Qwen-Image-Edit, instructions, edit, usage guide - Alibaba Cloud Model Studio

Qwen-Image-Edit supports multi-image input and output. It can accurately modify text in images, add, delete, or move objects, change the actions of subjects, transfer image styles, and enhance details.

Getting started

This example shows how to use qwen-image-edit-max to generate two edited images from three input images and a prompt.

Input prompt: The girl in Image 1 wears the black dress from Image 2 and sits in the pose from Image 3.

Input image 1	Input image 2	Input image 3	Output images (multiple images)

Before making a call, obtain an API key and set the API key as an environment variable.

To call the API using the SDK, install the DashScope SDK. The SDK is available for Python and Java.

The Qwen image editing models support one to three input images. qwen-image-edit-max and qwen-image-edit-plus can generate one to six images. qwen-image-edit can generate only one image. The URLs for the generated images are valid for 24 hours. Be sure to download the images to your local device promptly.

Python

import json
import os
import dashscope
from dashscope import MultiModalConversation

# The following is the URL for the Singapore region. If you use a model in the Beijing region, replace the URL with: https://dashscope.aliyuncs.com/api/v1
dashscope.base_http_api_url = 'https://dashscope-intl.aliyuncs.com/api/v1'

# The model supports one to three input images.
messages = [
    {
        "role": "user",
        "content": [
            {"image": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/thtclx/input1.png"},
            {"image": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/iclsnx/input2.png"},
            {"image": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/gborgw/input3.png"},
            {"text": "Make the girl from Image 1 wear the black dress from Image 2 and sit in the pose from Image 3."}
        ]
    }
]

# The Singapore and Beijing regions use separate API keys. Get an API key: https://www.alibabacloud.com/help/en/model-studio/get-api-key
# Replace with your API key if the environment variable is not set: api_key="sk-xxx"
api_key = os.getenv("DASHSCOPE_API_KEY")

# qwen-image-edit-max and qwen-image-edit-plus series support outputting 1 to 6 images. This example shows how to output 2 images.
response = MultiModalConversation.call(
    api_key=api_key,
    model="qwen-image-edit-max",
    messages=messages,
    stream=False,
    n=2,
    watermark=False,
    negative_prompt=" ",
    prompt_extend=True,
    size="1024*1536",
)

if response.status_code == 200:
    # To view the full response, uncomment the following line.
    # print(json.dumps(response, ensure_ascii=False))
    for i, content in enumerate(response.output.choices[0].message.content):
        print(f"URL of output image {i+1}: {content['image']}")
else:
    print(f"HTTP status code: {response.status_code}")
    print(f"Error code: {response.code}")
    print(f"Error message: {response.message}")
    print("For more information, see the documentation: https://www.alibabacloud.com/help/en/model-studio/error-code")

Response example

{
    "status_code": 200,
    "request_id": "fa41f9f9-3cb6-434d-a95d-4ae6b9xxxxxx",
    "code": "",
    "message": "",
    "output": {
        "text": null,
        "finish_reason": null,
        "choices": [
            {
                "finish_reason": "stop",
                "message": {
                    "role": "assistant",
                    "content": [
                        {
                            "image": "https://dashscope-result-hz.oss-cn-hangzhou.aliyuncs.com/xxx.png?Expires=xxx"
                        },
                        {
                            "image": "https://dashscope-result-hz.oss-cn-hangzhou.aliyuncs.com/xxx.png?Expires=xxx"
                        }
                    ]
                }
            }
        ],
        "audio": null
    },
    "usage": {
        "input_tokens": 0,
        "output_tokens": 0,
        "characters": 0,
        "height": 1536,
        "image_count": 2,
        "width": 1024
    }
}

Java

import com.alibaba.dashscope.aigc.multimodalconversation.MultiModalConversation;
import com.alibaba.dashscope.aigc.multimodalconversation.MultiModalConversationParam;
import com.alibaba.dashscope.aigc.multimodalconversation.MultiModalConversationResult;
import com.alibaba.dashscope.common.MultiModalMessage;
import com.alibaba.dashscope.common.Role;
import com.alibaba.dashscope.exception.ApiException;
import com.alibaba.dashscope.exception.NoApiKeyException;
import com.alibaba.dashscope.exception.UploadFileException;
import com.alibaba.dashscope.utils.JsonUtils;
import com.alibaba.dashscope.utils.Constants;

import java.io.IOException;
import java.util.Arrays;
import java.util.Collections;
import java.util.HashMap;
import java.util.Map;
import java.util.List;

public class QwenImageEdit {

    static {
        // The following URL is for the Singapore region. If you use a model in the Beijing region, replace the URL with https://dashscope.aliyuncs.com/api/v1.
        Constants.baseHttpApiUrl = "https://dashscope-intl.aliyuncs.com/api/v1";
    }
    
    // The Singapore and Beijing regions use separate API keys. To obtain an API key, visit https://www.alibabacloud.com/help/en/model-studio/get-api-key.
    // If you have not configured the environment variable, replace the following line with your Model Studio API key: apiKey="sk-xxx"
    static String apiKey = System.getenv("DASHSCOPE_API_KEY");

    public static void call() throws ApiException, NoApiKeyException, UploadFileException, IOException {

        MultiModalConversation conv = new MultiModalConversation();

        // The model supports one to three input images.
        MultiModalMessage userMessage = MultiModalMessage.builder().role(Role.USER.getValue())
                .content(Arrays.asList(
                        Collections.singletonMap("image", "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/thtclx/input1.png"),
                        Collections.singletonMap("image", "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/iclsnx/input2.png"),
                        Collections.singletonMap("image", "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/gborgw/input3.png"),
                        Collections.singletonMap("text", "Make the girl from Image 1 wear the black dress from Image 2 and sit in the pose from Image 3.")
                )).build();
        // qwen-image-edit-max and qwen-image-edit-plus support outputting 1 to 6 images. This example shows how to output 2 images.
        Map<String, Object> parameters = new HashMap<>();
        parameters.put("watermark", false);
        parameters.put("negative_prompt", " ");
        parameters.put("n", 2);
        parameters.put("prompt_extend", true);
        parameters.put("size", "1024*1536");

        MultiModalConversationParam param = MultiModalConversationParam.builder()
                .apiKey(apiKey)
                .model("qwen-image-edit-max")
                .messages(Collections.singletonList(userMessage))
                .parameters(parameters)
                .build();

        MultiModalConversationResult result = conv.call(param);
        // To view the complete response, uncomment the following line.
        // System.out.println(JsonUtils.toJson(result));
        List<Map<String, Object>> contentList = result.getOutput().getChoices().get(0).getMessage().getContent();
        int imageIndex = 1;
        for (Map<String, Object> content : contentList) {
            if (content.containsKey("image")) {
                System.out.println("URL of output image " + imageIndex + ": " + content.get("image"));
                imageIndex++;
            }
        }
    }

    public static void main(String[] args) {
        try {
            call();
        } catch (ApiException | NoApiKeyException | UploadFileException | IOException e) {
            System.out.println(e.getMessage());
        }
    }
}

Sample response

{
    "requestId": "46281da9-9e02-941c-ac78-be88b8xxxxxx",
    "usage": {
        "image_count": 2,
        "width": 1024,
        "height": 1536
    },
    "output": {
        "choices": [
            {
                "finish_reason": "stop",
                "message": {
                    "role": "assistant",
                    "content": [
                        {
                            "image": "https://dashscope-result-sz.oss-cn-shenzhen.aliyuncs.com/xxx.png?Expires=xxx"
                        },
                        {
                            "image": "https://dashscope-result-sz.oss-cn-shenzhen.aliyuncs.com/xxx.png?Expires=xxx"
                        }
                    ]
                }
            }
        ]
    }
}

curl

The following command uses the URL for the Singapore region. If you use a model in the China (Beijing) region, replace the URL with: https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation

curl --location 'https://dashscope-intl.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation' \
--header 'Content-Type: application/json' \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--data '{
    "model": "qwen-image-edit-max",
    "input": {
        "messages": [
            {
                "role": "user",
                "content": [
                    {
                        "image": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/thtclx/input1.png"
                    },
                    {
                        "image": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/iclsnx/input2.png"
                    },
                    {
                        "image": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/gborgw/input3.png"
                    },
                    {
                        "text": "Make the girl from Image 1 wear the black dress from Image 2 and sit in the pose from Image 3."
                    }
                ]
            }
        ]
    },
    "parameters": {
        "n": 2,
        "negative_prompt": " ",
        "prompt_extend": true,
        "watermark": false,
        "size": "1024*1536"
    }
}'

Response example

{
    "output": {
        "choices": [
            {
                "finish_reason": "stop",
                "message": {
                    "role": "assistant",
                    "content": [
                        {
                            "image": "https://dashscope-result-sz.oss-cn-shenzhen.aliyuncs.com/xxx.png?Expires=xxx"
                        },
                        {
                            "image": "https://dashscope-result-sz.oss-cn-shenzhen.aliyuncs.com/xxx.png?Expires=xxx"
                        }
                    ]
                }
            }
        ]
    },
    "usage": {
        "width": 1536,
        "image_count": 2,
        "height": 1024
    },
    "request_id": "bf37ca26-0abe-98e4-8065-xxxxxx"
}

Download images to your local device

# Install the requests library first: pip install requests
import requests


def download_image(image_url, save_path='output.png'):
    try:
        response = requests.get(image_url, stream=True, timeout=300)  # Set timeout
        response.raise_for_status()  # Raise an exception if the HTTP status code is not 200.
        with open(save_path, 'wb') as f:
            for chunk in response.iter_content(chunk_size=8192):
                f.write(chunk)
        print(f"Image downloaded successfully to: {save_path}")

    except requests.exceptions.RequestException as e:
        print(f"Image download failed: {e}")


image_url = "https://dashscope-result-sz.oss-cn-shenzhen.aliyuncs.com/xxx.png?Expires=xxx"
download_image(image_url, save_path='output.png')

import java.io.FileOutputStream;
import java.io.InputStream;
import java.net.HttpURLConnection;
import java.net.URL;
 
public class ImageDownloader {
    public static void downloadImage(String imageUrl, String savePath) {
        try {
            URL url = new URL(imageUrl);
            HttpURLConnection connection = (HttpURLConnection) url.openConnection();
            connection.setConnectTimeout(5000);
            connection.setReadTimeout(300000);
            connection.setRequestMethod("GET");
            InputStream inputStream = connection.getInputStream();
            FileOutputStream outputStream = new FileOutputStream(savePath);
            byte[] buffer = new byte[8192];
            int bytesRead;
            while ((bytesRead = inputStream.read(buffer)) != -1) {
                outputStream.write(buffer, 0, bytesRead);
            }
            inputStream.close();
            outputStream.close();
 
            System.out.println("Image downloaded successfully to: " + savePath);
        } catch (Exception e) {
            System.err.println("Image download failed: " + e.getMessage());
        }
    }
 
    public static void main(String[] args) {
        String imageUrl = "http://dashscope-result-bj.oss-cn-beijing.aliyuncs.com/xxx?Expires=xxx";
        String savePath = "output.png";
        downloadImage(imageUrl, savePath);
    }
}

Model selection recommendations

qwen-image-edit-max series: A flagship image editing model with highly stable and extensive editing capabilities. Recommended for scenarios that require high-quality editing.
qwen-image-edit-plus series: Provides powerful general-purpose editing capabilities and excels in text editing, industrial design, geometric reasoning, and character consistency.

qwen-image-edit does not support multi-image output, output resolution adjustment, or prompt rewriting. Use qwen-image-edit-plus instead.

For the models supported in each region, see Model list.

Input instructions

Input images (messages)

The messages parameter is an array that must contain a single object. This object must include the role and content properties. The role property must be set to user. The content property must include both image (one to three images) and text (one editing instruction).

The input images must meet the following requirements:

The supported image formats are JPG, JPEG, PNG, BMP, TIFF, WEBP, and GIF.
The output image is in PNG format. For animated GIFs, only the first frame is processed.
For best results, the image resolution should be between 384 and 3072 pixels for both width and height. A low resolution may result in a blurry output, while a high resolution increases processing time.
The size of a single image file cannot exceed 10 MB.

"messages": [
    {
        "role": "user",
        "content": [
            { "image": "Public URL or Base64 data of Image 1" },
            { "image": "Public URL or Base64 data of Image 2" },
            { "image": "Public URL or Base64 data of Image 3" },
            { "text": "Your editing instruction, for example: 'The girl in Image 1 wears the black dress from Image 2 and sits in the pose from Image 3'" }
        ]
    }
]

Image input order

When you provide multiple input images, their order is defined by their sequence in the array. The editing instruction must correspond to the order of the images in the content field, such as 'Image 1' and 'Image 2'.

Input image 1

Input image 2

Output image

Replace the clothes of the girl in Image 1 with the clothes of the girl in Image 2.

Replace the clothes of the girl in Image 2 with the clothes of the girl in Image 1.

Image input methods

Public URL

You can provide a publicly accessible image URL that supports the HTTP or HTTPS protocol.
Example value: https://xxxx/img.png.

Base64 encoding

Convert the image file to a Base64-encoded string and concatenate it in the following format: data:{mime_type};base64,{base64_data}.

{mime_type}: The media type of the image, which must correspond to the file format.
{base64_data}: The Base64-encoded string of the file.
Example value: data:image/jpeg;base64,GDU7MtCZz... (The example is truncated for demonstration purposes.)

For complete code examples, see Python SDK and Java SDK.

More parameters

Adjust the generation results using the following optional parameters:

n: The number of images to generate. The default value is 1. The qwen-image-edit-max and qwen-image-edit-plus series of models support generating one to six images. qwen-image-edit supports generating only one image.
negative_prompt: Describes content to exclude from the image, such as "blur" or "extra fingers". This parameter helps optimize the quality of the generated image.
watermark: Specifies whether to add a "Qwen-Image" watermark to the bottom-right corner of the image. The default value is false. The following image shows the watermark style:
seed: The random number seed. The value must be an integer from [0, 2147483647]. If this parameter is not specified, the algorithm generates a random number to use as the seed. Using the same seed value helps ensure consistent generation results.

The following optional parameters are available only for the qwen-image-edit-max and qwen-image-edit-plus series of models:

size: The resolution of the output image. The format is width*height, such as "1024*2048". The width and height can range from 512 to 2048 pixels. If this parameter is not set, the output image has a resolution close to 1024*1024 and an aspect ratio similar to the original image. If you provide multiple input images, the last image is used as the reference for the aspect ratio.
prompt_extend: Enables or disables the prompt rewriting feature. The default value is true. If enabled, the model optimizes the prompt. This feature can significantly improve the results for simple or less descriptive prompts.

For a complete list of parameters, see Qwen-Image-Edit API reference.

Overview

Multi-image fusion

Input image 1

Input image 2

Input image 3

Output image

The girl in Image 1 wears the necklace from Image 2 and carries the bag from Image 3 on her left shoulder.

Subject consistency

Input image	Output image 1	Output image 2	Output image 3
	Change to a certificate photo with a blue background. The person wears a white shirt, a black suit, and a striped tie.	The person wears a white shirt, a gray suit, and a striped tie, with one hand on the tie, against a light-colored background.	The person wears a black hoodie with "千问图像" in a thick brush font, leans against a guardrail with sunlight on their hair, and a bridge and the sea are in the background.

	Place this air conditioner in the living room, next to the sofa.	Add mist coming from the air conditioner's vent, extending to the sofa, and add green leaves.	Add the white handwritten text "自然新风畅享呼吸" at the top.

Sketch creation

Input image

Output image

Generate an image that matches the detailed shape outlined in Image 1 and follows this description: A young woman smiles on a sunny day. She wears round brown sunglasses with a leopard print frame. Her hair is neatly tied up, she wears pearl earrings, a dark blue scarf with purple star patterns, and a black leather jacket.

Generate an image that matches the detailed shape outlined in Image 1 and follows this description: An elderly man smiles at the camera. His face is wrinkled, his hair is messy in the wind, and he wears round-framed reading glasses. He has a worn-out red scarf with star patterns around his neck and is wearing a cotton-padded jacket.

Creative product generation

Input image

Output image

Make this bear sit under the moon (represented by a light gray crescent outline on a white background), holding a guitar, with small stars and speech bubbles with phrases such as "Be Kind" floating around.

Print this design on a T-shirt and a paper tote bag. A female model is displaying these items. The woman is also wearing a baseball cap with "Be kind" written on it.

A hyper-realistic 1/7 scale character model, designed as a commercial finished product, is placed on a desk with an iMac that has a white keyboard. The model stands on a clean, round, transparent acrylic base with no labels or text. Professional studio lighting highlights the sculpted details. On the iMac screen in the background, the ZBrush modeling process for the same model is displayed. Next to the model, place a packaging box with a transparent window on the front, showing only the clear plastic shell inside. The box is slightly taller than the model and reasonably sized to hold it.

This bear is wearing an astronaut suit and pointing into the distance.

This bear is wearing a gorgeous ball gown, with its arms spread in an elegant dance pose.

This bear is wearing sportswear, holding a basketball, with one leg bent.

Generate image from depth map

Input image

Output image

Generate an image that matches the depth map outlined in Image 1 and follows this description: A blue bicycle is parked in a side alley, with a few weeds growing from cracks in the stone in the background.

Generate an image that matches the depth map outlined in Image 1 and follows this description: A worn-out red bicycle is parked on a muddy path, with a dense primeval forest in the background.

Generate image from keypoints

Input image

Output image

Generate an image that matches the human pose outlined in Image 1 and follows this description: A Chinese woman in a Hanfu is holding an oil-paper umbrella in the rain, with a Suzhou garden in the background.

Generate an image that matches the human pose outlined in Image 1 and follows this description: A young man stands on a subway platform. He wears a baseball cap, a T-shirt, and jeans. A train is speeding by behind him.

Text editing

Input image

Output image

Input image

Output image

Replace 'HEALTH INSURANCE' on the Scrabble tiles with '明天会更好'.

Change the phrase "Take a Breather" on the note to "Relax and Recharge".

Input image

Output image

Change "Qwen-Image" to a black ink-drip font.

Change "Qwen-Image" to a black handwriting font.

Change "Qwen-Image" to a black pixel font.

Change "Qwen-Image" to red.

Change "Qwen-Image" to a blue-purple gradient.

Change "Qwen-Image" to candy colors.

Change the material of "Qwen-Image" to metal.

Change the material of "Qwen-Image" to clouds.

Change the material of "Qwen-Image" to glass.

Add, delete, modify, and replace

Capability	Input image	Output image
Add element		Add a small wooden sign in front of the penguin that says "Welcome to Penguin Beach".
Delete element		Remove the hair from the plate.
Replace element		Change the peaches to apples.
Portrait modification		Make her close her eyes.
Pose modification		She raises her hands with palms facing the camera and fingers spread in a playful pose.

Viewpoint transformation

Input image

Output image

Input image

Output image

Get a front view.

Face left.

Get a rear view.

Face right.

Background replacement

Input image

Output image

Change the background to a beach.

Replace the original background with a realistic modern classroom scene. In the center of the background is a traditional dark green or black blackboard. The Chinese characters "通义千问" are neatly written on the blackboard in white chalk.

Old photo processing

Capability

Input image

Output image

Old photo restoration and colorization

Restore the old photo, remove scratches, reduce noise, enhance details, high resolution, realistic image, natural skin tone, clear facial features, no distortion.

Intelligently colorize the image based on its content to make it more vivid.

Billing and rate limiting

For free quota and pricing, see Model list and pricing.

For rate limits, see Qwen (Qwen-Image).

Billing details:

Billing is based on the number of successfully generated images. Failed model calls or processing errors do not incur fees or consume the free quota.
You can enable the 'Free quota only' feature to avoid extra charges after your free quota is used up. For more information, see Free quota for new users.

API reference

For the input and output parameters of the API, see Qwen - image editing.

Error codes

If a call fails, see Error messages for troubleshooting.

FAQ

Q: What languages do the Qwen image editing models support?

A: They officially support Simplified Chinese and English. Other languages may work, but results are not guaranteed.

Q: When uploading multiple reference images with different aspect ratios, how is the output aspect ratio determined?

A: By default, the output image's aspect ratio matches that of the last image in the input array. This behavior can be overridden by specifying the parameters.size parameter.

Q: How do I view model usage?

A: For more information, see Bill query and cost management.

For more information, see Image generation FAQ.