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]
24.3
arena score
#19 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Stable Diffusion 3.5 Large Turbo
10.7
arena score
#61 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [dev]
0%
win rate
Ties
0%
Stable Diffusion 3.5 Large Turbo
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Perfect adherence to all spatial instructions and object placement.
- + Highly realistic textures for glass, wood, and paper.
- + Excellent handling of refractions through the glass, showing the plant as requested.
- − None notable.
Stable Diffusion 3.5 Large Turbo
- + Sharp lighting and shadows coming from the correct direction.
- + Clean, modern aesthetic.
- − Failed the spatial placement by putting the red book inside the cube instead of on top.
- − The blue sphere is sitting on the book rather than the table inside the cube.
- − The plant is to the side rather than behind the cube, minimizing its visibility through the glass.
Verdict: FLUX.2 [dev] significantly outperformed Stable Diffusion 3.5 Large Turbo by following every spatial instruction in the prompt, including the complex request of a plant visible through the glass. Stable Diffusion 3.5 Large Turbo failed to place the red book on top of the cube, instead placing both the book and the sphere inside it.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to photorealistic requirements with natural skin textures and realistic rain details.
- + Strong composition that captures the 'candid' and 'imperfect framing' requested.
- + Accurate motion blur and bokeh on the background traffic.
- − The hands have minor anatomical artifacts near the bike's handlebars.
Stable Diffusion 3.5 Large Turbo
- + Successfully includes the red bicycle and the elderly man.
- + Follows the basic subject matter of the prompt.
- − The image has a plastic, CGI-like appearance that fails the 'no stylization' requirement.
- − Traffic in the background is static, missing the requested motion blur.
- − The man's hair and clothing have unnatural, smooth textures.
Verdict: FLUX.2 [dev] produced a highly realistic, cinematic street photograph that perfectly captured the requested atmosphere of light rain and natural textures. In contrast, Stable Diffusion 3.5 Large Turbo failed to deliver the photographic quality, producing a stylized, almost illustrative result with static cars and simplified details. FLUX.2 better understood the nuance of the 'candid' and '50mm' lens instructions.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the 'beads in hair' prompt element.
- + The skin textures, dirt, and scarring look realistic and grimy.
- + Lighting from the torch feels integrated with the scene and reflects naturally on the metal.
- − The torch is partially cut off on the left side, slightly impacting the composition.
Stable Diffusion 3.5 Large Turbo
- + Strong engraving detail on the armor.
- + Distinct bokeh effect in the background.
- − Failed to include the requested beads in the braids.
- − The skin and facial hair have a plastic, over-smoothed digital look compared to the prompt's request for 'lifelike'.
- − The blood/scars appear painted on like face paint rather than physical wounds.
Verdict: FLUX.2 [dev] followed the prompt more accurately, particularly with the inclusion of small beads in the braids and the realistic 'battle-worn' skin textures. Stable Diffusion 3.5 Large Turbo produced a much more stylized, almost video-game-like image that failed to capture the gritty, lifelike detail requested.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the menu layout prompt, including categorized sections.
- + High-quality, realistic food photography integrated into the design.
- + Professional typography and use of vibrant color accents for a modern look.
- − Some text elements contain garbled or nonsensical characters.
- − Minor repetition in section headers like 'PIZZAU' and 'PIZZA'.
Stable Diffusion 3.5 Large Turbo
- + Creative use of a grid for food imagery.
- + Clean minimalist aesthetic with plenty of white space.
- − Layout looks like an artistic collage rather than a functional restaurant menu.
- − Food imagery looks rubbery and artificial compared to real photography.
- − Typographical hierarchy is weak, and text is largely unreadable/placeholder lines.
Verdict: FLUX.2 [dev] successfully creates a professional, functional menu design that follows all prompt instructions regarding layout and sections. In contrast, Stable Diffusion 3.5 Large Turbo produces a more abstract collage that lacks the professional structure and high-quality food photography found in the FLUX output.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography rendering with the requested fiery, glowing effects.
- + Superior prompt adherence for the 'exploded' view requirement.
- + Highly photorealistic textures on the bun, meat, and vegetables.
- − The starburst shape for the price is slightly irregular but still recognizable.
Stable Diffusion 3.5 Large Turbo
- + Dynamic lighting with interesting smoke and ember effects.
- + High contrast and vibrant colors.
- − Failed to include any of the requested text elements.
- − Did not produce an 'exploded' burger, showing an assembled one instead.
- − Visual style leans toward illustrative/AI-art rather than photorealism.
Verdict: FLUX.2 [dev] followed every instruction perfectly, including complex text rendering and the specific 'exploded' layout of the burger. In contrast, Stable Diffusion 3.5 Large Turbo completely ignored all text requirements and failed to create the exploded view, resulting in a generic (though visually striking) burger image.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the complex text requirements with perfect spelling.
- + Highly realistic chalk texture with natural smudging and varied pressure.
- + Authentic handwritten look that perfectly matches the requested aesthetic.
- − The 'w' in 'with' is slightly obscured by a chalk smudge.
Stable Diffusion 3.5 Large Turbo
- + Cozy and clean aesthetic composition.
- + Good use of lighting and shelf staging.
- − Multiple spelling errors across the entire board ('trulale', 'risctto', 'ocopue').
- − Text looks like a digital font rather than natural chalk handwriting.
- − Failed to follow the date and pricing instructions accurately.
Verdict: FLUX.2 [dev] followed every detail of the prompt, providing accurate text, realistic chalk textures, and a convincing handwritten style. In contrast, Stable Diffusion 3.5 Large Turbo struggled with spelling, failed to capture the requested handwriting style, and produced text that looks like a clean digital overlay rather than chalk.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent photorealism and cinematic lighting
- + High level of detail in the astronaut suit and horse's fur coat
- + Natural integration of the horse and rider within the space background
- − Failed the negative constraint; the astronaut is riding the horse, not the other way around
- − The moon lighting does not match the scene's primary light source
Stable Diffusion 3.5 Large Turbo
- + Clean, graphic composition with a strong sense of movement
- + Correctly captures the 'surreal' aspect requested in the prompt
- − Failed the negative constraint; the astronaut is riding the horse
- − The anatomy of the horse's rear leg is distorted and unnatural
- − The celestial body in the top left has a low resolution, blurred texture
Verdict: Both FLUX.2 [dev] and Stable Diffusion 3.5 Large Turbo failed the negative constraint to have the 'horse on top.' However, FLUX.2 [dev] is the superior image due to its professional cinematic lighting, realistic textures, and overall coherence, whereas Stable Diffusion produced significant anatomical artifacts in the horse's legs.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the 'human looking at phone' prompt
- + High detail in the capybara's fur and clothing texture
- + Realistic framing showing the interior and the street simultaneously
- − The capybara's paws look slightly more like human hands than actual capybara feet
Stable Diffusion 3.5 Large Turbo
- + Dynamic lighting and lens bokeh in the background
- + Good capybara anatomy on the paws
- − Failed to include the phone for the businesswoman
- − The interior perspective is slightly cramped and less realistic
- − The businesswoman is out of focus and lacks the bored expression detail
Verdict: FLUX.2 [dev] significantly outperformed Stable Diffusion 3.5 Large Turbo by capturing all elements of the prompt, specifically the businesswoman's bored expression and use of a phone, which were completely missing in the latter. FLUX.2 [dev] also provided a more convincing 'New York' feel through its layout and lighting.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Perfect adherence to the text prompts, including the main title, scroll banner, and all event details.
- + Highly polished and atmospheric Gothic aesthetic with excellent cinematic lighting.
- + Superior integration of the web and thorn border elements that frame the central composition well.
- − The parchment texture is slightly repetitive near the outer edges.
Stable Diffusion 3.5 Large Turbo
- + Clean graphic design with clear silhouettes for the trees and bats.
- + Effective use of symmetry and negative space.
- − Failed to include the required scroll banner and the specific event details (Date, Time, Location).
- − The title text is truncated to just 'Halloween Party' and uses a basic font rather than an elegant Gothic style.
- − The lighting feels flat and more like a cartoon than the requested 'cinematic' and 'moody' night sky.
Verdict: FLUX.2 [dev] significantly outperformed Stable Diffusion 3.5 Large Turbo by following the prompt's complex instructions perfectly, including all requested text and event details. Stable Diffusion 3.5 Large Turbo failed to render the sub-header and the specific event details, while also delivering a less detailed and atmospheric image. FLUX.2 [dev] captured the 'vintage gothic' mood much more effectively with its use of texture and lighting.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Perfect adherence to text instructions with correct spelling of 'SUSHI'.
- + Clean, professional 3D rendering with soft lighting and high-quality textures.
- + Accurate depiction of the Japanese flag and isometric diorama layout.
- − The sushi pieces are slightly crowded on the plate.
Stable Diffusion 3.5 Large Turbo
- + Bold, vibrant colors that fit a 3D cartoon style.
- + Interesting use of physical signage for the text elements.
- − Mispelled 'SUSHI' as 'SIIHI' on the plate.
- − The flag is generic/incorrect and does not match the Japanese flag requested.
- − Failed the request for the text to be at the top-center of the square format.
Verdict: FLUX.2 [dev] followed every instruction perfectly, including the specific text placement, correct spelling, and accurate national flag. Stable Diffusion 3.5 Large Turbo struggled with the spelling of 'SUSHI' and failed to provide the correct flag, resulting in a less polished and accurate diorama.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the prompt, including all four specific animal species.
- + Superior lighting effects with realistic god rays and backlighting on fur.
- + Highly realistic textures for fur, grass, and butterflies.
- − Added a second bunny not explicitly requested in the prompt.
- − The 'tumbling' action is more of a static seated group pose.
Stable Diffusion 3.5 Large Turbo
- + Successfully captured the 'big expressive eyes' and joyful vibe.
- + Vibrant colors and high-contrast backlighting.
- − Failed to include a rabbit or a distinct fox, showing mostly cat/dog hybrids.
- − Lacks the '8K masterpiece' level of detail, appearing more like a digital illustration or vector art than a photorealistic scene.
- − Anatomical issues with paws blending into each other.
Verdict: FLUX.2 [dev] followed the prompt much more accurately by including all four distinct animal types with high photorealism and beautiful atmospheric lighting. Stable Diffusion 3.5 Large Turbo struggled with species variety and produced a much more stylized, less detailed image that lacked the requested realism.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography including the grave accent in 'Caffè'.
- + Minimalist vector aesthetic that perfectly matches the requested logo style.
- + Clean, professional banner integration.
- − The 'steam' icon at the top is very basic compared to the rest of the logo.
Stable Diffusion 3.5 Large Turbo
- + Beautiful warm color gradients and lighting.
- + Detailed 'steam' illustration that looks elegant.
- − Missed the spelling of 'Caffè' by adding an extra 'e' and messing up the letter 'a' in Florian.
- − Included a mug handle which was not requested and clashes with the cloche dome idea.
- − The logo feels cluttered and less like a minimalist vector emblem.
Verdict: FLUX.2 [dev] followed the prompt more accurately, producing a clean, professional vector logo with correct spelling and typography. Stable Diffusion 3.5 Large Turbo struggled with the text rendering and introduced unnecessary elements like a mug handle that diluted the minimalist cloche dome concept.
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 Large Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the requested grid-based infographic structure.
- + Crisp vector iconography that accurately represents the Saturn V and Lunar Module.
- + Highly legible typography and correct use of the requested color palette.
- − Includes several redundant or garbled labels beyond the requested six steps.
- − The text 'SATURN ICON' and 'Trarquility' contain minor spelling or redundancy errors.
Stable Diffusion 3.5 Large Turbo
- + Strong aesthetic appeal with a vintage mid-century modern infographic style.
- + Good use of the muted NASA-inspired color palette.
- + Creative use of geometric shapes to frame the narrative.
- − Failed to follow the requested 6-step sequence accurately.
- − Significant issues with text rendering, including major spelling errors like 'Apoll.o'.
- − The astronaut on the moon appears to be standing on a ladder-like structure that is physically illogical.
Verdict: FLUX.2 [dev] followed the specific structural instructions much better, providing a clear 6-step process with mostly accurate icons for each phase of the mission. Stable Diffusion 3.5 Large Turbo created a beautiful artistic poster but failed the technical infographic requirements, miscounting the steps and failing on almost every word of text.
Next steps
Explore each model
Distilled version of SD 3.5 Large that generates high-quality images in just 4 steps, offering faster inference and reduced costs