Black Forest Labs' compact, open-source image generation model with sub-second inference, optimized for production and near real-time applications with multi-reference support
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [klein] 4B
#32 of 62 in Text-to-Image
Stable Diffusion 3.5 Large
#28 of 62 in Text-to-Image
Where the votes landed
FLUX.2 [klein] 4B
0%
win rate
Ties
0%
Stable Diffusion 3.5 Large
0%
win rate
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Perfect adherence to spatial instructions including the placement of the book on top of the cube.
- + Realistic rendering of soft light and shallow depth of field.
- + Accurate refraction and reflections within the glass material.
Stable Diffusion 3.5 Large
- + High resolution and Sharp focus on the object textures.
- + Good rendering of the glass cube edges and surface dust details.
- − Failed the spatial prompt: the red book is inside/under the sphere instead of on top of the cube.
- − The sphere appears to be floating unnaturally.
Verdict: FLUX.2 [klein] 4B followed the prompt instructions perfectly, correctly placing the red book on top of the glass cube and the sphere inside. Stable Diffusion 3.5 Large failed the spatial challenge, placing the book at the bottom and failing to put anything on top of the cube despite producing a sharp image.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent depiction of raining atmosphere with visible streaks and wet texture.
- + Subject's posture and skin texture feel very natural and realistic.
- + Beautiful bokeh and shallow depth of field effects.
- − Major anatomical and mechanical error where the bike frame passes through the man's leg.
- − The bike's drivetrain (chain and gears) is physically impossible.
- − Lack of requested motion blur on passing cars.
Stable Diffusion 3.5 Large
- + Better structural integrity of the bicycle and character interaction.
- + Captures the requested weathered skin texture on the arms and face.
- + Includes subtle motion blur on the background car as requested.
- − The man's right arm (hands) is poorly rendered and blobby.
- − The background elements (bus/car) are slightly distorted or simplified.
- − The rain effect looks a bit like a static overlay rather than a 3D scene element.
Verdict: Stable Diffusion 3.5 Large is the winner primarily because FLUX.2 [klein] 4B suffers from a significant glitch where the bicycle frame clips through the man's leg. While FLUX.2 has a more cinematic lighting and atmospheric rain effect, Stable Diffusion 3.5 Large followed more specific prompt instructions like the motion blur on vehicles and maintains a more logical physical composition between the subject and the bike.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent depiction of small colorful beads in the braids
- + Clear, high-contrast engravings on the chest plate
- + Very lifelike eyes with realistic wetness and reflections
- − The lighting feels slightly artificial, with the torches looking like static background props
- − The skin texture is a bit smooth despite the requested dirt and scars
Stable Diffusion 3.5 Large
- + Intricate armor engraving and highly detailed chainmail/fabric underlayers
- + More naturalistic and cinematic lighting integration
- + Higher level of visible dirt and skin texture suitable for a 'battle-worn' descriptor
- − The hair braids appear thicker and lack the specifically requested 'small beads'
- − The depth of field is very shallow, making some incidental sparks look like blurry artifacts
Verdict: Both models followed the prompt well, but Stable Diffusion 3.5 Large delivered a more cinematic and gritty result that better captured the 'battle-worn' aesthetic through skin texture and complex layering. While FLUX.2 [klein] 4B was more literal with the 'small beads' instruction and had excellent facial symmetry, Stable Diffusion 3.5 Large's superior armor detail and lighting make it the stronger image.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Stronger adherence to the 'grid' food photography layout.
- + Clean use of white space that feels like a modern menu.
- + Higher quality food photography that looks appetizing and professional.
- − Nonsense text and misspellings for section titles (e.g., 'Fraza', 'Mdains').
- − The font weights are slightly inconsistent across sections.
Stable Diffusion 3.5 Large
- + Excellent typography for the main 'Menu' header.
- + The layout has a more sophisticated editorial feel.
- + Captures the 'bold sans-serif' requirement very well.
- − The 'grid' format is localized to the columns rather than an integrated layout.
- − Contains significant spelling errors in headers (e.g., 'Appetizrs', 'Maimaes').
- − The vertical composition is less practical for a standard restaurant menu.
Verdict: FLUX.2 [klein] 4B followed the prompt's layout instructions more closely, providing a standard, highly readable menu format with professional food photography. Stable Diffusion 3.5 Large produced a more artistic, editorial design with superior main headers, but the overall layout is less functional as a menu and the food photography is more repetitive.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent typography rendering for the secondary messages.
- + High photorealistic detail on the burger textures.
- + Good use of the starburst graphic for the price.
- − Serious spelling error in the primary title 'MAGIC BURGER'.
- − The burger is not 'exploded' with components suspended separately as requested; it is a stacked burger.
Stable Diffusion 3.5 Large
- + Strong visual aesthetics with the fire and coal elements.
- + High detail on the melting cheese and grill textures.
- + Dynamic lighting that fits the fiery theme.
- − Failed to include any of the requested text/titles.
- − Did not fulfill the 'exploded burger' requirement, showing a mostly assembled stack.
- − Included extra ingredients like pickles not specifically highlighted in the prompt.
Verdict: Both models failed the specific structural instruction of an 'exploded' burger, instead providing a standard floating stack. FLUX.2 [klein] 4B is better as a complete advertisement because it included the requested text and price starburst, despite the significant misspelling in the main title. Stable Diffusion 3.5 Large produced a more atmospheric background with the coals and flames but ignored all text-related prompt instructions.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent chalk texture and natural handwriting style across all text.
- + Accurate transcription of the long-form prompt, including the date 2026.
- + Good balance and realistic smudging on the chalkboard surface.
- − Several spelling errors like 'Truffel', 'Musheram', 'Ootpous', and 'Buter'.
- − Slight spacing issue in the top header word 'SPECIALS'.
Stable Diffusion 3.5 Large
- + Great environmental composition and café atmosphere.
- + Clean typography that fits a professional café aesthetic.
- + High resolution and realistic interior lighting.
- − Failed significantly on prompt adherence regarding text content and date (shows 2024).
- − Text is largely illegible gibberish in the sub-sections.
- − The font appears more digital/uniform than the requested natural chalk handwriting.
Verdict: FLUX.2 [klein] 4B is the clear winner for its superior prompt adherence and realistic chalk rendering, successfully including the specific year (2026) and the full menu item descriptions. While Stable Diffusion 3.5 Large produced a more aesthetically pleasing background environment, it failed to deliver on the core text requirements, resulting in incorrect dates and illegible text.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent anatomical clarity of the horse and astronaut
- + High resolution with sharp textures on the spacesuit and horse's coat
- + Clean composition with a beautiful starlit background
- − The lighting on the horse and rider doesn't match the planet-based light source below
- − The 'surreal' aspect is limited strictly to the juxtaposition of the horse in space
Stable Diffusion 3.5 Large
- + Stronger cinematic atmosphere with dynamic lighting and stardust effects
- + Captures the 'surreal' instruction more effectively with the cloud-like space trails
- + Great sense of movement and energy in the composition
- − Significant anatomical issues with the horse's front legs and harness
- − Blurrier textures and more artifacts compared to Model A
Verdict: FLUX.2 [klein] 4B produces a much cleaner and more physically coherent image with superior detail in the suit and horse, but it feels a bit static. Stable Diffusion 3.5 Large offers a more creative and cinematic interpretation of 'surreal', but it suffers from anatomical distortions and lower image clarity. FLUX.2 is the likely winner for its overall polish and professional finish.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent adherence to the complex multi-subject scene including the passenger.
- + Photorealistic lighting and interior textures.
- + Nails the specific 'bored' expression requested for the passenger.
Stable Diffusion 3.5 Large
- + High saturation and vibrancy in colors.
- + Detailed rendering of the capybara's fur and whiskers.
- + Clear focus on the central character.
- − Completely failed to include the passenger in the back seat.
- − The car interior appears oversized compared to the driver.
- − Anthropomorphized the capybara with human legs and pants, which was not requested.
Verdict: FLUX.2 [klein] 4B follows every detail of the prompt, successfully capturing the surreal interaction between the capybara driver and the bored businesswoman passenger. Stable Diffusion 3.5 Large failed the primary compositional requirement by omitting the passenger entirely and adding unrequested elements like human legs to the capybara.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Atmospheric cinematic lighting with a moody color palette
- + Excellent border integration using webs and thorns
- + Strong central focus on a high-quality jack-o-lantern
- − Significant spelling errors in the title text
- − Missing the year digit '2' in the date and misspelled 'Location'
- − The text is generally distorted
Stable Diffusion 3.5 Large
- + Great 'burnt parchment' edge effect
- + Accurate spelling of the main title and most of the banner text
- + Dynamic composition with a large moon and twisted trees
- − Failed to include the specific event details (Date, Time, Location) at the bottom
- − Text alignment and font choice feel less 'gothic' and more modern/digital
- − More of a collage feel rather than a polished cinematic poster
Verdict: FLUX.2 [klein] 4B produces a much more atmospheric and visually coherent image that perfectly captures the 'moody night sky' and 'thorny border' requested, but it fails significantly on text legibility. Stable Diffusion 3.5 Large succeeds in spelling the main title correctly and provides a great parchment aesthetic, but it completely ignores the specific event details requested for the bottom of the invitation. FLUX.2 is the likely winner for its superior artistic composition and adherence to the visual style, despite the typos.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent adherence to the 'minimal' aesthetic and clean background.
- + Text is rendered directly on the background as requested.
- + The materials look like soft, refined 3D miniatures with a clean PBR sheen.
- − The flag icon is incorrect (looks like Austria rather than Japan).
- − The sushi fish textures are a bit too smooth and plastic-like.
Stable Diffusion 3.5 Large
- + Successfully included a correct Japanese flag.
- + Beautifully detailed textures on the rice and fish.
- + Great interpretation of the 'diorama base' with a layered platform.
- − Failed the text placement instruction by putting it on a sign rather than 'at top-center' on the background.
- − Ignored the instruction for 'minimal garnish', resulting in a cluttered composition.
- − The 45-degree isometric angle is slightly off compared to Model A.
Verdict: FLUX.2 [klein] 4B followed the stylistic and layout instructions much more accurately, creating a clean, centered isometric image with the text in the correct location. Stable Diffusion 3.5 Large produced much better individual textures and a correct flag, but failed on the specific layout requirement (text placement) and the 'minimal' constraint, resulting in a busy scene instead of a clean miniature diorama. FLUX.2 is the winner for better instruction adherence despite the flag error.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent fur texture and lighting on the animals
- + Clear and vibrant wildflower meadow following the 8K aesthetic
- − Completely missed the baby bunny requirement, substituting for a second kitten
- − The fox's anatomy is slightly elongated and awkward
Stable Diffusion 3.5 Large
- + Successfully included all four requested animals: puppy, kitten, bunny, and fox kit
- + Captures the tumbling/chasing action more dynamically than Model A
- + Beautiful use of bokeh and backlighting to create the sunrise atmosphere
- − Slightly lower resolution/clarity in the background compared to Model A
- − The kitten's facial features look slightly anthropomorphized
Verdict: Stable Diffusion 3.5 Large is the clear winner as it adhered to all parts of the prompt, specifically by including the bunny which FLUX.2 [klein] 4B omitted. While FLUX.2 has slightly sharper details, Stable Diffusion 3.5 Large better captured the joyful, tumbling movement requested.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Clean vector emblem style with professional line work
- + Subtle, realistic paper texture on the background
- + Excellent color palette consistency
- − Included significant spelling errors in the brand name ('FLAXTION')
- − Redundant 'Est. 1720' text appearing twice
Stable Diffusion 3.5 Large
- + Correctly spelled the brand name (with a minor accent addition)
- + Intricate vintage border details and ornamentation
- + Good interpretation of the cloche dome with visible contents/steam
- − Typography within the banner is slightly warped
- − The steam graphic is a bit heavy-handed compared to the minimalist request
- − A few artifacts in the vector shapes under the cloche
Verdict: Stable Diffusion 3.5 Large is the winner primarily because it successfully spelled the brand name correctly, whereas FLUX.2 [klein] 4B hallucinated a different name ('FLAXTION'). While FLUX.2 offered a cleaner vector aesthetic, Stable Diffusion 3.5 Large delivered a more complete and accurate response to the prompt's specific text requirements.
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.”
AI Judge Analysis
FLUX.2 [klein] 4B
- + Successfully captures a flat vector aesthetic with clean iconography.
- + Generally follows the requested NASA-inspired color palette.
- + Includes specific names like Aldrin and Collins as requested by the prompt context.
- − Poor typography with significant misspellings in the header and labels.
- − The Saturn V rocket anatomy is anatomically incorrect for a vector style.
- − Icon layout is a bit scattered and lacks a clear linear flow for the 6 steps.
Stable Diffusion 3.5 Large
- + Highly detailed infographic layout that feels professional and dense.
- + Better integration of the 'navy' color as a primary background element.
- + Good use of fine lines and technical-looking callouts.
- − Incorrectly includes a space shuttle instead of the Saturn V rocket for the Apollo 11 mission.
- − The text is mostly illegible gibberish despite the complex layout.
- − The composition is cluttered and fails to clearly present the 6 requested sequential steps.
Verdict: FLUX.2 [klein] 4B better follows the specific iconography requirements and the flat-vector style, although its text rendering is poor. Stable Diffusion 3.5 Large creates a more visually complex 'poster' feel but fails significantly on historical accuracy by depicting a space shuttle for a mission that used a Saturn V rocket. FLUX.2 is the winner for better adhering to the design style and subject matter logic.
Explore each model
Stability AI's 8.1-billion parameter Multimodal Diffusion Transformer (MMDiT) text-to-image model featuring improved image quality, typography, complex prompt understanding, and resource-efficiency