Qwen Image

What is Qwen Image?#

Qwen Image is Alibaba's base text-to-image model and the lightest-cost entry in the Qwen Image lineup. The standout features are in-frame text rendering for Chinese and English plus multi-image fusion with up to ten references in a single call. The callable surface gives you an 800-character prompt, an optional 500-character negative prompt, a smart prompt rewriter (prompt_extend, on by default), an optional seed, and five aspect ratios (1:1, 16:9, 9:16, 4:3, 3:4). Reference images are optional, so the same model covers both pure text-to-image and reference-driven editing without a different endpoint.

Key features of Qwen Image#

Bilingual in-frame text rendering

Reliable Chinese and English typography inside the image. Useful for posters, retail signage, and bilingual packaging where the type has to read at thumb size.

Up to 10 reference images per call

Pass an array of URLs alongside the prompt for editing or multi-image fusion. Useful for compositing, reference-driven edits, and brand-locked sequences.

Smart prompt rewriting

The `prompt_extend` toggle (on by default) routes your prompt through Alibaba's rewriter for better adherence. Turn it off when you need wording preserved verbatim.

Negative prompts and seed control

Up to 500 characters of negative prompt to keep stray props out, plus an integer seed for reproducible runs. Useful for A/B copy variants and regression checks.

Best for#

Use cases#

Build a poster set with bilingual headlines by passing the Chinese and English copy in the prompt; Qwen Image holds both writing systems legible in the same frame. Run an ecommerce catalog refresh by fusing up to ten reference shots in a single call (product, lifestyle, environment) into one composited still. Swap an editorial hero from landscape to portrait by holding the prompt and seed steady and changing only the aspect ratio. Use the negative prompt to clear stray props from game UI drafts: cleaner icons in fewer iterations.

API examples#

Call Qwen Image from any language by POSTing to /v1/tasks. Full parameter docs live at docs.unifically.com/models/image/alibaba/qwen-image.

curl -X POST https://api.unifically.com/v1/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "alibaba/qwen-image",
    "input": {
      "prompt": "A serene Japanese garden with cherry blossoms",
      "aspect_ratio": "16:9"
    }
  }'

Successful submission returns a task_id. Poll GET /v1/tasks/<task_id> or set a callback_url on the request to receive the finished result.