Alibaba's Qwen image editing model for instruction-based image modifications and transformations
Settled by community votes across 6 shared challenges, with an AI judge weighing in on each.
Qwen Image Edit 2509
#32 of 32 in Image Editing
Wan 2.6
#28 of 62 in Text-to-Image
Where the votes landed
Qwen Image Edit 2509
0%
win rate
Ties
0%
Wan 2.6
0%
win rate
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Man and Car in California
Editing“Make a photo of the man driving the car down the California coastline”
AI Judge Analysis
Qwen Image Edit 2509
- + Successfully placed the car and a driver in a California coastline setting.
- + Maintained the car's exterior design accurately.
- − The driver's face and features are poorly rendered and blurry.
- − The lighting on the car doesn't quite match the golden hour background.
Wan 2.6
- + Excellent preservation of the man's facial features and specific clothing (plaid coat, scarf).
- + Highly realistic composition with dynamic motion blur and beautiful scenery.
- + The car's interior and exterior are seamlessly integrated into the new environment.
- − The car is slightly cropped at the front compared to the original composition.
- − The wheel design has changed slightly from the source image.
Verdict: Wan 2.6 is the clear winner as it successfully preserved the identity of both the man and the car while placing them in a new, high-quality environment. Qwen Image Edit 2509 failed to maintain the man's specific facial features and clothing, providing a generic silhouette instead.
Outfit Transfer Challenge
Editing“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.”
AI Judge Analysis
Qwen Image Edit 2509
- + Successfully replicates the full outfit including jeans, boots, and watch.
- + Follows the pose from Image 2 to integrate the clothing naturally.
- − Failed to preserve the person's face, blending it with the man from Image 2.
- − Changed the person's body shape and proportions significantly.
- − The overall image quality is lower with over-saturated colors and harsh lighting.
Wan 2.6
- + Excellent preservation of the original person's face, skin markings (vitiligo), and sand textures.
- + Maintains the high resolution and photographic style of the source image.
- + Correctly applies the coat, scarf, and sunglasses from Image 2.
- − The scarf pattern and color are slightly modified compared to the source.
- − Does not show the full outfit (jeans and shoes) as requested, though it maintains the original framing.
Verdict: Qwen Image Edit 2509 failed the primary preservation task by blending the faces of the two men, resulting in a different person entirely. Wan 2.6 successfully dressed the original person while maintaining his identity and the image's overall quality, despite not zooming out to include the shoes.
Bald man challenge
Image Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
Qwen Image Edit 2509
- + Successfully added a large volume of hair as requested.
- + Preserved the background and clothing perfectly.
- − The hair texture has a plastic, artificial sheen and appears 'painted on' in several areas.
- − The hairline is very high and lacks a natural transition to the forehead.
- − The lighting on the new hair does not perfectly match the environmental lighting.
Wan 2.6
- + Expertly rendered hair with realistic texture and flow.
- + The hairline and sideburns integrate seamlessly with the original facial hair and skin.
- + The lighting and highlights on the hair perfectly match the existing scene's light source.
- − The forehead height was slightly reduced, subtly altering the face shape compared to the original.
Verdict: Wan 2.6 provided a much more realistic and high-quality edit, creating hair that looks genuinely natural with convincing texture and a seamless hairline. Qwen Image Edit 2509 added the requested hair volume but suffered from artificial-looking textures and a less believable integration with the subject's face.
Over-the-top cartoon caricature
Editing“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.”
AI Judge Analysis
Qwen Image Edit 2509
- + Maintains the subject's denim jacket from the source image
- + Successfully incorporates a hockey rink into the newsroom setting
- + Accurately replicates the selfie pose of the source image
- − The dog is very small and lacks detail
- − Contains nonsensical text in the speech bubble and on-screen
Wan 2.6
- + Includes multiple dogs as requested in the prompt
- + Clearly identifies the TV anchor profession with a microphone and headset
- + High-quality, clean illustrative style
- − Does not preserve the clothing or pose from the source image
- − The subject's facial features do not resemble the source woman as closely as Model A
Verdict: Qwen Image Edit 2509 is the superior editor because it preserves the subject's clothing and selfie pose from the source image while creatively blending the newsroom and hockey rink. Wan 2.6 generates a generic cartoon illustration that ignores the specific composition and attire of the original image.
Studio Ghibli Anime Style
Editing“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”
AI Judge Analysis
Qwen Image Edit 2509
- + Excellent preservation of the original pose and character expressions.
- + Uses a clean cel-shaded line art style consistent with anime aesthetics.
- + Maintains the structural details of the background and foreground characters perfectly.
- − The colors lean more toward a standard manga/webtoon style rather than the requested soft Ghibli palette.
- − The pink dress change diverges from the iconic red color of the source meme unnecessarily.
Wan 2.6
- + Expertly captures the Ghibli-esque hand-painted watercolor texture.
- + Features a dreamy, soft lighting effect with sparkling particles that enhances the nostalgic mood.
- + Preserves the original red color of the woman's dress while applying the requested art style.
- − The character faces have a more generic 'pretty' aesthetic that loses some of the original expressive nuance.
- − The background is slightly more washed out compared to the first model.
Verdict: Both models successfully transformed the 'distracted boyfriend' meme into an illustration. Qwen Image Edit 2509 is better at preserving the exact geometry and expressions of the source image, but Wan 2.6 far exceeds it in capturing the specific 'Studio Ghibli' aesthetic through its soft watercolor textures and gentle lighting effects. Wan 2.6 is the winner for better adherence to the specific stylistic prompts.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
Qwen Image Edit 2509
- + Successfully added distinct flying leaves and dynamic hair movement.
- + Maintained high image consistency with the source.
- + Added motion to the dog's tail to enhance the lively feel.
- − The hair movement looks a bit stylized and wavy rather than naturally windswept.
- − The autumn-colored leaves clash slightly with the green summer background.
Wan 2.6
- + The hair movement looks very natural and realistic.
- + Preserved the original colors and lighting of the scene perfectly.
- + Added leaves that match the colors of the existing trees.
- − The 'flying leaves' are sparse and less noticeable than in model A.
- − Did not add motion to the dog, which was part of the 'lively' request.
Verdict: Both models followed the instructions well, but Qwen Image Edit 2509 provided a more 'energetic' feel by adding more flying leaves and dynamic movement to both the woman's hair and the dog's tail. Wan 2.6 achieved a more realistic and subtle effect with the hair, but it was less transformative overall based on the prompt's request for dynamic energy.
Explore each model
Alibaba's multimodal generation model from the Wan AI suite, supporting text-to-video, image-to-video, reference-to-video with audio, and text-to-image, in both Chinese and English