P-Image Edit PrunaAI vs Qwen Image Edit 2509 Alibaba

P-Image Edit

• + Excellent preservation of the car's original design and proportions
• + Highly realistic motion blur on the road and wheels
• + Correct identification of the subject's hairstyle from the second source image

• − The driver is positioned a bit low in the seat, making him hard to see
• − Slightly less vibrant color palette compared to model B

Qwen Image Edit 2509

• + Creative use of hair movement to imply wind while driving
• + Vibrant and cinematic lighting and color grading
• + Dynamic composition with a strong sense of speed

• − Minor distortion of the car's body length and rear proportions
• − Artifacts on the road surface lines and a less realistic motion blur

P-Image Edit

• + Successfully integrated character features like face, sunglasses, and scarf.
• + Maintained the dynamic pose from Image 1.
• + Correctly swapped the clothing style to match Image 2's black outfit.

• − Significant anatomy errors, including a foot emerging from the head area and distorted limbs.
• − Added long feminine hair that contradicts the character from Image 2.
• − Poor blending between different photographic elements resulting in a 'cut-and-paste' look.

Qwen Image Edit 2509

• + Perfect preservation of the original Image 1 background and lighting.

• − Failed the core task of replacing the character from Image 1 with the one from Image 2.
• − Produced an almost identical copy of Image 1 with zero character editing.
• − Did not incorporate the scarf, sunglasses, or clothing from Image 2.

P-Image Edit

• + Successfully extracted and applied the coat and scarf layers while maintaining the subject's face.
• + Accurately identifies the need for jewelry, adding a chunky gold necklace.
• + Good preservation of the background and wooden structure.

• − Included bright yellow/blue shoes that were not present in either source image.
• − The subject's facial features and vitiligo patterns changed significantly compared to Image 1.
• − The scarf pattern is simplified compared to the source.

Qwen Image Edit 2509

• + Successfully applied the sunglasses from Image 2 to the subject's face.
• + Maintains the pose and slouching lean from Image 2 more effectively than Model A.
• + Higher fidelity reproduction of the plaid scarf pattern.

• − The person's face was completely altered to look like a fusion of both men rather than keeping the base person.
• − The shoes are generic black boots not found in the source image.
• − Overall image suffers from high contrast and loss of detail in shadow areas.

P-Image Edit

• + Excellent preservation of the original facial features and environment
• + Naturally rendered curly hair texture
• + Perfect lighting match with the source image

• − The hairline on the forehead looks slightly superimposed rather than integrated

Qwen Image Edit 2509

• + Natural integration of the hairline and sideburns
• + Dynamic hair volume that fits the character's style
• + Accurately preserved the jacket and background

• − Subtle smoothing effect on the skin of the forehead compared to the original
• − Slightly glossy/synthetic look to some hair strands

P-Image Edit

• + Excellent caricature translation that feels professional and polished.
• + Cleverly integrates all three requirements: TV news desk, hockey rink backdrop, and dog accessories.
• + Maintains a high level of facial resemblance to the source image within the artistic style.

• − The dog element is a bit small and cluttered with an odd red object on its back.
• − The hand holding the microphone has perspective issues with the fingers.

Qwen Image Edit 2509

• + Successfully captures the 'exaggerated and humorous' aspect of a caricature with bold expressions.
• + Preserves the selfie-style framing and the specific clothing from the source image better than Model A.
• + Includes all elements: hockey, TV environment, and a dog.

• − The facial exaggeration borders on grotesque rather than charming.
• − The text in the speech bubble is nonsensical gibberish.
• − The dog is very small and lacks detail compared to the rest of the scene.

P-Image Edit

• + Excellent preservation of the original subjects' facial structures and expressions.
• + Strong adherence to the requested 'dreamy background' with soft, painterly clouds.
• + Keeps the specific plaid pattern of the man's shirt very close to the source.

• − Art style leans more toward generic modern webtoon/manhwa than specific Ghibli aesthetics.
• − The background changes the urban setting of the original into a park, which was not explicitly requested.

Qwen Image Edit 2509

• + Expertly captures the Ghibli character design style, including eye shapes and soft shading.
• + Preserves the original urban background elements (street, bus) while applying the requested style.
• + Perfectly captures the 'warm, nostalgic' lighting and pastel color palette requested.

• − Faces are generalized into the anime style, losing some of the specific likeness of the source individuals.
• − Minor line work jitter on the plaid shirt pattern.

P-Image Edit

• + Excellent preservation of the subject's face and original features.
• + Highly realistic integration of flowing hair that maintains the texture of the original hair.

• − The 'flying leaves' appear as small, indistinct orange and yellow specks rather than detailed leaves.
• − Lower count of leaves compared to the other model results in slightly less 'energetic' feel.

Qwen Image Edit 2509

• + Successfully added detailed, recognizable autumn leaves as requested.
• + Good hair motion that effectively follows the implied wind direction.

• − Noticeable over-smoothing and slight alteration of the subject's facial features, making her look different from the source image.
• − The hair rendering is a bit more illustrative and less realistic in texture compared to the source.