Alibaba's Qwen image editing model for instruction-based image modifications and transformations
Settled by community votes across 7 shared challenges, with an AI judge weighing in on each.
Qwen Image Edit 2511
20.2
arena score
#16 of 32 in Image Editing
Skill signature
· Image Editing
Seedream 4.0
24.7
arena score
#15 of 62 in Text-to-Image
Vote tally
Where the votes landed
Qwen Image Edit 2511
88.6%
win rate
Ties
0.0%
Seedream 4.0
11.4%
win rate
88.6%
0.0% ties
11.4%
Shared challenges
7
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Man and Car in California
EditingEdit instruction
“Make a photo of the man driving the car down the California coastline”
Source
Qwen Image Edit 2511
Seedream 4.0
98% wins
0% ties
2% wins
AI Judge Analysis
Qwen Image Edit 2511
- + Excellent preservation of the subject's face and hair details
- + Perfectly recreates the specific plaid pattern of the coat
- + High visual quality with realistic lighting and interior details
- − Changes the car model internally (dashboard/steering wheel no longer match a Rolls-Royce)
Seedream 4.0
- + Strongly preserves the exterior model of the original car (Rolls-Royce Phantom Drophead Coupe)
- + Excellent composition capturing the movement of the car on the coastline
- + Good integration of the driver into the scene
- − Loss of fine detail in the subject's face and hair
- − The plaid pattern of the coat is significantly simplified and altered
Verdict: This is a close competition between two different approaches. Qwen Image Edit 2511 focuses on preserving the man's identity and clothing with incredible accuracy, although it sacrifices the specific car model's interior details. Seedream 4.0 does a better job of preserving the car's exterior identity and the sense of motion, but lose the likeness of the man and the detail of his outfit. Qwen Image Edit 2511 is the winner for its superior rendering of the subject's face and textures.
Pose & Character Mashup
EditingEdit instruction
“Use Image 1 as the exact pose reference and Image 2 as the character reference. Recreate the person/character from Image 2 in the exact dynamic pose and body position from Image 1. Keep the exact face, hair, clothing style/details, and expression from Image 2. Match the lighting and environment of Image 1. The final image must show the character from Image 2 performing the precise action/pose from Image 1 with perfect anatomy and natural integration.”
Source
Qwen Image Edit 2511
Seedream 4.0
AI Judge Analysis
Qwen Image Edit 2511
- + Retains the background from Image 2 correctly.
- + Maintains high resolution for the character components.
- − Failed the edit instruction completely by simply overlaying both characters.
- − Did not transfer the character features from Image 2 onto the pose from Image 1.
- − Anatomy of the woman is broken where she overlaps with the man.
Seedream 4.0
- + Successfully attempted to combine the character elements from Image 2 with the pose from Image 1.
- + Accurately recreates the yellow background and red ottoman from the pose reference.
- + Incorporates the clothing style (black sweatshirt, scarf) from the character reference.
- − The face is severely distorted and loses the likeness of the man in Image 2.
- − The pose is not an exact match to Image 1, particularly the leg positioning.
- − The hair is a messy blend of the two subjects rather than a clean recreation of the character.
Verdict: Qwen Image Edit 2511 failed the fundamental task of transferring a character to a pose, instead producing a simple collage where both original subjects are visible. Seedream 4.0 followed the instructions much better by attempting to merge the character's identity and clothing with the target pose and environment, although the facial rendering is poor and the anatomy is slightly altered. Seedream 4.0 is the winner for actually performing the requested image edit rather than creating a composite.
Outfit Transfer Challenge
EditingEdit instruction
“Use Image 1 as the base person. Dress them in the exact elaborate outfit from Image 2 (including all layers, accessories, jewelry, and shoes). Carefully adapt the clothing to the body shape and pose in Image 1 while maintaining realistic fabric behavior, correct proportions, and perfect lighting/shadow matching. Keep the person’s exact face, hair, and background completely unchanged.”
Source
Qwen Image Edit 2511
Seedream 4.0
AI Judge Analysis
Qwen Image Edit 2511
- + Successfully replicates the outfit style, colors, and pattern from Image 2.
- + Matches the background and lighting reasonably well.
- + Includes shoes that were not visible in the full-body source of the person.
- − Completely failed to preserve the source person's identity, face, and unique skin features (vitiligo).
- − Replaced the subject's face with the face of the person from Image 2.
Seedream 4.0
- + Perfectly preserves the source person's unique face, vitiligo, and hair pattern.
- + Accurately applies the outfit including the specific scarf pattern, coat, and accessories like the watch.
- + Maintains the original background and pose with high fidelity.
- − The facial hair (moustache/goatee) from the person in Image 2 was erroneously applied over the source person's face.
- − Minor artifacts around the edges of the hair.
Verdict: Qwen Image Edit 2511 failed the primary objective of source preservation, replacing the subject's face and identity entirely with the person from the clothing reference image. Seedream 4.0 successfully transferred the complex outfit while keeping the original person's distinctive skin and facial features, only failing slightly by including some of the reference's facial hair.
Bald man challenge
Image EditingEdit instruction
“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
Qwen Image Edit 2511
Seedream 4.0
83% wins
0% ties
17% wins
AI Judge Analysis
Qwen Image Edit 2511
- + Excellent texture and realistic hair volume
- + Very natural integrated hairline
- + Perfectly preserved the background and lighting
- − Subtle change to the eye/brow region makes the face look slightly different from the source
Seedream 4.0
- + Successfully added a full head of hair
- + Maintained the face shape and character well
- − Hair texture looks slightly fuzzy or synthetic compared to the sharp beard
- − The hairline is very straight and lacks a natural recession or transition
- − Lower technical resolution on the hair fibers
Verdict: Qwen Image Edit 2511 produced a much more realistic result with convincing hair texture and a believable hairline, though it slightly softened some facial details. Seedream 4.0 successfully followed the prompt but the hair appears somewhat like a wig due to its uniform texture and lack of a natural transition at the forehead.
Over-the-top cartoon caricature
EditingEdit instruction
“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
Source
Qwen Image Edit 2511
Seedream 4.0
AI Judge Analysis
Qwen Image Edit 2511
- + Strong caricature art style with an exaggerated bobblehead look
- + Includes an anchor desk setting and a dog clearly
- + Good preservation of the subject's facial features in a cartoon format
- − Completely misses the 'hockey' requirement
- − Nonsensical text in the speech bubble
- − The candle on the desk is an odd, irrelevant detail
Seedream 4.0
- + Includes all elements: anchor desk, dog, and hockey sticks/jerseys
- + Effectively blends the caricature into a realistic environment
- + Excellent preservation of the source image's clothing style and background elements
- − The hand holding the phone is slightly deformed
- − The facial exaggeration is a bit less 'humorous' and more slightly distorted compared to Model A
Verdict: Seedream 4.0 is the clear winner as it successfully incorporated all three required elements (news anchor, dogs, and hockey), whereas Qwen Image Edit 2511 completely forgot the hockey theme. Seedream 4.0 also did a better job of maintaining the source image's context by keeping the denim shirt and living room background while layering in the caricature and professional props.
Studio Ghibli Anime Style
EditingEdit instruction
“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
Source
Qwen Image Edit 2511
Seedream 4.0
40% wins
0% ties
60% wins
AI Judge Analysis
Qwen Image Edit 2511
- + Excellent preservation of the original subjects' facial features and expressions
- + Accurate reproduction of the clothing patterns, specifically the plaid shirt
- + Maintains the depth and complexity of the urban background
- − The art style leans more towards generic digital manhwa/anime rather than the specific Studio Ghibli aesthetic
- − The addition of sparkles feels a bit distracting and not characteristic of the requested style
Seedream 4.0
- + Perfectly captures the Studio Ghibli 'watercolor' and hand-painted texture look
- + Uses the soft pastel palette and dreamy lighting requested in the prompt
- + Excellent translation of the characters into the simplified anime character designs associated with the studio
- − Sacrifices the specific facial identities of the people in the original image for generic character faces
- − The background detail is almost entirely lost in favor of a watercolor wash
Verdict: Seedream 4.0 followed the stylistic instructions much more accurately, producing a result that truly feels like a Studio Ghibli watercolor illustration, despite losing the likeness of the original subjects. Qwen Image Edit 2511 preserved the original image details and identities much better but failed to fully capture the distinctive Ghibli art style, resulting in a more generic digital illustration.
Golden Hour Stroll
Image EditingEdit instruction
“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
Qwen Image Edit 2511
Seedream 4.0
56% wins
0% ties
44% wins
AI Judge Analysis
Qwen Image Edit 2511
- + Excellent adherence to the 'hair blowing in the wind' instruction
- + Bright and colorful leaves add high energy to the scene
- + Perfectly preserves the woman's face and original clothing
- − Hair strands look a bit stiff and vectorized rather than natural
- − The dog's tail has a slightly awkward jagged edge after the edit
Seedream 4.0
- + The hair motion looks more realistic and natural
- + Good overall preservation of the source image colors and lighting
- − The woman's facial features were altered, losing the likeness of the original
- − The flying leaves are subtle and less energetic compared to the other model
Verdict: Qwen Image Edit 2511 followed all instructions perfectly, creating a high-energy scene while maintaining the exact likeness of the original subject. Seedream 4.0 provided more natural-looking hair movement, but it failed to preserve the woman's face correctly and the overall impact of the added motion was less dynamic.
Next steps
Explore each model
ByteDance's image generation model with integrated text-to-image and image editing capabilities in a unified architecture, supporting up to 4K resolution