Qwen Image 3.0 Text-to-Image API by Alibaba
Generates images from a text prompt at resolutions up to 2048×2048, with automatic prompt rewriting and prompt-guided resolution selection, building on Qwen strength in complex text rendering and precise prompt adherence
Qwen Image 3.0 Text-to-Image को Alibaba द्वारा विकसित किया गया है। Atlas Cloud (Atlas Cloud AI LLC द्वारा संचालित) इस तक पहुँच प्रदान करता है, इसका स्वामी नहीं है। सभी ट्रेडमार्क उनके संबंधित स्वामियों की संपत्ति हैं।
Qwen-Image 3.0 (Text-to-Image)
Qwen-Image 3.0 is a high-fidelity text-to-image model that generates images from detailed text prompts. Built-in prompt expansion (DPE/APE) enriches short prompts, while flexible resolution and reproducible seeds make it suitable for both rapid exploration and controlled iteration.
Invite-beta model for interface testing. Output does not represent final quality.
🌟 Key Features
✏️ Text-to-Image
Generate high-fidelity images from text prompts alone, with strong semantic understanding and photorealistic quality.
🧠 Prompt Expansion (DPE / APE)
prompt_extend intelligently rewrites short prompts. Choose direct (DPE, single-pass) or agent (APE, agentic pipeline).
📐 Flexible Resolution
Specify size as width*height (up to 2048×2048), or omit it to let the model pick automatically from the prompt.
🔁 Reproducible Results
Pin a seed value to reproduce or iterate on a specific generation, enabling systematic prompt refinement.
🖼 Batch Generation
Request multiple images per call with n to explore variations at once.
⚙️ Parameters
| Parameter | Required | Description |
|---|---|---|
prompt | ✅ | Text description of the image to generate. |
size | ❌ | Output resolution as width*height (e.g. 2048*2048). Omit for automatic. Range: 512×512–2048×2048. |
n | ❌ | Number of images to generate (default: 1). |
negative_prompt | ❌ | Content that should NOT appear in the image. |
seed | ❌ | Seed for reproducible output, range [0, 2147483647]. Omit for random. |
prompt_extend | ❌ | Enable intelligent prompt rewriting (default: true). |
prompt_extend_mode | ❌ | direct (DPE, default) or agent (APE). |
💡 Prompt Tips
A vertical outdoor portrait in warm afternoon light: a young woman turns to the camera with a bright smile, holding a large bouquet of orange and peach roses, backlit hair rimmed with gold, a blurred sunlit city street behind her, warm film grain.
- Be specific about subject, environment, lighting, and camera angle for best results.
- Fix
seedwhen iterating on a prompt to isolate the effect of each change. - Keep
prompt_extendon for short prompts; turn it off when you want the model to follow your exact wording. - Request multiple images with
nto explore variations in a single call.
🎯 Use Cases
- Creative Illustration — Generate concept art, characters, and environments from descriptive text prompts.
- Marketing Visuals — Produce eye-catching posters, banners, and ad creatives quickly.
- Social Media Content — Generate scroll-stopping visuals in portrait or square formats for any platform.
- Rapid Prototyping — Iterate on creative directions at scale before committing to production assets.
- Design Exploration — Batch-generate variations to compare styles, compositions, and color palettes.
📝 Notes
promptis the only required field (besidesmodel).- The platform model name is
qwen-image-3.0-text-to-image; requests are dispatched upstream asqwen-image-3.0. sizeuses awidth*heightformat (asterisk); valid range is 512×512–2048×2048.- Task status values:
created,processing,completed,failed. - Generated image URLs are returned in
outputsonce status iscompleted, and are valid for about 24 hours — download promptly.
