition works, then adjust lighting, wardrobe or expression in small steps. The negative prompt lets you rule things out explicitly — extra people in the background, text, tattoos you did not ask for — rather than hoping the model leaves them out.
Editing works as it does on the standard tier: attach a picture and describe the change, and the rest of the frame stays. Reference mode also unlocks more output shapes than a fresh start, a bigger square among them, which matters when the final frame needs size. Each extra reference adds to the cost, so bring only the ones the edit depends on.
Use Qwen Image Edit 3.0 to find an idea cheaply, then move the prompt here for the version you keep. If you need very large or ultrawide output, Seedream 5.0 Lite covers that; for a flat price across every shape, look at Bria FIBO 1.5.
Qwen Image 3.0 Pro in action
Qwen Image 3.0 Pro outputs with the exact prompts behind them.

Text to image
Natural light portrait of a freckled woman in her thirties with auburn curls and green eyes, wearing an oversized knit sweater, sitting by a rainy cafe window, gentle smile, soft bokeh, photorealistic

Text to image
Black-and-white glamour photograph of a muscular man in his thirties, shirtless in tailored trousers and suspenders, leaning in a doorway, hard side light sculpting his torso, classic Hollywood style

Text to image
Cyberpunk mercenary woman in her thirties with an undercut, a chrome cybernetic arm and glowing tattoos, cropped tactical jacket, standing on a rooftop above a sprawling neon megacity in the rain, cinematic concept art

Text to image
Cinematic interior of a 1960s jazz club, a sultry singer in her thirties in a sequined emerald gown at a vintage microphone, smoky haze, a single spotlight, band silhouettes behind her, warm tungsten film look

Text to image
Watercolor and ink illustration of a mermaid woman in her thirties wearing a top of woven seaweed and pearls, long teal hair flowing over her shoulders, sitting on a rock in a moonlit cove, shimmering scales, loose painterly washes

Text to image
Minimalist flat lay of champagne-colored silk lingerie, a pearl necklace and a glass of rosé on a white marble surface, soft daylight, editorial product photography, no text, no logo

Text to image
A woman in her thirties relaxing in a steaming stone hot spring, wrapped in a white towel with bare shoulders, eyes closed, snowy mountains behind her, soft mist, serene and sensual
Before
AfterEdit
Repaint the entire image as a classical oil painting on canvas, with thick visible brushstrokes, a painterly background and warm gallery tones, so it no longer looks like a photo; keep the composition and the likeness
When to pick Qwen Image 3.0 Pro
Pick it when
- A composition already works and you want to lock the seed, then change one word per run to see exactly what it does.
- You need to exclude specific unwanted elements, like background figures, text or tattoos, with a short negative prompt instead of hoping.
- The final frame has to be large and you can start from a reference image, which unlocks more shapes and a bigger square.
Pick another model when
- You are still exploring loose ideas and do not need seed control yet; the cheaper Qwen Image Edit 3.0 is the place to start.
- You need very large or ultrawide output from a fresh prompt; Seedream 5.0 Lite reaches those sizes.
- You want one flat price whatever shape the layout needs; Bria FIBO 1.5 prices every shape the same.
Prompt or reference
Starting fresh offers a square plus one wide and one tall shape. Attaching a picture opens up more shapes, including a larger square, so a frame that has to end up big is worth starting from an image. Every picture past the first adds to the cost, so attach only what the edit needs.
How to use Qwen Image 3.0 Pro on Vellria
- Prompt: up to 8,000 characters in the panel above.
- To edit or combine pictures, attach 1–3 reference images; leave them out to start from text.
- Cost: 10 credits — see pricing.
Example prompts
Studio portrait of an adult woman with short silver hair, bare shoulders, wrapped in a white sheet, seamless grey backdrop, beauty dish lighting, crisp detail. Negative: text, watermark, jewelry.
An adult couple in a dim jazz bar booth, her hand on his collar, cigarette smoke curling in a spotlight, 1960s film photograph. Negative: extra people, modern objects.
Fantasy art of an adult elven sorceress in a sheer gown standing in a moonlit forest pool, bioluminescent plants, painterly detail.
Edit: keep the person and pose, replace the backdrop with a sunlit greenhouse full of orchids.
Noir illustration of an adult private detective, shirt unbuttoned, slouched in a desk chair under a ceiling fan, venetian blind shadows across the room.
Tips
- Fix the seed as soon as a composition works, then change one thing per run so you can see its effect.
- Keep negative prompts short and concrete; listing ten vague dislikes weakens all of them.
- Start from an image when the final frame has to be large: reference mode offers bigger shapes than a fresh start.
- Copy a prompt that worked on the standard Qwen tier here unchanged before you start adjusting it.
Specifications
| Size (Text to image): | 1024×1024 (1:1)1536×1024 (3:2)1024×1536 (2:3) |
|---|---|
| Size (Image to image): | 1024×1024 (1:1)1280×960 (4:3)960×1280 (3:4)1280×720 (16:9)720×1280 (9:16)1440×1440 (1:1, max.) |
| Reference images: | 1–3 reference images |
See also: uncensored AI image generator · uncensored image editor.
Pricing
Text to image
| Size | Credits |
|---|---|
| 1024×1024 (1:1) | 10 credits |
| 1536×1024 (3:2) | 10 credits |
| 1024×1536 (2:3) | 10 credits |
Image to image
| Size | Credits |
|---|---|
| 1024×1024 (1:1) | 10 credits |
| 1280×960 (4:3) | 10 credits |
| 960×1280 (3:4) | 10 credits |
| 1280×720 (16:9) | 10 credits |
| 720×1280 (9:16) | 10 credits |
| 1440×1440 (1:1, max.) | 10 credits |
+1 credit per extra image
Qwen Image 3.0 Pro vs Qwen Image Edit 3.0 and Qwen Image Edit 2511
Qwen Image 3.0 Pro is the higher tier of Alibaba's Qwen-Image 3.0 on Vellria. It keeps the line's literal reading of the prompt and its edit mode, at a higher price and with a shorter list of fresh-image sizes than the standard tier.
| Model | Price | Max size | Reference images |
|---|---|---|---|
| Qwen 3 Pro Uncensored This page | 10 credits | 1440×1440 | Up to 3 |
| Qwen 3 Uncensored | 8 credits | 2048×2048 | Up to 3 |
| Qwen Image Edit 2511 Uncensored | 6 credits | 1280×960 | Up to 3 |
Prices and limits come from the live model catalog.
Qwen Image 3.0 Pro or Qwen Image Edit 3.0?
- Price: 8 credits (Qwen Image 3.0 Pro: 10 credits)
- Max size: 2048×2048 (Qwen Image 3.0 Pro: 1440×1440)
The standard tier is the place to explore: it costs less per image and offers larger and more varied fresh sizes. Bring a prompt here when a frame is worth a second pass on the Pro tier.
Qwen Image 3.0 Pro or Qwen Image Edit 2511?
- Price: 6 credits (Qwen Image 3.0 Pro: 10 credits)
- Max size: 1280×960 (Qwen Image 3.0 Pro: 1440×1440)
- Extra reference images: No extra charge (Qwen Image 3.0 Pro: +1 credit per extra image)
- Prompt length: 10,000 characters (Qwen Image 3.0 Pro: 8,000 characters)
- Settings: Seed (Qwen Image 3.0 Pro: Seed, Negative prompt)
Choose Qwen Image Edit 2511 for quick instruction edits at a lower flat price, with every reference picture included. It has no negative prompt, which Pro offers.
Qwen Image 3.0 Pro API: price and a first request
Qwen Image 3.0 Pro costs 10 credits over the API, the same rate as in Studio. Model id vellria-qwen-3-pro on POST /v1/generate/image.
curl https://vellria.com/v1/generate/image \
-H "Authorization: Bearer vll_..." \
-H "Content-Type: application/json" \
-d '{
"model": "vellria-qwen-3-pro",
"prompt": "…",
"size": "1024*1024"
}'Parameters, polling and uploads: API documentation · NSFW image API.
FAQ
Does Qwen Image 3.0 Pro allow NSFW?
On Vellria, yes, for fictional adult content: nudity and explicit scenes between adults are allowed and the prompt is sent as written. Minors are refused automatically and real people without consent are prohibited.
Is Qwen Image 3.0 Pro free?
Free to start: a new account created with Google gets 20 free credits, valid for 7 days (about 10 images on Z-Image Turbo Uncensored), and every signed-in account gets free daily uses of the NSFW prompt generator and the image to prompt tool. The gift is enough to try Qwen Image 3.0 Pro, which costs 10 credits. After that every run is paid from prepaid credits at the rate this page shows before you start, and a run that fails at the provider is refunded automatically. Is there a free uncensored AI generator? covers the cheapest way to a first result.
What does the Pro tier add over Qwen Image Edit 3.0?
A fixed seed and a negative prompt, plus more shape options when you start from a reference. The way it reads prompts is the same.
Can I use Qwen Image 3.0 Pro through an API?
Yes. The API section on this page shows a ready request with this model's id; the full reference is in the docs.
Is it the open-source Qwen-Image model?
It belongs to the same Alibaba Qwen-Image family, but this page runs the hosted 3.0 Pro version rather than the earlier open checkpoints, so there is nothing to download.



