Black Forest Labs' distilled 9 billion parameter image generation model with sub-second inference and multi-reference support
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [klein] 9B
20.6
arena score
#13 of 32 in Image Editing
Skill signature
· Image Editing
Stable Diffusion 3.5 Medium
16.8
arena score
#56 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [klein] 9B
0.0%
win rate
Ties
0.0%
Stable Diffusion 3.5 Medium
100.0%
win rate
0.0%
0.0% ties
100.0%
Shared challenges
13
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Geometric Composition
Text-to-Image“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent photorealism and high texture detail on the wooden table and red book.
- + Accurate glass refraction and lighting effects.
- + Superior compositional balance and depth of field.
- − The sphere appears to be floating inside the glass rather than resting on the bottom, though this is a common AI interpretation.
Stable Diffusion 3.5 Medium
- + Successfully includes all requested elements in the scene.
- + The light source is clearly coming from the left as requested.
- − The glass cube has distorted edges and lacks the optical realism of Model A.
- − The red book looks more like a flat felt-covered slab than a bound book.
- − The overall image quality is lower with some digital noise and less refined textures.
Verdict: FLUX.2 [klein] 9B is the clear winner due to its exceptional photorealistic quality and realistic rendering of light through glass. While both models followed the prompt instructions and included all objects, Stable Diffusion 3.5 Medium produced a much flatter image with noticeably less detail in the textures and physics.
Candid Street Photography
Text-to-Image“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent skin texture and facial detail showing realistic aging.
- + Accurate bicycle components including the chain, gears, and brakes.
- + High clarity and beautiful depth of field that separates the subject from the background.
- − The cars in the background look static rather than having the requested motion blur.
- − The raindrops are rendered as static white streaks that look slightly overlayed.
Stable Diffusion 3.5 Medium
- + Captures the 'candid' and 'imperfect framing' requested in the prompt with a wider street view.
- + Better representation of 'motion blur' on the vehicle to the left.
- + Atmospheric lighting with vibrant reflections on the wet pavement.
- − Anatomical issues with the man's hands and the bicycle's structural integrity.
- − Lower overall resolution and clarity compared to Model A.
- − The man appears to be hunched over rather than actively 'repairing' the bike.
Verdict: FLUX.2 [klein] 9B is the clear winner due to its superior anatomical correctness and technical detail, producing a highly realistic 50mm-style portrait. While Stable Diffusion 3.5 Medium better captured the 'candid' atmosphere and motion blur, it failed significantly on the structural details of the man and the bicycle.
Fantasy Warrior
Text-to-Image“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent depiction of ornate engraving and weathered leather textures.
- + Logical and clear composition with integrated torch lighting elements.
- + Highly detailed facial features including realistic scarring and beads in hair.
- − The bokeh sparks appear a bit static and uniform in the background.
Stable Diffusion 3.5 Medium
- + Strong cinematic lighting with dramatic warm and cool contrast.
- + Intense character expression that captures the battle-worn theme well.
- + Solid rendering of the shallow depth of field effects.
- − The hair braids appear busy and somewhat messy compared to the prompt's description.
- − The armor engraving is less defined and lacks the intricate detail seen in the competitor.
Verdict: FLUX.2 [klein] 9B followed the prompt with superior attention to detail, especially regarding the 'ornate engraved plate armor' and the specific 'beads' in the hair. While Stable Diffusion 3.5 Medium produced a very atmospheric and cinematic image, it lacked the textural clarity and precise adherence to the material descriptions provided in the prompt.
Modern Clean Menu
Text-to-Image“Modern minimalist restaurant menu design, white background with colorful food photos in grid, sections for appetizers/pizza/mains, bold sans-serif fonts, vibrant accents, clean professional layout for casual dining.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Strong typography that balances bold headers with readable list items
- + Consistent and high-quality food photography across the grid
- + Includes specific requested sections like Pizza and Mains clearly
- − Repeats the 'Appetizers' and 'Mains' labels unnecessarily
- − Contains several gibberish words and merged letters in the title
Stable Diffusion 3.5 Medium
- + Excellent grid layout that integrates photos more thoroughly across the page
- + High variety of food imagery that fits a 'casual dining' theme
- + Consistent minimalist aesthetic
- − Font choice is a bit thin and less professional than Model A
- − Text is largely illegible and lacks the requested section headers
- − Overall composition feels a bit repetitious without a clear title/branding area
Verdict: FLUX.2 [klein] 9B produces a more professional and realistic menu layout with the bold fonts and clear sections requested, despite some repetition in the headers. Stable Diffusion 3.5 Medium creates a nice grid of photos but fails to include the specific requested text sections and the typography is significantly weaker.
Magic Burger Explosion: Fiery Photorealism Challenge
Text-to-Image“Ad for 'Magic Burger'. Dynamic, exploded burger with all components (bun, patty, cheese, lettuce, tomato, sauce) suspended in mid-air. Emphasize photorealistic detail and a sense of motion. Dark, fiery background with glowing embers. Integrate text: 'MAGIC BURGER' as a prominent title, 'LIMITED TIME ONLY' as a secondary message, and '€6.99' in a starburst, all rendered with a fiery, glowing effect.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent adherence to the 'exploded burger' requirement with layers clearly separated in mid-air.
- + Superior text rendering with the requested fiery, glowing effect applied consistently across all elements.
- + Highly dynamic composition with sauce droplets and embers creating a strong sense of motion.
- − The burger bun texture looks slightly plastic-like compared to realistic bread.
- − The placement of the '€6.99' starburst slightly overlaps the bottom bun, though it remains legible.
Stable Diffusion 3.5 Medium
- + Realistic texture on the sesame seed bun and the beef patty.
- + Effective use of actual fire in the background to create high contrast.
- − Failed the 'exploded burger' prompt, as the ingredients are mostly stacked normally rather than suspended in air.
- − The text is small, lacks the requested fiery glow effect, and the starburst is a simple black-and-white graphic.
- − The 'LIMITED TIME ONLY' text contains a minor rendering artifact under the 'M'.
Verdict: FLUX.2 [klein] 9B followed the prompt instructions much more closely, successfully creating the 'exploded' view and applying the fiery glow effect to all text elements as requested. Stable Diffusion 3.5 Medium failed to separate the burger components and provided very basic, non-stylized text that didn't match the advertising aesthetic requested.
Chalkboard Menu
Text-to-Image“Handwritten-style chalkboard menu in a cozy café, all text rendered in the exact same realistic chalk handwriting style with natural variations in letter size, slight slant, and chalk texture — no printed or digital fonts anywhere on the board. Title at the top in elegant cursive chalk handwriting: ‘TODAY’S SPECIALS – APRIL 30, 2026’. Below it, three menu items also in the same handwritten chalk style: ‘Truffle Mushroom Risotto – $24’, ‘Grilled Octopus with Lemon & Herbs – $28’, ‘Brown Butter Chocolate Chip Cookies – $9’. At the very bottom, smaller text in the identical handwritten chalk style (slightly smaller but still clearly legible with the same handwriting characteristics): ‘All items made fresh daily • Ask about our gluten-free options’. Warm ambient café lighting, visible chalk dust and smudges, realistic handwriting imperfections, no clean printed text anywhere.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent text legibility and accuracy with nearly perfect spelling
- + Realistic chalk texture including believable smudges and dust on the board
- + Natural handwritten cursive style that adheres perfectly to the prompt
- − Minor spelling error in the footer text 'fress' instead of 'fresh'
Stable Diffusion 3.5 Medium
- + Stylistic chalk outlines and flourishes
- + Good wood frame texture and background context
- − Extremely poor text rendering with significant gibberish and misspellings
- − Mangled date and prices that do not match the prompt requirements
- − Messy layout with overlapping and unreadable text elements
Verdict: FLUX.2 [klein] 9B followed the prompt instructions precisely, delivering highly legible and accurate handwritten text with a very convincing chalk aesthetic. Stable Diffusion 3.5 Medium failed significantly on text coherence, producing a jumble of nonsensical words and symbols despite attempting the correct artistic style.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent visual depth and cinematic lighting.
- + High level of detail in the horse's anatomy and the astronaut's suit.
- + Rich, vibrant background with multiple celestial bodies.
- − Failed the specific spatial instruction; the astronaut is on top of the horse.
- − Generated nonsensical text at the bottom of the image.
Stable Diffusion 3.5 Medium
- + Successfully placed the horse on top of the astronaut as requested by the prompt.
- + Clean composition with a clear horizon line.
- − The transition between the horse and astronaut is physically incoherent and messy.
- − The astronaut's legs appear detached and incorrectly placed behind the horse.
- − Lower level of detail compared to Model A.
Verdict: While FLUX.2 [klein] 9B produces a much higher quality, cinematic image, it completely ignores the negative constraint/spatial instruction of 'horse on top, not vice versa'. Stable Diffusion 3.5 Medium successfully interprets the difficult spatial request, placing the horse above the astronaut, but suffers from significant anatomical glitched and lower overall visual fidelity.
The Capybara Taxi Driver
Text-to-Image“Photorealistic scene inside a yellow New York taxi at night. A capybara is driving, wearing a yellow taxi driver cap and a dark jacket. It has a calm, professional expression and both front paws on the steering wheel. In the back seat sits a human businesswoman in a coat, looking at her phone with a completely normal, bored expression (as if this is just another normal ride). Through the windows you can see the streets of Manhattan at night with blurred lights. Realistic taxi interior, photorealistic, detailed fur and fabric, 35mm lens, night lighting with reflections, shallow depth of field.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
0% wins
0% ties
100% wins
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent photorealism with convincing textures on the capybara fur and human skin.
- + Perfect adherence to the human character prompt, showing a bored businesswoman on her phone.
- + Realistic interior lighting that feels consistent with a night scene.
- − The passenger is sitting in the front passenger seat instead of the back seat.
- − Minor text artifact on the driver's cap.
Stable Diffusion 3.5 Medium
- + Successfully placed the passenger in the back seat as requested.
- + The capybara's expression is very centered and focus-oriented.
- − The passenger is not looking at a phone and does not appear to be a 'businesswoman'.
- − Poor hand/paw anatomy with strange elongated claws/fingers.
- − Composition is awkward with the car roof cutting off the top of the frame.
Verdict: FLUX.2 [klein] 9B produces a far more realistic image with superior lighting and texture, though it failed the spatial instruction to place the passenger in the back seat. Stable Diffusion 3.5 Medium followed the backseat instruction but failed on nearly every other detail regarding the passenger and overall image quality. FLUX.2 remains the better choice due to its high level of detail and accurate rendering of the 'bored businesswoman' character.
The Halloween Invitation
Text-to-Image“Vintage gothic Halloween party invitation. Dark parchment poster, spooky border with webs and thorns, central glowing jack-o-lantern, bats, twisted trees, moody night sky. Add elegant gothic title text saying "Halloween Party Invitation", a small scroll banner saying "You are invited to a night of frights", and event details at the bottom: Date: 30.10.2026 Time: 7pm Location: The Arches, NYC Spooky but polished, cinematic lighting, square format.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Perfect text rendering for all requested titles, banners, and event details.
- + Excellent aesthetic composition with the central glowing jack-o-lantern and thorn border.
- + Consistent gothic atmosphere with high-quality cinematic lighting.
- − The thorns are slightly repetitive in the border design.
Stable Diffusion 3.5 Medium
- + Includes the requested parchment textures and twisted trees.
- + Has good color contrast between the pumpkins and the background.
- − Major spelling errors in the title and body text (e.g., 'Halloweeen Inviloween').
- − Failed to follow the banner scroll instruction for the specified phrase.
- − Composition feels cluttered and less professional than the prompt requested.
Verdict: FLUX.2 [klein] 9B followed every instruction perfectly, including complex text requirements and specific layout elements like the thorn border and scroll banner. Stable Diffusion 3.5 Medium struggled significantly with text rendering and overall graphic design coherence, resulting in many spelling errors and a messy layout. FLUX.2 is the clear winner for its superior polish and prompt adherence.
Isometric Miniature Diorama Scenes
Text-to-Image“Create a clear, 45° top-down isometric miniature 3D cartoon scene of Japan's signature dish: sushi, with soft refined textures, realistic PBR materials, gentle lighting, on a small raised diorama base with minimal garnish and plate. Solid light blue background. At top-center: 'JAPAN' in large bold text, 'SUSHI' below it, small flag icon. Perfectly centered, ultra-clean, high-clarity, square format.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent typography with clean, bold font rendering.
- + Highly accurate 3D miniature/diorama aesthetic with soft PBR-style textures.
- + Perfect adherence to the isometric perspective and background requirements.
Stable Diffusion 3.5 Medium
- + Vibrant colors with high contrast.
- + Accurate sushi topping texture (ikura/roe).
- − Failed to include the 'JAPAN' text at the top as requested.
- − The font layout for 'SUSHI' is messy with overlapping/ghosting characters.
- − Missing the flag icon and the small raised diorama base.
Verdict: FLUX.2 [klein] 9B followed every instruction perfectly, including the complex text layout, the isometric perspective, and the specific diorama base. In contrast, Stable Diffusion 3.5 Medium failed to render all mandated text, omitted the flag, and the text it did render was distorted and poorly placed.
Adorable Baby Animals in Sunny Meadow
Text-to-Image“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent fur texture and realistic anatomy for all animals including the fox.
- + Successfully included multiple butterflies with intricate details.
- + Beautiful lighting with visible god rays and dew-like sparkling effects.
- − Failed to include the baby bunny, showing only three of the four requested animals.
Stable Diffusion 3.5 Medium
- + High saturation and vibrant colors that enhance the 'wholesome' vibe.
- + Good composition with a clear sunrise background.
- − Failed to include the baby bunny, providing only three animals.
- − The kitten and fox share very similar facial structures and ears, losing distinct species characteristics.
- − Anatomical issues, particularly with the puppy's back legs/tail and the fox's front paw merge.
Verdict: Both models failed to include all four requested animals, with both FLUX.2 and Stable Diffusion 3.5 Medium missing the baby bunny. FLUX.2 [klein] 9B is the superior image due to its much higher level of detail in the fur, more realistic lighting, and better anatomical accuracy. Stable Diffusion 3.5 Medium suffered from species blurring where the cat and fox looked nearly identical.
Vintage Cafe Logo
Text-to-Image“Vintage minimalist restaurant logo for "Caffè Florian", retro cloche dome with steam and "Est. 1720" banner, classic typography, warm brown and cream tones, subtle texture on light background, vector emblem style.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Perfect adherence to text spelling and date
- + Clean, professional vector lines suitable for a real logo
- + Accurate depiction of a cloche dome and steam
- − Slightly generic 'stock' feel to the illustration
Stable Diffusion 3.5 Medium
- + More detailed vintage engraving style
- + Dynamic use of scrolls and banners
- − Spelling errors in primary name ('Florrian')
- − Incorrect date ('Est 170' instead of 1720)
- − Cluttered composition that makes it difficult to read as a minimalist logo
Verdict: FLUX.2 [klein] 9B followed the prompt instructions perfectly, producing a clean, usable logo with no spelling errors and a clear cloche icon. Stable Diffusion 3.5 Medium went for a more intricate illustrative style but failed significantly on the text rendering, misspelling the restaurant name and missing a digit in the date.
Apollo 11: Journey to Tranquility
Text-to-Image“Create a clean, modern vector infographic poster about the Apollo 11 mission. NASA-inspired palette (navy, white, muted red, light gray). Flat-vector style, crisp lines, consistent iconography, subtle gradients only. Steps (stop at landing): 1. Launch (Saturn Vicon) 2. Earth Orbit (Earth + orbit ring icon) 3. Translunar (trajectory arc icon) 4. Lunar Orbit (Moon + orbit ring icon) 5. Descent (lunar module descending icon) 6. Landing (lunar module on the surface icon) Small supporting elements (minimal text): • Crew strip: three silhouette icons with only last names: Armstrong, Aldrin, Collins. • Landing site marker: Moon pin labeled "Tranquility" only. Layout constraints: generous margins, large readable labels, clean background with subtle stars. Vector-only, print-poster look, high resolution.”
FLUX.2 [klein] 9B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 9B
- + Strong adherence to the requested flat-vector infographic style
- + Clean and legible layout with clear iconography for each stage
- + Excellent use of the specified NASA-inspired color palette
- − Several spelling errors in text labels such as 'Launchh', 'Eardht', and 'Landining'
- − Incorrectly includes portraits instead of purely focusing on the mission steps
Stable Diffusion 3.5 Medium
- + Creative use of negative space and orbital curves
- + Includes the exact mission step headers requested
- − Icons are cluttered and do not clearly represent the specific steps described
- − Typography is messy and contains significant nonsensical text string artifacts
- − Composition feels unbalanced with a large abstract sun/planet element
Verdict: FLUX.2 [klein] 9B is the clear winner for its superior execution of the flat-vector style and logical infographic flow, despite some minor spelling errors. Stable Diffusion 3.5 Medium struggled with the clean layout and produced icons that were difficult to interpret and text that was largely gibberish.
Next steps
Explore each model
Stability AI's 2.5-billion parameter Multimodal Diffusion Transformer with improvements (MMDiT-X) text-to-image model optimized for consumer hardware, featuring improved image quality, typography, and complex prompt understanding