Black Forest Labs' premium multimodal flow transformer with greatly improved prompt adherence and typography generation for in-context image generation and editing without compromise on speed
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
FLUX.1 Kontext [max]
23.7
arena score
#23 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.1 Kontext [max]
0%
win rate
Ties
0%
P-Image Edit
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.”
FLUX.1 Kontext [max]
P-Image Edit
AI Judge Analysis
FLUX.1 Kontext [max]
- + Successfully added thick hair that matches the subject's existing beard texture
- + Preserved the background and clothing nearly perfectly
- + Maintained the facial identity and lighting of the original image
- − The hairline is a bit harsh and lacks fine transition hairs
- − The top of the hair has some minor digitizing artifacts
P-Image Edit
- + Excellent preservation of the facial features and glasses
- + High level of detail in the individual curls
- + Consistent lighting across the new hair
- − The hair texture (tightly coiled curls) does not match the straight/wavy texture of the existing beard
- − The hair-to-skin transition on the forehead looks slightly superimposed
- − A few curls on the perimeter have a repetitive, artificial appearance
Verdict: Both models did an excellent job of preserving the source image context and facial identity. FLUX.1 Kontext [max] is the winner because it provided hair that logically matches the man's existing beard style and color, whereas P-Image Edit chose a contrasting curly texture that feels less natural for this specific individual.
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
FLUX.1 Kontext [max]
P-Image Edit
AI Judge Analysis
FLUX.1 Kontext [max]
- + Excellent character likeness preserved while applying a sticker-book style.
- + Includes all requested elements: news studio, microphone, dog, and hockey stick
- − The hockey element is minimally represented by a floatation stick at the top.
- − Added glasses that were not in the original photo.
P-Image Edit
- + Successfully creates an elaborate, lively scene with high engagement.
- + Detailed integration of hockey through players and rink background.
- − The facial resemblance is less accurate than the first model.
- − Contains cluttered elements and unintelligible text graphics.
Verdict: FLUX.1 Kontext [max] maintains a better likeness to the source subject and uses a cohesive illustration style, whereas P-Image Edit creates a more detailed background that better represents the 'hockey' aspect. FLUX.1 Kontext [max] is preferred for capturing the subject's essence despite the simpler composition.
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
FLUX.1 Kontext [max]
P-Image Edit
AI Judge Analysis
FLUX.1 Kontext [max]
- + Excellent adherence to the Studio Ghibli art style including eye shape and facial features.
- + Accurately replicates the paper texture and watercolor-style coloring requested.
- + Perfectly preserves the composition and poses of the original meme.
- − The hand visible on the bottom right of the man's shirt is slightly malformed compared to the source.
P-Image Edit
- + Successfully creates a 'dreamy' background with cotton-candy clouds.
- + Maintains high structural consistency with the source image characters.
- − Fails to capture the specific Studio Ghibli aesthetic, leaning toward a generic modern digital illustration style.
- − The lighting and textures feel overly smooth and digital rather than 'hand-painted'.
- − Characters' facial features remain too realistic for the requested style.
Verdict: FLUX.1 Kontext [max] is the clear winner as it successfully translated the image into the iconic Studio Ghibli art style with watercolor textures and specific character design choices. P-Image Edit maintained the original facial structure well but produced a generic digital painting that largely ignored the 'hand-painted' and 'Ghibli' stylistic requirements.
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.”
FLUX.1 Kontext [max]
P-Image Edit
AI Judge Analysis
FLUX.1 Kontext [max]
- + Excellent preservation of the woman's face and body features
- + Natural integration of blowing hair strands
- + Subtle but effective movement in the dog's fur and tail
- − Some minor anatomical distortion on the hand reaching for the dog
- − The number of leaves is slightly low for a 'dynamic' feel
P-Image Edit
- + Successfully captures a very dynamic wind effect in the hair
- + Plentiful and colorful flying leaves for high energy
- + Good overall preservation of the source image background and subject
- − The hair movement looks a bit artificial/stiff at the ends
- − Small artifacts near the hairline where it meets the forehead
Verdict: Both models followed the instructions well, but FLUX.1 Kontext [max] produced a more realistic result by subtly altering the woman's posture and the dog's tail to match the energetic feel, whereas P-Image Edit primarily focused on the hair and leaves while keeping the rest of the subject static. FLUX.1 Kontext [max] is the winner for its superior blend of hair physics and overall image coherence.
Next steps
Explore each model
PrunaAI's sub-1-second multi-image editing model supporting up to 5 reference images with state-of-the-art quality