Black Forest Labs' open-weights image generation model with frontier performance, available for non-commercial local deployment
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev]
24.5
arena score
#18 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Stable Diffusion 3.5 Medium
16.8
arena score
#56 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [dev]
0%
win rate
Ties
0%
Stable Diffusion 3.5 Medium
0%
win rate
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent spatial reasoning with the blue sphere resting naturally on the bottom.
- + Highly realistic rendering of glass thickness and refractive properties.
- + Accurate lighting consistently applied from the left window source.
- − The plant is conceptually behind the cube but slightly overlaps the glass edges.
Stable Diffusion 3.5 Medium
- + Adheres to all requested objects including the red book and green plant.
- + Good color saturation and vibrance across the scene.
- − The blue sphere is levitating unnaturally in the center of the cube.
- − The red book appears to be merging with the top glass panel of the cube.
- − Texture quality on the wooden table is somewhat soft and lacks detail.
Verdict: FLUX.2 [dev] produces a much more physically coherent scene where the sphere respects gravity and the glass has believable volume. Stable Diffusion 3.5 Medium struggles with object placement, resulting in a levitating sphere and distorted contact points between the book and the cube.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the 'imperfect framing' and 'motion blur from passing cars' prompts.
- + Extremely realistic skin texture and raindrops on the jacket.
- + Anatomically correct hands and complex mechanical bicycle details.
- − The red bicycle frame is partially obscured by another bike-like structure in the foreground.
Stable Diffusion 3.5 Medium
- + Strong composition with atmospheric reflections and lighting.
- + Correct subject matter and color palette.
- − Significant anatomical errors with the hands and feet blending into the bike.
- − Lacks the requested 'motion blur' from passing cars as the cars in the background look static.
- − Visible AI artifacts on the man's face and the bicycle's basket.
Verdict: FLUX.2 [dev] significantly outperforms the competitor by delivering a photograph that looks authentic and high-resolution, following specific technical requests like motion blur and realistic skin texture. Stable Diffusion 3.5 Medium struggles with anatomy, particularly where the man's hands and feet interact with the bicycle, and fails to produce the motion blur requested in the prompt.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent depiction of ornate engraved plate armor with high-quality textures
- + Faithfully follows the request for hair beads and leather strap details
- + Naturalistic face with realistic battle-worn elements like scars and dirt
- − The torch in the background is a bit distracting in the composition
Stable Diffusion 3.5 Medium
- + Intense facial expression and strong dramatic lighting
- + Detailed skin texture including pores and freckles
- − Failed to include the requested beads in the braided hair
- − The hair is overly dark and some braids appear messy or poorly defined near the forehead
- − The armor engraving is less detailed compared to its competitor
Verdict: FLUX.2 [dev] followed the prompt more comprehensively, specifically including the beads in the hair and showing much higher skill in rendering the ornate engravings of the armor. Stable Diffusion 3.5 Medium provided a very intense portrait with great skin textures, but missed key prompt elements like the beads and had less intricate detailing on the straps and cloth layers.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography rendering with readable fonts
- + Clear logical hierarchy for sections
- + Professional use of color accents and white space
- − Some minor gibberish in long-form text descriptions
- − Mains section is very short compared to the others
Stable Diffusion 3.5 Medium
- + Includes a high variety of food images
- + Consistent grid layout for many menu items
- − Typography is distorted and nearly illegible
- − Layout feels cluttered and repetitive
- − Poor text alignment and spacing
Verdict: FLUX.2 [dev] successfully captures the modern minimalist aesthetic with a professional multi-column layout and clean, bold typography. In contrast, Stable Diffusion 3.5 Medium struggles significantly with text legibility and produces a cluttered layout that lacks the professional design polish requested by the prompt.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the 'exploded' layout with clearly separated layers.
- + Perfect rendering of text and graphics with the requested fiery, glowing effect.
- + High level of photorealistic detail in the food textures and sauce splashes.
- − The bun texture is slightly generic compared to the patty.
Stable Diffusion 3.5 Medium
- + Strong fiery background with intense lighting.
- + Good burger texture, particularly the sesame seed bun.
- − Failed the 'exploded' layout request, showing a mostly assembled burger.
- − Text is small, plain white, and lacks the requested fiery/glowing effect.
- − The price starburst is a simple black and white graphic that clashes with the scene.
Verdict: FLUX.2 [dev] followed all prompt instructions perfectly, including the complex exploded layout and specific fiery text effects. Stable Diffusion 3.5 Medium failed to deconstruct the burger layers and missed the stylistic requirements for the typography and price tag.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent text rendering with near-perfect spelling and legibility.
- + Highly realistic chalk texture and smudging effects on the blackboard.
- + Natural composition follows all prompt instructions including the specific third menu item.
- − Slight alignment issues in the price for the second item where it seems orphaned.
Stable Diffusion 3.5 Medium
- + Creative use of decorative cursive elements.
- + Good frame and background texture.
- − Significant spelling errors and gibberish text throughout the board.
- − Failed to follow the date instruction '2026'.
- − Poor layout with overlapping and messy chalk lines that make it hard to read.
Verdict: FLUX.2 [dev] significantly outperforms Stable Diffusion 3.5 Medium by following every text instruction precisely, including the specific menu items and date. While Stable Diffusion 3.5 Medium struggles with legibility and basic spelling, FLUX.2 [dev] produces a highly realistic, professional-looking chalkboard that matches the requested aesthetic perfectly.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent anatomical realism for both the horse and the spacesuit.
- + Cinematic lighting with a vibrant, detailed nebula background.
- − The composition is somewhat conventional for a surreal prompt.
Stable Diffusion 3.5 Medium
- + Strong adherence to the surreal atmosphere mentioned in the prompt.
- + Interesting use of shadows on the planet surface below the horse.
- − Significant anatomical issues with the horse's legs and the astronaut's lower body.
- − The background stars appear somewhat grainy and less refined.
Verdict: FLUX.2 [dev] provides a high-fidelity, cinematic image with realistic textures and lighting, whereas Stable Diffusion 3.5 Medium attempts a more conceptual layout but suffers from anatomical distortions. FLUX.2 [dev] is the preferred choice for its superior visual clarity and professional execution.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the 'bored businesswoman on phone' prompt detail
- + High level of photorealism and natural lighting consistency
- + Accurate depiction of a capybara's anatomy and fur texture
- − The capybara's paws look more like human-primate hands than capybara feet
Stable Diffusion 3.5 Medium
- + Strong symmetrical composition and vibrant colors
- + Creative cap badge detail on the driver's hat
- − Failed to include the businesswoman looking at her phone
- − The capybara's facial proportions are slightly stylized and squirrel-like
- − The paws are resting on the dashboard/top of wheel rather than gripping it naturally
Verdict: FLUX.2 [dev] followed the complex narrative instructions much better than Stable Diffusion 3.5 Medium, specifically by including the businesswoman on her phone with a bored expression. FLUX.2 [dev] also achieved a more convincing photorealistic look for a taxi interior at night, whereas Stable Diffusion 3.5 Medium missed key prompt elements and felt more like a studio portrait.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Perfect text rendering for all requested strings including dates and locations.
- + Highly cohesive gothic aesthetic with elegant typography and cinematic lighting.
- + Strong adherence to composition details like the thorny border and scroll banner.
- − The pumpkin is black/dark grey rather than an orange jack-o-lantern, though it fits the 'moody' prompt.
Stable Diffusion 3.5 Medium
- + Bright, classic orange jack-o-lanterns that pop against the background.
- + Good use of twisted trees and cobwebs to frame the central parchment.
- − Significant spelling errors in every line of text (e.g., 'Halloweeen Inviloween', 'Loccation').
- − Missing requested text elements like '7pm' and corrupted year '226'.
- − Lacks the requested scroll banner for the secondary text.
Verdict: FLUX.2 [dev] followed every instruction perfectly, producing a professional-grade invitation with flawless text and a sophisticated gothic atmosphere. Stable Diffusion 3.5 Medium struggled deeply with the typography, resulting in numerous spelling errors and a cluttered layout that failed to include all the prompt's specific details.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the isometric diorama style requested.
- + Perfect text rendering and inclusion of the Japanese flag icon.
- + Higher quality 3D textures with clear separation of sushi ingredients.
- − The color palette is slightly muted compared to the vibrant 'cartoon' request.
Stable Diffusion 3.5 Medium
- + Successfully used a solid light blue background.
- + Clean text rendering for the word 'SUSHI'.
- − Failed to include the diorama base requested in the prompt.
- − Did not include the Japanese flag icon.
- − The 'JAPAN' text is small and lacks the 'large bold' styling requested.
Verdict: FLUX.2 [dev] followed every instruction in the prompt, including the specific isometric diorama base, the flag icon, and the hierarchy of the text. Stable Diffusion 3.5 Medium failed to create the diorama structure and missed the flag icon entirely, resulting in a much simpler composition. FLUX.2 [dev] is the clear winner for its superior prompt adherence and visual complexity.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the list of animals including both the bunny and several butterflies.
- + Very natural, warm god rays and soft lighting that feels photorealistic.
- + High-quality fur textures and realistic anatomy for each animal.
- − The animals are sitting relatively still rather than actively 'tumbling' as requested.
Stable Diffusion 3.5 Medium
- + Bright, vibrant colors that evoke a joyful vibe.
- + Dynamic poses that suggest a bit more movement than the other model.
- − Failed to include the baby bunny requested in the prompt.
- − The fur textures look over-sharpened and more like an illustration than a 'hyper-photorealistic' 8K masterpiece.
- − The fox and kitten have anatomically strange facial features and flattened ears.
Verdict: FLUX.2 [dev] significantly outperforms Stable Diffusion 3.5 Medium by including all requested animals (including the bunny) and maintaining a high level of photorealism. Stable Diffusion 3.5 Medium has a more 'AI-plastic' look with unnatural sharpening and anatomical inconsistencies.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography with perfect spelling and correct accent marks.
- + Clean vector emblem style that adheres strictly to the minimalist logo prompt.
- + Sophisticated use of subtle texture on the background for a vintage feel.
- − The steam element is very simplified, though it fits the minimalist style.
Stable Diffusion 3.5 Medium
- + Ornate illustrative style with nice cross-hatching detail.
- + Creative integration of a coffee cup at the bottom.
- − Failed spelling of 'Florian' as 'Florrian'.
- − Incorrect date 'Est 170' instead of 1720.
- − Illustration is cluttered and busy, failing the 'minimalist' requirement of the prompt.
Verdict: FLUX.2 [dev] followed every instruction perfectly, delivering a professional-grade minimalist logo with accurate text and a clean vector aesthetic. Stable Diffusion 3.5 Medium failed on typography, spelling, and adherence to the minimalist style, producing a cluttered illustration with multiple text errors.
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 [dev]
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [dev]
- + Clean vector aesthetic that matches the requested flat-vector style.
- + Excellent adherence to the NASA-inspired color palette.
- + Includes high-quality icons for the Saturn V and Lunar Module.
- − The sequence of steps is jumbled and not in a logical mission order.
- − Significant text corruption on several labels despite correctly spelling others.
- − Confuses 'Descent' with a moon icon and puts the actual landing icon under a gibberish label.
Stable Diffusion 3.5 Medium
- + Creative use of a large circular element to anchor the composition.
- + Follows a vertical linear sequence for the mission steps.
- + Maintains the requested muted color palette well.
- − Very low text legibility and numerous spelling errors.
- − Visuals do not correspond to the labels (e.g., Earth icons for the 'Landing' step).
- − Failed to incorporate specific requested icons like the Saturn V or Lunar Module.
Verdict: FLUX.2 [dev] is the clear winner because it successfully captures the modern infographic aesthetic and produces high-quality icons that look professional, despite the labels being out of order and containing some typos. Stable Diffusion 3.5 Medium fails to produce recognizable mission icons, suffers from severe text corruption, and uses circular Earth/Moon motifs for every step instead of the specific mission hardware requested.
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