PrunaAI's sub-1-second multi-image editing model supporting up to 5 reference images with state-of-the-art quality
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
P-Image Edit
9.5
arena score
#29 of 32 in Image Editing
Skill signature
· Image Editing
Z-Image Turbo
25.3
arena score
#12 of 62 in Text-to-Image
Vote tally
Where the votes landed
P-Image Edit
0%
win rate
Ties
0%
Z-Image Turbo
0%
win rate
Shared challenges
4
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Bald man challenge
Image EditingEdit instruction
“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
P-Image Edit
Z-Image Turbo
AI Judge Analysis
P-Image Edit
- + Excellent preservation of the source background, clothing, and facial features.
- + Followed the instruction to add a full head of hair.
- + Matches the lighting well on the new hair texture.
- − The hair texture has a somewhat repetitive, 'rendered' look in the curls.
- − The transition between the forehead and the new hairline is a bit abrupt.
Z-Image Turbo
- + Successfully preserved the clothing and overall composition.
- − Failed the main prompt to provide a full, thick head of hair, only adding a light buzz cut.
- − Removed the subject's glasses which was not requested.
- − Significantly changed the background landscape from the source image.
Verdict: P-Image Edit was far more successful, as it actually fulfilled the request to add a full head of hair while preserving the identity and background of the source image perfectly. Z-Image Turbo failed to add long hair, removed the subject's glasses, and altered the background scenery, making it a poor edit of the original source.
Over-the-top cartoon caricature
EditingEdit instruction
“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.”
Source
P-Image Edit
Z-Image Turbo
AI Judge Analysis
P-Image Edit
- + Successfully incorporates all elements: tv anchor, hockey, and dogs.
- + Excellent caricature style that maintains the subject's likeness while exaggerating features.
- + High level of detail and creative composition that tells a humorous story.
- − The text on the background screens is nonsensical/garbled.
- − Slightly messy rendering of the small dog on the desk.
Z-Image Turbo
- + Almost perfectly preserves the original source image's realism and subject.
- − Completely fails to follow the principal instruction to create a caricature.
- − Fails to incorporate the TV anchor or hockey elements.
- − Only adds a very small, blurry dog in the background.
Verdict: P-Image Edit fully embraced the prompt, creating a vibrant and funny caricature that expertly blended the subject's face with the requested themes of hockey and news anchoring. In contrast, Z-Image Turbo failed nearly every aspect of the edit instruction, providing a nearly identical photo to the source with only a minor, low-quality addition of a dog. P-Image Edit is the clear winner for its creativity and prompt adherence.
Studio Ghibli Anime Style
EditingEdit instruction
“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”
Source
P-Image Edit
Z-Image Turbo
AI Judge Analysis
P-Image Edit
- + Successfully applied the Ghibli anime aesthetic with clean lines and soft colors.
- + Maintains the exact composition and character poses while adding a dreamy background.
- + Captures the hand-painted, nostalgic mood requested in the prompt.
- − The characters' faces look more like modern manhwa/manga than classic Studio Ghibli character designs.
- − The background changes the entire setting from a city street to a garden path.
Z-Image Turbo
- + Preserves the likeness and texture of the original people extremely well.
- − Completely failed to apply the requested edit style.
- − The image remains a photograph rather than an illustration.
- − Did not incorporate soft pastel colors or dreamy, hand-painted textures.
Verdict: P-Image Edit followed the instructions perfectly, transforming the famous meme into a cohesive anime-style illustration while retaining the recognizable composition. Z-Image Turbo failed the style transfer entirely, providing what appears to be a slightly modified or high-definition version of the original photo instead of an illustration.
Golden Hour Stroll
Image EditingEdit instruction
“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
P-Image Edit
Z-Image Turbo
AI Judge Analysis
P-Image Edit
- + Excellent adherence to the 'hair blowing in wind' instruction with a dramatic effect
- + Effectively adds numerous falling leaves across the entire composition
- + High source preservation of the background and the dog's features
- − The wind effect on the hair is slightly symmetrical and looks a bit digital
- − The lighting on the added leaves is a bit flat compared to the scene
Z-Image Turbo
- + Subtle and realistic integration of falling leaves into the lighting of the scene
- + Preserves the overall identity of the woman and dog well
- − Fails to meaningfully change the hair to reflect the requested wind effect
- − The foliage on the left side was altered/degraded compared to the source image
- − The 'dynamic motion' and 'energetic feel' are much weaker than the competitor
Verdict: P-Image Edit followed the instructions much more effectively, specifically creating a dramatic wind effect in the hair and adding a visible quantity of flying leaves. Z-Image Turbo was too conservative, failing to significantly alter the hair and unintentionally modifying the background plants while trying to add a few leaves.
Next steps
Explore each model
Tongyi-MAI's 6-billion parameter distilled text-to-image model optimized for speed, achieving high-quality generation in 8 steps or fewer with support for bilingual text rendering