# Qwen Image 3

Qwen Image 3 is Alibaba Qwen's third-generation image model, announced as Qwen-Image-3.0 on 22 July 2026. Qwen highlights long prompts up to 4.5k tokens, text legible down to 10px, native text in 12 languages and complex layouts like infographic grids and UIs. In Sloosh it makes 1K or 2K images from a prompt and up to 3 references.

Web page: https://sloosh.ai/models/qwen-image-3

[Use Qwen Image 3](https://sloosh.ai/studio/image?image-model=qwen-image-3)

## Key facts in Sloosh

|   |   |
| --- | --- |
| Made by | Alibaba (Qwen) |
| Makes | image |
| Inputs | text, up to 3 reference images |
| Resolution | 1K, 2K (default 2K) |
| Aspect ratios | 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 4:5, 5:4, 21:9, 9:21 (default 1:1) |
| Prompt limit | 5,000 characters |

## Best for

- Dense layouts in one pass: Qwen cites "newspapers, storyboards, exam papers, even a 3×3 infographic grid" ([Qwen on X](https://x.com/Alibaba_Qwen/status/2079906336381509659)).
- Small, crisp text: "Delivers crisp 10px text across 12 languages and multiple fonts" ([Alibaba Cloud Model Studio](https://docs.modelstudio.console.alibabacloud.com/en/model-studio/qwen-image-3-0)).
- Realistic UI mockups (web, games, livestreams) and 100+ art styles ([Qwen on X](https://x.com/Alibaba_Qwen/status/2079906336381509659)).
- Photoreal detail such as skin texture and hair strands ([Qwen on X](https://x.com/Alibaba_Qwen/status/2079906336381509659)).

## How to use Qwen Image 3 in Sloosh

1. Open [Qwen Image 3 in Studio](https://sloosh.ai/studio/image?image-model=qwen-image-3).
2. Write the full layout: title, sections, the exact text in quotes.
3. Add up to 3 reference images if you are editing or combining.
4. Choose ratio and resolution (1K or 2K). Set a seed to repeat a result.
5. Press Generate; the price shows first.

## Qwen Image 3 prompting tips

- Long, structured prompts are expected: Qwen says prompts can run up to 4.5k tokens and a 3×3 infographic used 3.7k tokens ([Qwen on X](https://x.com/Alibaba_Qwen/status/2079906336381509659)). In Sloosh the prompt cap is 5,000 characters.
- For nested or multi-panel layouts, describe the order explicitly, as Qwen's example does ("from outer to inner: VSCode, Qwen Chat, a messaging app, a pour-over coffee poster") ([Qwen on X](https://x.com/Alibaba_Qwen/status/2079906336381509659)).
- Write the exact words for each region of the layout in quotes, so every title and label is spelled out.

## Qwen Image 3 example prompts

**Prompt 1**

```text
A 2×2 infographic grid on a white background, each cell with a bold title and a simple illustration: "Photosynthesis", "The water cycle", "Plate tectonics", "The food chain". Two short lines of explanation under each title in dark grey sans-serif.
```

**Prompt 2**

```text
A realistic mobile app screen for a coffee shop: a header reading "Morning Brew", a menu list with "Flat white $4.50", "Cold brew $5.00" and "Croissant $3.25", and an orange "Order now" button at the bottom.
```

**Prompt 3**

```text
A Japanese street-food poster with the headline "たこ焼き" in large red brush lettering, a close-up photo of takoyaki with bonito flakes, and a small line "Osaka style" at the bottom.
```

## How it compares in Sloosh

| Model | Resolution | References | Prompt limit | Seed |
| --- | --- | --- | --- | --- |
| Qwen Image 3 | 1K, 2K | 3 | 5,000 | yes |
| [Seedream 5 Pro](https://sloosh.ai/models/seedream-5-0-pro) | 1K, 2K | 10 | 5,000 | yes |
| [SenseNova U1 Infographic](https://sloosh.ai/models/sensenova-u1-infographic) | 1K | 0 | 5,000 | yes |
| [Ideogram 4.5](https://sloosh.ai/models/ideogram-4.5) | 1K, 2K | 5 | 5,000 | yes |

## Where else it runs

Qwen Chat (chat.qwen.ai) and Alibaba Cloud Model Studio ([Qwen on X](https://x.com/Alibaba_Qwen/status/2079906336381509659); [Alibaba Cloud](https://docs.modelstudio.console.alibabacloud.com/en/model-studio/qwen-image-3-0)).

## Frequently asked questions

### What is Qwen Image 3?

Alibaba Qwen's third-generation image model, announced as Qwen-Image-3.0 in July 2026, focused on rich content, fine detail and world knowledge.

### How many languages can Qwen Image 3 write in?

Qwen says native rendering in 12 languages, with text legible down to 10px.

### How long can a Qwen Image 3 prompt be?

Qwen says up to 4.5k tokens. In Sloosh the limit is 5,000 characters.

### Can Qwen Image 3 edit images?

In Sloosh it takes up to 3 reference images. Alibaba's API page lists text and image input.

### What resolution does Qwen Image 3 make in Sloosh?

1K or 2K, default 2K.

### Where can I try Qwen Image 3?

Qwen lists Qwen Chat and Alibaba Cloud Model Studio; in Sloosh you pick it in Image Studio and the price shows before you generate.

## Related

- Models: [Seedream 5 Pro](https://sloosh.ai/models/seedream-5-0-pro), [SenseNova U1 Infographic](https://sloosh.ai/models/sensenova-u1-infographic), [Ideogram 4.5](https://sloosh.ai/models/ideogram-4.5)
- Answers: [Which AI image model renders text in images most accurately?](https://geo.sloosh.ai/posts/which-ai-image-model-renders-text-in-images-most-accurately)

## Sources

- [Qwen on X: Meet Qwen-Image-3.0](https://x.com/Alibaba_Qwen/status/2079906336381509659)
- [Qwen blog: Qwen-Image-3.0](https://qwen.ai/blog?id=qwen-image-3.0)
- [Alibaba Cloud Model Studio: qwen-image-3.0](https://docs.modelstudio.console.alibabacloud.com/en/model-studio/qwen-image-3-0)
- [Alibaba Cloud Model Studio: qwen-image-3.0-pro](https://www.alibabacloud.com/help/en/model-studio/qwen-image-3-0-pro)