Model · Alibaba
Qwen Image, generation and editing in one place
Qwen Image is Alibaba's image model family, from the Qwen lineage. It is in the catalogue in three generations, each available as text-to-image and as editing of an existing image.
What Qwen Image brings
Three generations to choose from
Qwen Image 2, Qwen Image 2 Pro and Qwen Image 3 sit in the same picker. One brief goes to all three, and you keep the one that fits.
Editing with a sentence
Each generation has an editing version: supply an image and describe the change — a background, a colour, an object — with no mask to trace.
A negative prompt
On top of the brief, Qwen accepts a list of what to avoid. Useful to rule out a recurring defect without weighing down the main prompt.
Text inside the image
The Qwen Image lineage is known for rendering text inside images, including non-Latin scripts. Worth testing on a poster or a package.
The Qwen Image engines available here
Generation and editing, for each version.
- Qwen Image 2Text to image, and editing (Qwen Image 2 Edit).
- Qwen Image 2 ProText to image, and editing (Qwen Image 2 Pro Edit).
- Qwen Image 3The most recent version, as text to image and as editing.
What Beemm Vision adds
An image model is only one step of a production.
- Compare Qwen Image, Nano Banana and Seedream on one brief.
- Take the result into the image editor: background removal, upscale, erase.
- Replay a workflow over a list with Scale mode.
- Ship the graph as an app for the team.
- One credit balance, and no Alibaba Cloud account.
Getting the most out of Qwen Image
Put the text to display in quotes
For a poster, a headline or a label, write the exact text in quotes in the prompt and say where it should appear. Proofread every letter before delivering: even a model known for text can get a long word wrong.
Keep the negative prompt short
The negative prompt is there to rule out a specific defect — a watermark, distorted hands, a busy background. A long generic list dilutes the effect; three or four targeted terms are enough.
Qwen, Nano Banana or Seedream?
No model wins everywhere, and the result depends as much on the subject as on the model. On the canvas, one brief goes to all three and the renders sit side by side: you choose on your image, not on a leaderboard.
Frequently asked questions
- Can Qwen Image edit an existing image?
- Yes: Qwen Image 2, 2 Pro and 3 each have an editing version. Supply the start image and describe the change.
- Which Qwen Image version should I choose?
- The most recent is Qwen Image 3. The safest approach is still to run one brief on all three versions and compare the results side by side on the canvas.
- Do I need an Alibaba Cloud account for Qwen Image?
- No. Beemm Vision calls the provider for you and bills in credits on a single balance shared across the whole catalogue.
Try Qwen Image
Open the canvas and pick Qwen Image. The full catalogue is available from the free plan.
Keep reading
- Nano Banana on Beemm VisionUse Google's Nano Banana 2 and Nano Banana Pro image models on one canvas: legible in-image text, conversational editing, output up to 4K.
- Kling on Beemm VisionGenerate video with Kling V3 and Kling O3 4K: text, image or reference to video, up to 15 seconds, multi-shot sequences and generated audio.
- Seedream on Beemm VisionGenerate with ByteDance's Seedream and Seedream 5 Pro: photographic rendering, instruction-based editing, output up to 4K, on one canvas.
- The full catalogueEvery image and video model available in the canvas.