Start building

Qwen Image 3 Edit

alibaba/qwen-image-3/edit

Alibaba Qwen Image 3 image editing model with support for up to three reference images.

Use in agentGet API key
Base price
$0.05USD / run
Execution
async
Model type
image
Input fields
5

Try the model

Playground

Open playground
Input
The editing instruction describing what changes to make.

PNG, JPEG, WebP, or GIF · 20 MiB maximum each · up to 3

List of input image URLs for editing. Supports 1 to 3 images.
The aspect ratio of the generated image. Allowed values: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16.
The format of the generated image. Allowed values: jpeg, png.
Random seed for reproducibility (0-2147483647).
OutputReady

Your output will appear here

Complete the inputs, then click Run.

Specifications

Pricing

Base price
$0.05 / run
Billing formula
0.042 + len(params.images ?? []) * 0.00315

Context & modalities

Input
Schema-defined
Output
image

Capabilities

Chat
Not supported
Vision
Not supported
Reasoning
Not supported
Structured output
Not supported
Function calling
Not supported
Audio input
Not supported

Access

Provider
Alibaba
Model ID
alibaba/qwen-image-3/edit
Execution
async
API
Unified Run API
Endpoint
/v1/run

API README

Qwen Image 3 Edit

Qwen Image 3 Edit is an image-editing model for applying precise written changes to one or more supplied visuals. It is designed to understand both the content of the reference and the requested transformation, so subjects, layout, style, and textual elements can be revised without rebuilding the entire image from scratch.

The model is useful for product variations, poster updates, localized typography, scene restyling, object-level changes, and iterative art direction. State what must change and what must remain fixed; concrete wording about position, scale, lighting, color, and text placement gives the editing process a clearer target.

Highlights

  • Instruction-driven image editing. Applies a written transformation while the source image continues to anchor the result.
  • Complex layout understanding. Handles compositions with several regions, subjects, or design elements that must remain coordinated.
  • Small and multilingual text. Supports detailed typography-oriented edits, including compact text and multiple writing systems.
  • Photographic detail preservation. Retains fine visual structure while changing selected content, style, or presentation.

Pricing

ChargeBilling unitPrice
Base edit requestPer request$0.042
Reference-image processingPer supplied image$0.00315

When to Use

✅ Good fit❌ Consider alternatives
The project needs revising an existing visual assetThe goal is a different media task or endpoint
The available inputs match the required local schemaRequired source media or permissions are unavailable
The brief can specify subject, composition, style, and deliveryThe result must be deterministic at pixel or sample level
The supported formats and controls match final placementDelivery requires unsupported dimensions, codecs, or duration
An asynchronous generated result fits the workflowA live, frame-synchronous, or real-time response is mandatory

Prompt Guide

Write the request as a production brief: identify the main subject or source material, state the intended transformation, describe composition or timing, and finish with style, atmosphere, and delivery constraints. Keep preservation requirements separate from requested changes, and use only fields exposed by this route.

{
  "prompt": "用途：precise-object-edit。图像1是唯一编辑目标。这是中国高中数学试卷批改真实性测试。\n严格保留原图的木桌、相机视角、纸张轮廓与阴影，以及全部黑色印刷文字、题号、公式、图形和版式；禁止重绘、改写、移动或删除原试卷内容。只新增以下红色教师签字笔笔迹：\n1. 在试卷右上角写“112/150”并用红笔圈住。\n2. 在四个部分标题旁依次写“45/60”“15/20”“10/20”“42/70”。\n3. 仅在第3、14、25题题号旁画红色叉号；仅在第1、13、21题题号旁画红色勾号。\n4. 在第25题旁写“计算错误”，在第22题旁写“过程不完整”。\n红笔必须像中国高中教师真实批卷：自然手写、轻微倾斜、压感和墨迹浓淡变化，位置准确且不遮挡题干。不要添加学生答案、手、人物、文具、印章、水印或其他内容。除指定红笔内容外，画面保持不变。",
  "images": [
    "https://example.com/reference.jpg"
  ],
  "seed": 20260724,
  "aspect_ratio": "21:9",
  "output_format": "jpeg"
}

Technical Specs

SpecValue
Model IDalibaba/qwen-image-3/edit
Input fieldsseed (integer)<br>images (array)<br>prompt (string)<br>aspect_ratio (string; 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16)<br>output_format (string; jpeg, png)
Required inputprompt, images
Output fieldsurl, content_type
ExecutionAsynchronous job

Related Models

  • alibaba/happy-horse/video-edit
  • alibaba/wan/2.2/image-to-video/turbo
  • alibaba/wan/2.1/text-to-video

Start building

Send your first request

OpenAI-compatible endpoint with unified authentication and usage tracking.

Production API
Unified Run API endpoint
https://api.sandbase.ai/v1/run
Model ID
alibaba/qwen-image-3/edit
# The quoted heredoc keeps Unicode and shell metacharacters unchanged.
result=$(curl --fail-with-body --silent \
  -X POST "https://api.sandbase.ai/v1/run" \
  -H "Authorization: Bearer $SANDBASE_API_KEY" \
  -H "Content-Type: application/json" \
  --data-binary @- <<'SANDBASE_JSON'
{
  "model": "alibaba/qwen-image-3/edit",
  "seed": 20260724,
  "images": [
    "https://media.sandbase.ai/files/3b376bfd-65d4-4653-817b-51098dee1045.png"
  ],
  "prompt": "用途：precise-object-edit。图像1是唯一编辑目标。这是中国高中数学试卷批改真实性测试。\n严格保留原图的木桌、相机视角、纸张轮廓与阴影，以及全部黑色印刷文字、题号、公式、图形和版式；禁止重绘、改写、移动或删除原试卷内容。只新增以下红色教师签字笔笔迹：\n1. 在试卷右上角写“112/150”并用红笔圈住。\n2. 在四个部分标题旁依次写“45/60”“15/20”“10/20”“42/70”。\n3. 仅在第3、14、25题题号旁画红色叉号；仅在第1、13、21题题号旁画红色勾号。\n4. 在第25题旁写“计算错误”，在第22题旁写“过程不完整”。\n红笔必须像中国高中教师真实批卷：自然手写、轻微倾斜、压感和墨迹浓淡变化，位置准确且不遮挡题干。不要添加学生答案、手、人物、文具、印章、水印或其他内容。除指定红笔内容外，画面保持不变。",
  "aspect_ratio": "3:4",
  "output_format": "png"
}
SANDBASE_JSON
)
run_id=$(printf '%s' "$result" | jq -r .id)
for attempt in $(seq 1 120); do
  status=$(printf '%s' "$result" | jq -r .status)
  case "$status" in completed|failed|timeout) break ;; esac
  sleep 2
  result=$(curl --fail-with-body --silent \
    -H "Authorization: Bearer $SANDBASE_API_KEY" \
    "https://api.sandbase.ai/v1/run/$run_id")
done
status=$(printf '%s' "$result" | jq -r .status)
[ "$status" = completed ] || { echo "Generation ended: $status" >&2; exit 1; }
printf '%s\n' "$result"

Choose your model

Compare models

All language models
ModelContextInput / 1MOutput / 1MReleased
Qwen Image 3 EditThis model
Alibaba		———Jul 21, 2026
Qwen Image 3
Alibaba		———Jul 21, 2026
Qwen Image 3 Text to Image
Alibaba		———Jul 21, 2026
Head Swap
Alibaba		———Jun 22, 2026
Face Swap
Alibaba		———Jun 22, 2026
Wan 2.7 Pro Edit
Alibaba		———Apr 1, 2026