P-Image Edit PrunaAI vs Wan 2.7 Alibaba

P-Image Edit

• + Excellent preservation of the car's original design and color.
• + Accurately represents the man from the source image behind the wheel.
• + Dynamic motion blur on the road and wheels creates a realistic sense of speed.

• − The man's scale relative to the car is slightly too small.

Wan 2.7

• + Successfully places the car on a coastal highway.
• + Good lighting and shadows that match the outdoor environment.

• − Fails to use the specific man from the source image, replacing him with a generic figure.
• − The face of the driver is distorted and poorly rendered.
• − The car's dimensions and details are subtly altered from the source.

P-Image Edit

• + Successfully integrated face and scarf elements from Image 2.
• + Maintained a strong resemblance to the character's clothing color scheme and sunglasses.
• + Followed the complex leg and stool arrangement from Image 1.

• − Confused the hair logic by blending Image 1's long hair with the short-haired male character.
• − Anatomical failure in the upper arm, creating a floating red sleeve disconnected from the shoulder.
• − Inconsistent clothing application where some limbs have black sleeves and others have red.

Wan 2.7

• + Perfect preservation of the original Image 1 background and pose.
• + High visual quality with no weird artifacts or glitches.

• − Complete failure to perform any editing; the character from Image 2 was not integrated at all.
• − Instruction following is zero for the character swap task.

P-Image Edit

• + Excellent adherence to the specific clothing items (peacoat, plaid scarf, blue jeans) from Image 2.
• + Maintains the person's identity and facial features with high accuracy.
• + Correctly preserves the wood structure and environment from the base image.

• − Added a thick gold chain and yellow-laced boots that were not in either source image.
• − The body proportions appear slightly shortened and compressed compared to the original person's height.

Wan 2.7

• + Successfully preserves the subject's face, skin patterns (vitiligo), and general hair style.
• + Good integration of the person into the beach environment with realistic lighting.

• − Failed the clothing instruction entirely by generating a generic ornate coat instead of the specific peacoat/scarf outfit from Image 2.
• − Anatomical issues with the hands, including an extra thumb/finger structure on the right hand.
• − Source preservation is weak as it changed the subject's pose significantly.

P-Image Edit

• + Successfully adds a full head of hair with high density
• + Perfectly preserves all original elements of the scene, facial features, and clothing

• − The hair texture looks somewhat synthetic/plasticky and coiled
• − The hairline is slightly too sharp and the volume is a bit exaggerated for a realistic look

Wan 2.7

• + Extremely realistic hair texture and natural-looking density
• + Seamless integration of the hairline and temples with the existing beard and skin
• + Flawless preservation of the source image identity and environmental lighting

• − Slightly less 'full' than requested compared to the volume in the other model, though more realistic

P-Image Edit

• + Excellent caricature of the subject that maintains her likeness while exaggerating features for humor.
• + Creative integration of all three themes, placing her as an anchor directly on an ice rink with hockey players.
• + High quality illustration style with clean lines and a professional layout.

• − The dog element is a bit small and secondary compared to the hockey theme.
• − Text on the background screens is nonsensical.

Wan 2.7

• + Strong emphasis on the 'dogs' element of the prompt with multiple humorous canine characters.
• + Good preservation of the source image's pose (the outstretched selfie arm).
• + Legible and relevant text in speech bubbles.

• − The hockey element is minimal, represented only by a small puck and a miniature rink on the desk.
• − The caricature style feels slightly less polished and more generic than Model A.
• − The outfit was changed to a suit, losing the denim shirt from the original.

P-Image Edit

• + Excellent preservation of the original pose and character lineup.
• + Captures the dreamy background and pastel colors requested.
• + Maintains clean, illustrative lines consistent with modern anime styles.

• − The background is completely changed from an urban street to a park, losing source context.
• − Artificial-looking shadows on the faces.

Wan 2.7

• + Successfully preserves the original urban background while applying a painterly texture.
• + Achieves a beautiful watercolor/hand-painted aesthetic that feels very nostalgic.
• + Characters' facial expressions are more closely aligned with the original meme's intensity.

• − The color palette is slightly muted and yellowish compared to the requested 'soft pastel'.
• − Minor anatomical artifacts in the woman's hair on the left.

P-Image Edit

• + Excellent adherence to the 'hair blowing in wind' instruction with a dramatic, symmetrical effect.
• + Adds a large quantity of colorful leaves throughout the scene to enhance the sense of motion.
• + Maintains the identity of the person and dog perfectly.

• − The wind effect on the hair is slightly stylized and surreal rather than naturalistic.

Wan 2.7

• + Provides a more naturalistic wind effect on the hair that feels realistic to the environment.
• + Good source preservation, keeping the background and character details consistent.
• + The leaves integrated have varied motion blur, adding to the energetic feel.

• − The motion effect is more subtle compared to Model A.
• − Fewer leaves added toward the top of the frame than requested for a 'lively' feel.