Head to head
Esc

Models · slot A

to navigate to pick

FLUX.2 [dev] Flash fal P-Image Edit PrunaAI

Settled by community votes across 7 shared challenges, with an AI judge weighing in on each.

FLUX.2 [dev] Flash

27.1 arena score

#5 of 62 in Text-to-Image

Skill signature · Text-to-Image

P-Image Edit

9.5 arena score

#29 of 32 in Image Editing

Vote tally

Where the votes landed

FLUX.2 [dev] Flash

0%

win rate

Ties

0%

P-Image Edit

0%

win rate

Shared challenges 7

Challenge by challenge

The strongest take from each model on every shared challenge, with the AI judge's read.

Man and Car in California

Editing
Edit instruction

“Make a photo of the man driving the car down the California coastline”

Source
FLUX.2 [dev] Flash
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Excellent preservation of the man's identity and specific clothing
  • + High level of detail in the interior and textures
  • + Natural-looking lighting and motion blur on the background
  • The steering wheel is positioned incorrectly for a modern high-end vehicle
  • Changed the vehicle's interior from the original modern style to a vintage wooden dash

P-Image Edit

  • + Perfect preservation of the original car's exterior model and color
  • + Accurate scenic composition for the California coastline
  • + Strong sense of motion and speed
  • The man is barely visible and lacks the specific details/clothing from the source image
  • The driver appears too small for the scale of the car

Verdict: FLUX.2 [dev] Flash did an exceptional job of merging the two source images, specifically keeping the man's unique hairstyle and clothing intact while placing him in a believable driving position. P-Image Edit prioritized keeping the exact 3D model of the car from the first source image and a wide-angle view of the coast, but it failed to represent the person well. FLUX.2 is the winner for its superior ability to merge the subject identity with the new environment.

Pose & Character Mashup

Editing
Edit instruction

“Use Image 1 as the exact pose reference and Image 2 as the character reference. Recreate the person/character from Image 2 in the exact dynamic pose and body position from Image 1. Keep the exact face, hair, clothing style/details, and expression from Image 2. Match the lighting and environment of Image 1. The final image must show the character from Image 2 performing the precise action/pose from Image 1 with perfect anatomy and natural integration.”

Source
FLUX.2 [dev] Flash
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Successfully copies the face, sunglasses, and clothing from the character reference.
  • + Matches the background and lighting of the pose reference perfectly.
  • Fails to match the pose, showing the character standing upright instead of the complex contorted pose.
  • Creates a major hallucination artifact with a woman's head and long hair appearing behind the main character.

P-Image Edit

  • + Attempts the dynamic pose and maintains the character's clothing and facial features.
  • + Successfully blends the black sweater and scarf from the character with the red elements of the pose reference.
  • Anatomy is broken, particularly where the legs and torso connect.
  • Included long flowing hair that belonged to the woman in the pose reference instead of the male character's short hair.
  • The raised arm transitions into a leg/foot at the top right.

Verdict: Both models struggled with this complex pose-and-character merge. FLUX.2 [dev] Flash captured the character's identity well but completely ignored the required pose and added a disturbing floating head artifact, while P-Image Edit attempted the pose but resulted in surreal anatomical failures and incorrect hair. P-Image Edit is the marginal winner for at least attempting the pose requested, despite the messy execution.

Outfit Transfer Challenge

Editing
Edit instruction

“Use Image 1 as the base person. Dress them in the exact elaborate outfit from Image 2 (including all layers, accessories, jewelry, and shoes). Carefully adapt the clothing to the body shape and pose in Image 1 while maintaining realistic fabric behavior, correct proportions, and perfect lighting/shadow matching. Keep the person’s exact face, hair, and background completely unchanged.”

Source
FLUX.2 [dev] Flash
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Captured the textures of the coat and scarf well.
  • + Included some jewelry as requested.
  • Severely failed to preserve the person's face, creating a grotesque anatomical glitch with a second head.
  • Did not preserve the background accurately, adding blurry elements behind the subject.
  • Failed to include the jeans from the source outfit.

P-Image Edit

  • + Successfully preserved the subject's face, hair, and the background elements from Image 1.
  • + Accurately transferred the coat, scarf, and jeans from Image 2.
  • + Maintained the correct pose and perspective of the original subject.
  • The jewelry added does not match Image 2's subtle accessories, opting for a bright gold chain instead.
  • The shoes are not visible in Image 2, so the model hallucinated stylized boots that don't match the source.

Verdict: P-Image Edit is the clear winner as it successfully composite the two images while following the instruction to keep the person's face and background unchanged. FLUX.2 [dev] Flash completely failed the task by generating a distorted, multi-faced subject and altering the background.

Bald man challenge

Image Editing
Edit instruction

“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”

Before After
FLUX.2 [dev] Flash
Before After
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Excellent source preservation of the face, glasses, and clothing.
  • + The texture of the hair matches the beard's coarse, frizzy nature.
  • + Perfectly captures the request for high density.
  • The hairline on the forehead looks slightly pixelated or blurred upon close inspection.

P-Image Edit

  • + Successfully adds a full head of hair while preserving the identity and background.
  • + Defined curl texture provides a distinct high-quality look.
  • The hair texture (tightly coiled curls) does not match the straighter texture of the original beard.
  • The lighting on the top of the hair is slightly flatter than the source image's overhead sun.

Verdict: Both models did an excellent job of adhering to the prompt while leaving the rest of the image untouched. FLUX.2 [dev] Flash is the winner because the frizzy texture of the added hair perfectly matches the existing beard hair, creating a more cohesive and believable look than the tight curls in P-Image Edit.

Over-the-top cartoon caricature

Editing
Edit 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
FLUX.2 [dev] Flash
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Excellent caricature style that maintains a strong likeness to the source person.
  • + Creative and humorous integration of dogs as hockey players.
  • + High-quality rendering with clean line work and vibrant colors.
  • The transition from the tiny body to the large head is a bit abrupt even for a caricature.

P-Image Edit

  • + Preserved the subject's original denim shirt from the source image.
  • + Included a good variety of hockey and broadcasting elements.
  • + Good representation of the caricature art style.
  • Largely failed to incorporate the 'love for dogs' aspect, only featuring one small dog with an odd red growth/accessory on its back.
  • The likeness to the original person is slightly less accurate than Model A.
  • Text in the background is garbled.

Verdict: FLUX.2 [dev] Flash followed the instructions much more comprehensively, creating a humorous scene where dogs are actually part of the hockey game, which perfectly captures the requested 'exaggerated and humorous' tone. While P-Image Edit preserved the original clothing better, it almost entirely missed the requirement to include a love for dogs and had issues with background text and legibility.

Studio Ghibli Anime Style

Editing
Edit 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
FLUX.2 [dev] Flash
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Perfectly captures the Studio Ghibli essence with soft watercolors and pencil-like hatching.
  • + The character designs are highly reminiscent of classic Hayao Miyazaki films.
  • + Successfully creates a dreamy, nostalgic background with floral elements and a soft sun.
  • The faces, while stylized, lose some of the specific likeness of the people in the source photo.

P-Image Edit

  • + Preserves the likenesses of the original subjects more accurately than Model A.
  • + Successfully applies a hand-painted texture and pastel color palette.
  • + Maintains the composition of the original meme perfectly.
  • The art style feels more like a generic modern manhwa or digital illustration than the specific Ghibli aesthetic requested.
  • The lighting lacks the 'dreamy' quality found in Model A.

Verdict: While both models successfully translated the 'Distracted Boyfriend' meme into an illustration, FLUX.2 [dev] Flash followed the stylistic instruction much more effectively by replicating the specific hand-drawn, watercolor feel of Studio Ghibli. P-Image Edit produced a cleaner digital illustration that preserved the original faces better but failed to capture the specific nostalgic atmosphere requested.

Golden Hour Stroll

Image Editing
Edit instruction

“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”

Before After
FLUX.2 [dev] Flash
Before After
P-Image Edit

AI Judge Analysis

FLUX.2 [dev] Flash

  • + Successfully added a large number of flying leaves throughout the scene.
  • + Effective 'wind' effect on the hair that spreads it wide.
  • + Excellent preservation of the facial features and original background.
  • The hair effect looks slightly stiff or frozen rather than naturally flowing.
  • Some leaves appear slightly blurry or like static artifacts.

P-Image Edit

  • + Beautifully rendered flowing hair with realistic waves and light interactions.
  • + Leaves are colored vibrantly and distributed naturally.
  • + Perfect preservation of the original woman's face and the dog.
  • Fewer leaves than Model A, making the effect slightly less 'dynamic' in the background.

Verdict: Both models performed exceptionally well, successfully modifying the hair and adding leaves while perfectly preserving the identity of the person and dog from the source image. P-Image Edit (Model B) is the winner because the movement of the hair feels more organic and aesthetically pleasing, whereas FLUX.2 [dev] Flash (Model A) created a hair shape that feels a bit more forced and symmetrical.

Next steps

Explore each model