GPT Image 1 Mini OpenAI vs Qwen Image Edit 2511 Alibaba

GPT Image 1 Mini

• + Excellent preservation of the specific car model (Rolls-Royce) features.
• + Near-perfect consistency with the subject's clothing and facial features.
• + Natural-looking warm golden hour lighting that logically hits both the car and person.

• − The steering wheel is missing its center hub and looks slightly disconnected from the column.

Qwen Image Edit 2511

• + Successfully combines both source images into the requested scene.
• + Detailed interior rendering with wooden trim and instrument cluster.

• − The car model was changed significantly from the source image.
• − The subject's hairstyle was modified to be much more chaotic and lost some of its structured volume.
• − The subject appears to be looking directly at the camera while 'driving', which looks a bit staged/unnatural.

GPT Image 1 Mini

• + Successfully integrated the character's clothing and accessories into the scene.
• + Correctly matched the yellow studio lighting and background of Image 1.
• + Created a coherent image that logically approximates the requested pose.

• − Failed to replicate the exact dynamic torso angle and head position from Image 1.
• − The foot anatomy on the red box is poorly rendered with too many toes.

Qwen Image Edit 2511

• + Maintained the exact complex pose from Image 1 perfectly via compositing.
• + Preserved the character details of Image 2 accurately by keeping him in the background.

• − Failed the primary instruction to recreate the character FROM Image 2 IN the pose from Image 1.
• − The final result is a messy overlay of two people rather than a single character recreation.
• − The lighting on the dancer does not match the new dark background.

GPT Image 1 Mini

• + Excellent preservation of the subject's face and unique vitiligo features.
• + Good adaptation of the outfit to a new full-body pose.
• + Accurate scarf pattern and jacket style inspired by the source clothing.

• − The person's head shape and skin texture were slightly altered toward a more generic bald appearance.
• − The lighting on the subject is a bit flat compared to the original beach lighting.

Qwen Image Edit 2511

• + Successfully captured the blue hue of the coat and the specific sunglasses from Image 2.
• + Maintained the original background composition and wooden structure perfectly.

• − Completely failed to preserve the source person, replacing him with the man from Image 2.
• − Ignored the core instruction to keep the person's exact face and hair unchanged.

GPT Image 1 Mini

• + Successfully added a full head of hair that matches the requested density.
• + Maintained the overall color palette and environment.

• − Significantly altered the man's facial features, making him look like a younger/different person.
• − The hairline looks slightly artificial where it meets the forehead.
• − The lighting on the face was changed compared to the source.

Qwen Image Edit 2511

• + Excellently preserved the subject's face, glasses, and the entire background.
• + The curls have a very realistic texture and integrate well with the existing sideburns.
• + Perfectly honors the request to preserve facial features and lighting.

• − Small artifact where the hair meets the top center of the forehead.

GPT Image 1 Mini

• + Excellent adherence to all prompt elements including the dog, hockey stick, and puck.
• + Preserves the subject's likeness well while applying a distinct colored pencil caricature style.
• + Maintains the denim shirt from the original photo, bridging the source and the edit.

• − The hand holding the microphone is a bit simplified/clunky.

Qwen Image Edit 2511

• + Strong caricature style with vibrant, clean digital colors.
• + Captures the 'TV anchor' profession effectively with a detailed studio background.

• − Failed to include any representation of hockey, which was a specific requirement.
• − The text in the speech bubble is gibberish.
• − The subject's facial expression is slightly more generic and less like the source image than Model A.

GPT Image 1 Mini

• + Excellent adherence to the 'hand-painted textures' and 'soft pastel colors' keywords
• + Successfully captures the specific Ghibli-esque facial stylization and warm mood
• + Preserves the central composition and poses of the original meme perfectly

• − Focus is a bit soft across the entire image, losing some crispness in line work
• − Background details from the original are heavily simplified into a generic wash

Qwen Image Edit 2511

• + High resolution with clean, sharp line art reminiscent of modern anime
• + Preserves smaller details from the source like the sunglasses on the red dress and background figures
• + Excellent lighting and color palette that feels professional and polished

• − Character faces still look a bit too much like the real people, falling slightly short of a full Ghibli transformation
• − The 'sparkle' artifacts are a bit distracting and weren't explicitly requested

GPT Image 1 Mini

• + Excellent addition of numerous falling leaves across the whole scene
• + The hair physics look natural and wind-blown
• + Strong preservation of the overall composition

• − Noticeable facial feature shifts making the subject look slightly different
• − A small artifact appeared on the leash handle

Qwen Image Edit 2511

• + Superb preservation of the original face and dog characteristics
• + Dynamic, energetic hair styling that captures the wind effect well
• + Excellent integration of leaves with varying blur to simulate motion

• − The hair has a slightly stylized, overly symmetrical flow compared to Model A