Question 1

What is Qwen Image Edit 2511?

Accepted Answer

Qwen Image Edit 2511 is a 20B parameter open-source image editing model from Alibaba, released December 2025. It uses a unique dual-path architecture: Qwen2.5-VL captures what to change (semantics), while a VAE encoder preserves what to keep (appearance). The '2511' follows YYMM naming (November 2025 version). It supports natural language edits, multi-person group photos, bilingual text editing, and identity-preserving transformations.

Question 2

How much does Qwen Image Edit 2511 cost on Runflow?

Accepted Answer

Qwen Image Edit 2511 costs $0.03 per megapixel on Runflow. A standard 1024×1024 edit costs roughly $0.031. Cold start time is not billed. Volume discounts are available.

Question 3

What kinds of edits can it perform?

Accepted Answer

Qwen Image Edit 2511 supports: object removal/addition/replacement, background swaps, style transfer, lighting and material changes, bilingual text editing (English + Chinese), pose transformation, novel view synthesis, multi-person group photo fusion, product placement, old photo restoration, and sketch-to-image conversion. All driven by natural language — no masks or region selection needed.

Question 4

How does it compare to FLUX 2 Edit?

Accepted Answer

Qwen Image Edit 2511 is more controlled and deterministic compared to FLUX 2 Edit, which can produce outputs beyond the instruction. Qwen is more affordable ($0.03/MP), has superior bilingual text editing, and better identity preservation. FLUX 2 Edit can produce higher raw image quality in some cases and supports up to 10 reference images vs Qwen's 3.

Question 5

Can I pass multiple input images?

Accepted Answer

Yes. The image_urls parameter accepts an array of 1–3 images. This enables multi-person group photo fusion, product-in-scene placement, style reference transfer, and composite editing.

Question 6

Can I use Qwen Image Edit 2511 commercially?

Accepted Answer

Yes. Qwen Image Edit 2511 is released under the Apache 2.0 license — fully open-source with no restrictions on commercial use. You can use it via Runflow's API, self-host it, or fine-tune it with custom LoRAs.

Question 7

Do I need to manage GPUs?

Accepted Answer

No. Runflow handles all infrastructure, including GPU provisioning, scaling, and failover. You call our API and receive results — no DevOps required.

Use Case	Assessment	Notes
Portrait	Pending	Portrait editing quality
Fashion	Pending	Fashion editing quality
Product	Pending	Product editing quality
Creative	Pending	Creative editing quality

Parameter	Type	Required	Description
prompt	string	required	Natural language instruction describing the edit. Examples: 'change the background to a beach', 'make the lighting warmer'.
image_urls	array[string]	required	URLs of the source images to edit. Accepts 1–3 images for reference-based or multi-image editing.
negative_prompt	string	optional	Describe what to avoid in the edited output — artifacts, unwanted styles, or quality issues.
acceleration	string	optional	Speed/quality tradeoff. 'none' for maximum quality, 'regular' for balanced, 'high' for fastest editing.
image_size	object \| string	optional	Output dimensions. Defaults to input image dimensions if not specified.
num_images	integer	optional	Number of edited variations to generate per request. Range: 1–4.
num_inference_steps	integer	optional	Number of denoising steps. 20 for fast edits, 28 for balanced, 35–50 for max fidelity. Default: 28.
guidance_scale	number	optional	Controls how closely the model follows the edit instruction. 3–5 for subtle, 6–10 for dramatic. Default: 4.5.
seed	integer	optional	Reproducibility seed. Same seed + same inputs = identical output.
output_format	string	optional	Output image format. Options: "jpeg", "png", "webp". Default: "png".
enable_safety_checker	boolean	optional	If set to true, the safety checker will be enabled.
sync_mode	boolean	optional	If true, returns the image as a data URI.

Qwen Image Edit 2511 API

Qwen Image Edit family by Alibaba