image model · by Alibaba Qwen
Qwen-Image-Edit-Plus AI Image Editor
Edit images with surgical precision using Qwen-Image-Edit-Plus - Alibaba's 20-billion-parameter MMDiT editing model. Multi-image input, bilingual text editing, identity-preserving edits, and the kind of accuracy that turns AI editing into a real production tool.
Edited with Qwen-Image-Edit-Plus
A handful of edit-style samples produced inside Fliki. Outfit composite, product placement, pose transfer, identity-preserved scene, style transfer, multi-element composite, brand tile.

Prompt

Prompt

Prompt

Prompt

Prompt

Prompt

Prompt
Trusted by 50,000+ companies worldwide
What Qwen-Image-Edit-Plus delivers
Multi-image editing (1-3 inputs)
Provide one to three input images and a single prompt to compose them into a unified output. Place a person in a different scene wearing different clothes; combine product, model, and backdrop in one shot - all from a single instruction.
Bilingual text editing
Modify existing text inside images, add new copy, or translate between English and Chinese while preserving font, size, color, and layout. This is unique among open-source edit models and a major reason agencies pick Qwen for design-driven workflows.
Identity-preserving edits
Faces, product details, and brand elements stay intact through edits. Change a background, swap an outfit, or restyle a portrait without losing facial identity or product accuracy.
Dual-stream architecture
Qwen-Image-Edit-Plus uses Qwen2.5-VL for high-level semantic understanding (object identity, relationships, scene context) and a VAE encoder for low-level visual appearance (color, texture, lighting). The dual stream is what enables both semantic and pixel-level edits.
Up to six output variations per request
Generate one to six edited variations from a single prompt. Useful for A/B testing creative directions and giving teams multiple options on the first pass.
Style transfer and IP variations
Transform a photo into anime, oil painting, watercolor, or any artistic style. Generate IP variations with a consistent character across scenes - useful for branded content and serialized creative.
ControlNet support
Native ControlNet support (depth, edge, keypoint) gives art directors pixel-level control for local adaptations. Useful when an edit has to match strict design specs or reference geometry.
Apache 2.0 open source
Like Qwen Image, the edit model is released under Apache 2.0 with full commercial use rights. On Fliki, you get hosted access without infrastructure overhead.
September 2025 quality upgrade
Qwen-Image-Edit-Plus is the qwen-image-edit-2509 release - a meaningful step up from the original Qwen-Image-Edit on consistency, multi-image fusion, and text editing benchmarks. The Qwen team keeps shipping updates to the same model name on Fliki, so you stay on the latest weights without changing your prompt.
How it works
How to edit an image with Qwen-Image-Edit-Plus
Qwen-Image-Edit-Plus runs inside Fliki's AI image generator. Here's the six-step flow.

Write your prompt
Describe the change clearly. The model excels at specific instructions: "The girl in image 1 wears the dress from image 2 and stands in the pose from image 3." For text edits, specify the new text exactly as you want it rendered.

Select Qwen-Image-Edit-Plus as your model
Open Fliki's model selector and choose Qwen-Image-Edit-Plus. Your request routes to Alibaba's 20B-parameter MMDiT edit model.

Pick your aspect ratio
Default to the input image's aspect ratio or pick a different one (1:1, 16:9, 9:16, 4:3, 3:4) for the output.

Upload reference images
Required - upload between 1 and 3 input images (widths and heights between 512 and 2048 px per dimension). Use references to lock characters, products, or styles across the edit, or compose multiple inputs into a single coherent shot.

Set the resolution
Output is supported between 512 and 2048 pixels per side. Use lower resolutions for fast iteration; bump up for final delivery.

Generate
Hit Generate. Qwen-Image-Edit-Plus is paid-only on Fliki - upgrade your plan to access it alongside the rest of the premium catalog.
AI MODEL GALLERY
Built on the best AI models - ready inside Fliki
Every leading video, voice, and image model - integrated, unified, and tuned for creators. Generate with the latest AI video, AI voice, and AI image models from OpenAI, Google, Kling, Bytedance, ElevenLabs, and more - all from one place.






















Qwen-Image-Edit-Plus FAQ
Frequently asked questions
Everything you need to know about editing images with Qwen-Image-Edit-Plus inside Fliki.
What is Qwen-Image-Edit-Plus?
Qwen-Image-Edit-Plus (qwen-image-edit-2509) is Alibaba's 20-billion-parameter MMDiT image editing model. It supports multi-image input (1-3 images), bilingual English-Chinese text editing, identity preservation, and up to six output variations per request.
Is Qwen-Image-Edit-Plus free on Fliki?
No. Only Z Image Turbo and Flux 2 Klein are on Fliki's free tier. Qwen-Image-Edit-Plus requires a paid subscription.
How is Qwen-Image-Edit-Plus different from Qwen-Image-Edit?
Qwen-Image-Edit-Plus (the 2509 release) is the current Plus tier - it adds multi-image editing, multi-output generation, ControlNet support, and improved text editing on top of the base Qwen-Image-Edit model. It's the production-ready version most teams use.
Can it edit text inside images?
Yes. Bilingual text editing is one of its standout capabilities - modify, add, translate, or replace text in English and Chinese while preserving font, size, and style.
How many input images does it accept?
One to three input images per request. Use one for single-image edits, two or three for multi-image composition workflows.
What is its primary architecture?
A dual-input design - Qwen2.5-VL for semantic understanding plus a VAE encoder for visual appearance - feeding into the 20B-parameter MMDiT diffusion core. This dual-stream approach is what enables both high-level semantic edits and low-level pixel-precise modifications.
Still curious?
Try Fliki free in your browser, no credit card required.
Start freeHow does Qwen-Image-Edit-Plus compare to P-Image-Edit?
Qwen-Image-Edit-Plus has higher fidelity, larger model size (20B vs Pruna's smaller compressed model), and stronger bilingual text editing. P-Image-Edit is faster and more cost-efficient. Pick Qwen for hero edits and text-heavy work; pick P-Image-Edit for high-volume production.
Is it open source?
Yes. Qwen-Image-Edit-Plus is released under Apache 2.0, with full commercial use rights. Fliki provides hosted API access without managing weights or infrastructure.
AI image models
Discover more
Tools
Discover features
Edit your next image with Qwen-Image-Edit-Plus.
Upgrade your Fliki plan to unlock Qwen-Image-Edit-Plus and the rest of the premium catalog.
Upgrade to editFree forever plan · No credit card required · Cancel anytime