> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.empiriolabs.ai/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.empiriolabs.ai/_mcp/server. # Qwen Image 3.0 > Qwen image generation and editing with Base and Pro variants, multilingual typography, multi-image references, and controllable 1K or 2K output. ![Qwen Image 3.0](https://media.empiriolabs.ai/model-logos/qwen.png) [Alibaba Cloud](/providers/alibaba) · Image Generation `POST /v1/images/generations` Qwen image generation and editing with Base and Pro variants, multilingual typography, multi-image references, and controllable 1K or 2K output. ## At a glance | Field | Value | | ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Model id | `qwen-image-3-0` | | Model release date | 2026-07-21 | | Input modalities | Text, Image | | Output modalities | Image | | Context window | - | | Weight precision | - | | Region | Singapore | | Features | image\_generation, image\_editing, multi\_image, text\_rendering, multilingual\_text, dense\_layouts, photorealism | | Native inference | No | | New | Yes | | Supported endpoints | `POST /v1/images/generations` | | Alternate model ids | `qwen-image-3.0`, `qwen-image-3-0-pro`, `qwen-image-3.0-pro`, `alibaba/qwen-image-3-0`, `alibaba/qwen-image-3.0`, `alibaba/qwen-image-3-0-pro`, `alibaba/qwen-image-3.0-pro` | ## Pricing | Charge | Spec | Rate | | ---------------------------- | --------------- | ------- | | Base Output (1K or 2K) | per image | \$0.030 | | Pro Output up to 2.25MP (1K) | per image | \$0.040 | | Pro Output above 2.25MP (2K) | per image | \$0.075 | | Input Image (1K or 2K) | per input image | \$0.003 | ## Example request ```bash curl https://api.empiriolabs.ai/v1/images/generations \ -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \ -H 'Content-Type: application/json' \ -d '{"model": "qwen-image-3-0", "prompt": "a misty forest at dawn", "size": "1024x1024"}' ``` ## Parameters | Parameter | Type | Required | Default | Description | | -------------------- | ------- | -------- | ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `prompt` | string | no | - | Describe the image to create. With input images attached, this becomes the editing instruction. An image-only edit uses a neutral preservation prompt. | | `model_variant` | enum | no | `"qwen-image-3.0-pro"` | Choose Pro for stronger photorealism, dense layouts, fine multilingual typography, and complex edits, or Base for lower-cost generation. · Allowed: `qwen-image-3.0-pro`, `qwen-image-3.0` | | `image` | array | no | - | One to three reference images as URLs, data URIs, or uploaded image objects. | | `images` | array | no | - | Plural alias for multiple uploaded image references. | | `aspect_ratio` | enum | no | `"auto"` | Optional output aspect ratio. Auto is the default. With a fixed 1K or 2K resolution, Auto uses that tier's square size. With Resolution set to Auto, the model chooses the complete output size. · Allowed: `auto`, `1:1`, `16:9`, `9:16`, `3:2`, `2:3`, `4:3`, `3:4`, `5:4`, `4:5` | | `resolution` | enum | no | `"2K"` | Output pixel tier. Defaults to 2K, the maximum supported tier. The 1K tier stays at or below 2.25 megapixels, while 2K is above 2.25 megapixels. Auto lets the model choose when Aspect Ratio is Auto; with a selected aspect ratio, Auto uses the 1K preset. · Allowed: `auto`, `1K`, `2K` | | `size` | string | no | - | Exact output size as WIDTH*HEIGHT, for example 2048*1536. This overrides aspect\_ratio and resolution. | | `use_custom_size` | boolean | no | false | Use custom\_width and custom\_height instead of the aspect-ratio and resolution presets. | | `custom_width` | number | no | - | Custom output width in pixels. Total area must be between 512x512 and 2048x2048, with an aspect ratio from 1:8 to 8:1. · Range: 256 – 2048 | | `custom_height` | number | no | - | Custom output height in pixels. Total area must be between 512x512 and 2048x2048, with an aspect ratio from 1:8 to 8:1. · Range: 256 – 2048 | | `num_images` | number | no | `1` | Number of output images. Each output is billed separately. · Range: 1 – 6 | | `negative_prompt` | string | no | `""` | Describe visual elements or qualities to avoid in the output. | | `prompt_extend` | boolean | no | true | Expand the prompt automatically for stronger composition and detail. | | `prompt_extend_mode` | enum | no | `"direct"` | Direct works for generation and editing. Agent performs deeper prompt refinement and is available for text-to-image requests only. · Allowed: `direct`, `agent` | | `seed` | number | no | - | Reproducibility seed. Leave empty for a random result. · Range: 0 – 2147483647 | | `watermark` | boolean | no | false | Add the Qwen provenance watermark to the generated image. EmpirioLabs does not add a watermark of its own. | | `response_format` | enum | no | `"url"` | Return signed image URLs by default, or include base64 image data when b64\_json is requested. · Allowed: `url`, `b64_json` | ## Notes Qwen Image 3.0 combines Base and Pro image generation and editing in one model. Pro is selected by default, output starts at 2K, and aspect ratio starts in Auto. Base provides a lower-cost option for fast iteration. **Variants** * Choose qwen-image-3.0-pro for photorealism, dense layouts, fine multilingual typography, and complex multi-image edits. * Choose qwen-image-3.0 for lower-cost generation and routine edits. **Modes** * Text to image from a prompt. * Image editing and reference-guided generation from one to three images. * Direct prompt extension works for generation and editing. Agent prompt extension provides deeper refinement for text-to-image requests only. **Output and input limits** * One to six output images per request. * 1K, 2K, automatic, aspect-ratio presets, or an exact custom size. * Preset and custom sizes stay between 512x512 and 2048x2048 pixels, with each selected side up to 2048 pixels and aspect ratio from 1:8 to 8:1. Automatic sizing can choose its own dimensions within the 1K or 2K output tiers. * Input images can be JPG, JPEG, PNG, BMP, TIFF, WEBP, or GIF, up to 10 MB each. * Prompts support up to 4,500 tokens. **Billing** * Base uses one output rate across 1K and 2K. * Pro outputs up to 2.25 megapixels use the 1K rate. Larger Pro outputs use the 2K rate. * Every reference image adds the input-image rate shown above. **Performance** * 1K generations are optimized for interactive iteration. Complex 2K edits can take several minutes. **Content safety** Generated images pass an automated content-safety check. An image flagged as explicit is not returned. Because the image is generated before the check runs, a flagged generation is still billed. --- *Machine-readable schema:* `GET https://api.empiriolabs.ai/v1/models/qwen-image-3-0`.