Qwen Image
by Alibaba
Better text, finer textures, and your own LoRA on top.
What it is
Qwen Image 2512 is described as "an improved version of Qwen Image with better text rendering, finer natural textures, and more realistic human generation". Those three improvements point at the same use: pictures of people and materials that have to survive being looked at closely.
It supports LoRA, so a trained style or likeness applies here as well , which puts it in a small group with FLUX and Z-Image as the models in this studio that take your own work as input.
It is priced per megapixel rather than per image, so the cost follows the size you actually ask for.
What it is good at
Improved text rendering over the previous version, which matters the moment a picture contains a word rather than only a scene.
Finer natural textures is a stated improvement. Fabric, skin and stone are where a generated image usually gives itself away.
More realistic human generation, per its documentation , the hardest thing for an image model to get right.
A style or a likeness you trained once applies here, so a house look stops being something you re-describe.
Where it struggles
- Alibaba publishes no list of limitations for this model, so none is invented here.
- Priced per megapixel, so the bill follows resolution rather than image count. Worth knowing before you push the size up.
How to prompt it
Name the materials. Because texture is where this model is stronger than its predecessor, saying what things are made of , brushed steel, raw linen, wet slate , does more work here than describing a mood. For people, describe the light on the face rather than the expression.
An example that shows what that means
A close portrait of an older man in a heavy wool coat, side light from a window to the left picking up the weave of the fabric and the lines around his eyes, plain dark background, neutral colour, no retouching.
What we offer here
Straight from the catalogue, so this list never goes stale. The price is the provider’s published rate with our markup on top, and it is on the button before you run anything.
Questions
Can I use my own LoRA?
Yes. It is one of the models in the studio that accepts one, alongside FLUX and Z-Image Turbo.
How is it priced?
Per megapixel rather than per image, so the cost follows the resolution you ask for. The amount is on the button before you run.
Is it good for text in images?
Better than the version before it. For text as the main event, GPT Image 2, Ideogram and Seedream 5 are stronger choices.