GPT Image 1.5 OpenAI vs Qwen Image Edit 2509 Alibaba

GPT Image 1.5

• + Excellent preservation of the man's facial features and specific hairstyle from the source.
• + Maintains detailed interior car textures and realistic lighting.
• + Accurately places the car on a coastal road that matches the California coastline description.

• − The perspective of the car feels slightly flattened compared to the 3D space of the road.
• − The man appears to be in the passenger seat for a left-hand drive car while holding a steering wheel that is off-center.

Qwen Image Edit 2509

• + Dynamic composition with effective use of motion blur on the wheels/road.
• + Consistent wide-angle perspective and lighting across the entire scene.
• + Excellent preservation of the specific Rolls-Royce model from the source image.

• − The man's facial details are lost due to the distant framing and shadow.
• − The man's hair structure is significantly simplified compared to the source image.

GPT Image 1.5

• + Successfully transferred the character, clothing, scarf, and sunglasses from Image 2.
• + Accurately replicated the complex leg-crossing pose and environment from Image 1.
• + High resolution with great attention to detail in the scarf texture and facial features.

• − The head orientation is upright rather than tilted, losing the exact neck angle from the source pose.
• − Minor anatomical awkwardness where the left leg joins the body.

Qwen Image Edit 2509

• + Maintained the original image's pose and aspect ratio perfectly.

• − Completely failed to incorporate the character reference from Image 2.
• − Only applied a slight stylistic filter/upscale to the original Image 1 without any requested edits.
• − Ignored all instructions regarding character, clothing, and facial features.

GPT Image 1.5

• + Successfully replicates the exact plaid pattern and texture of the scarf from Image 2.
• + Maintains high resolution and realistic fabric lighting that matches the beach environment.

• − Failed to preserve the face and head, resulting in a low-angle crop that obscures the identity.
• − Missed the sunglasses and boots requested in the prompt.

Qwen Image Edit 2509

• + Successfully includes all accessories including the sunglasses, watch, and black boots.
• + Preserves the full composition of Image 1, including the person's face and the environment.

• − Significantly altered the facial features and hair of the subject from Image 1.
• − Lower visual quality with noticeable artifacts around the sunglasses and hands.

GPT Image 1.5

• + Natural hair texture that matches the beard
• + High level of preservation for the background and clothing
• + Realistic density and hairline that blends well with existing facial features

• − The hair has a slightly uniform, 'helmet' look around the edges
• − Slightly alters the shape of the upper forehead/brow

Qwen Image Edit 2509

• + Excellent full head of hair that meets the 'thick' requirement well.
• + Preserves the original facial features and glasses perfectly.
• + Background and lighting are maintained correctly.

• − The hair texture has a slightly plastic, 'wig-like' sheen that looks over-processed.
• − The transition at the temples is a bit abrupt compared to the natural beard.

GPT Image 1.5

• + Excellent caricature style that maintains facial resemblance to the source image
• + Creative inclusion of the dogs, specifically the one with the hockey helmet and stick
• + High visual quality with realistic lighting and professional news graphics

• − The fingers holding the microphone and papers are slightly distorted

Qwen Image Edit 2509

• + Successfully preserved the denim jacket and specific selfie pose from the source image
• + Coherent caricature art style with bold line work

• − The 'dog' is extremely small and lacks the detail/humor requested
• − Significant text artifacts in the speech bubble and background screen
• − The addition of a random male co-anchor was not requested

GPT Image 1.5

• + Excellent adherence to the 'hand-painted textures' and 'dreamy' lighting requested.
• + Superior translation of character expressions into an anime-inspired style while maintaining the source material's humor.
• + Great use of soft pastel colors and atmospheric blurring to create a nostalgic mood.

• − The woman in the foreground is significantly more blurred than in the original image.

Qwen Image Edit 2509

• + Keeps the foreground subject sharp, much like the original source image.
• + Preserves the details of the background elements, like the red bus, more clearly.

• − The art style feels more like a modern digital comic or manhwa than the requested Ghibli/hand-painted aesthetic.
• − Lighting is flat and lacks the 'warm, nostalgic' atmosphere characterized by Ghibli films.

GPT Image 1.5

• + Excellent source preservation of the background and characters.
• + Realistic hair motion that fans out naturally.
• + Uses motion blur on the leaves to enhance the feeling of speed and energy.

• − Leaves look a bit generic and consistent in color.

Qwen Image Edit 2509

• + Strong directional hair movement that clearly indicates wind direction.
• + Vibrant variety of leaf shapes (maple) and colors.
• + Good preservation of the original subjects' faces and clothing.

• − The hair has a slightly stylized, wavy pattern that looks less natural than the original texture.
• − The tail and some background elements were slightly altered in unnecessary ways.