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
22.3
arena score
#32 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Stable Diffusion 3.5 Medium
16.9
arena score
#57 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [klein] 4B
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 [klein] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent photo-realism with natural textures on the book cover and wooden table.
- + Accurate physical grounding with the sphere resting on the bottom of the cube.
- + Sophisticated lighting and reflections consistent with a window light source.
- − The plant is very out of focus compared to the requested 'visible through the glass' detail.
Stable Diffusion 3.5 Medium
- + Successfully positions the plant behind the cube as requested.
- + Bright, clear colors and strong contrast.
- − The sphere is floating unnaturally in the center of the cube.
- − The perspective of the cube's top and the book is slightly warped and inconsistent.
- − Visible white border artifacts along the edges of the book.
Verdict: FLUX.2 [klein] 4B produces a significantly more realistic and physically grounded image with convincing textures and lighting. While Stable Diffusion 3.5 Medium follows the spatial instruction for the plant better, it fails on basic physics by showing a floating sphere and exhibits lower image quality with warped geometry.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent anatomical realism in the man's face and hands
- + High coherence in the bicycle's mechanical structure
- + Strong adherence to the shallow depth of field and rain effects
- − Lacks the requested 'motion blur' from the passing cars
- − Framing feels a bit too balanced and professional for the 'imperfect framing' prompt
Stable Diffusion 3.5 Medium
- + Successfully captured the 'imperfect framing' with a more candid, snapshot aesthetic
- + Good lighting and reflections on the wet pavement
- + Colors feel more cinematic and atmospheric
- − The bicycle frame is physically impossible and lacks mechanical coherence
- − The man's hands are poorly rendered and blurry
- − Missing motion blur from passing cars
Verdict: FLUX.2 [klein] 4B is the superior image due to its high level of detail and physical coherence; the bicycle and the man's features are rendered with impressive realism. Stable Diffusion 3.5 Medium captures the candid mood and 'imperfect framing' slightly better, but fails significantly on anatomical and mechanical details, resulting in a distorted bicycle and hands.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent anatomical realism and skin texture
- + Highly detailed and legible plate armor engraving
- + Clear adherence to the braided hair with beads requirement
- − Lighting feels a bit flat and artificial despite the torches
- − The 'battle-worn' effect is relatively clean compared to the prompt
Stable Diffusion 3.5 Medium
- + Dramatic cinematic lighting and color contrast
- + More convincing 'battle-worn' appearance with grit and sweat
- + Intense, lifelike eye rendering
- − Missing the specific detail of beads in the hair
- − Armor engraving is less distinct and slightly more blurred
- − Slightly messy hair rendering where strands blend into the background
Verdict: FLUX.2 [klein] 4B produces a much cleaner image with superior detail in the armor and clear adherence to the beads requirement, though it feels somewhat staged. Stable Diffusion 3.5 Medium captures the 'battle-worn' mood and cinematic lighting much more effectively, but fails on the specific detail of the beads. FLUX is preferred for its technical clarity and prompt precision.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent photographic quality and color vibrancy in the food images
- + Professional typography and hierarchy that resembles a real menu
- + Clean, balanced grid layout with clear sectioning
- − Nonsense words used in section headers (e.g., 'FRAZZA', 'MAINS' becomes 'MDAINS')
- − Font choices in some areas feel a bit dated rather than purely modern minimalist
Stable Diffusion 3.5 Medium
- + Successfully includes numerous food photos in a structured grid
- + Consistent use of a single font style throughout the entire design
- − Poor image quality with significant artifacts in the food photography
- − Text is cramped and poorly aligned with many overlapping or illegible characters
- − The layout feels cluttered and lacks the professional finish requested in the prompt
Verdict: FLUX.2 [klein] 4B produces a much more professional and visually appealing menu with high-resolution food photography and a logical layout. While both models struggled with generating actual readable text, FLUX.2 [klein] 4B followed the 'clean professional' aesthetic whereas Stable Diffusion 3.5 Medium produced low-quality, muddy images and messy typography.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent photorealistic lighting and texture on the burger and starburst.
- + Perfectly renders secondary text 'LIMITED TIME ONLY' and price.
- + Vibrant and dynamic composition with glowing embers.
- − Severely garbled main title text ('MAAC AGIR BUIRGGER').
- − Burger is not an 'exploded' view; it is a whole assembled burger.
Stable Diffusion 3.5 Medium
- + Main title text 'MAGIC BURGER' is spelled correctly and clearly.
- + Good adherence to the fiery background and glowing price element.
- + Clean and professional font choices for the advertisement.
- − Failed to provide an 'exploded' view where components are suspended separately.
- − Burger texture and lighting look slightly less realistic than Model A.
- − Starburst graphic is a bit simplistic and lacks the requested fiery glow.
Verdict: Both models failed to deliver the 'exploded' view requested in the prompt, opting instead for a whole burger suspended in air. FLUX.2 [klein] 4B provides much better visual quality and realistic lighting, but its primary text rendering is incoherent. Stable Diffusion 3.5 Medium is the overall winner because it successfully renders all requested text accurately while still maintaining high visual appeal.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent chalk texture and realistic handwriting style.
- + High accuracy in spelling the majority of the complex prompt details.
- + Consistent cursive style that looks authentically hand-drawn.
- − Several typos in the small text and menu items like 'Musheram' and 'Ootrpous'.
- − An extra 'S' in 'SPECIALS' and a space issue in the title.
Stable Diffusion 3.5 Medium
- + Creative use of chalk outline and shading effects.
- + Good composition with various decorative elements and framing.
- − Severely failed text rendering with almost all words being unintelligible or misspelled.
- − Failed to follow the date and specific item instructions properly.
- − The text style looks more like digital vector art than realistic chalk handwriting.
Verdict: FLUX.2 [klein] 4B is the clear winner as it successfully rendered most of the requested text with a highly realistic chalk texture, despite some minor spelling errors. Stable Diffusion 3.5 Medium failed significantly on the prompt's core requirement of legible, specific text, resulting in a board filled with 'gibberish' characters.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [klein] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent anatomical details on the horse and space suit.
- + Great cinematic lighting and texture on the starfield and Earth atmosphere.
- + High image clarity with realistic textures.
- − Failed the negative constraint; the astronaut is on top of the horse, not vice versa.
Stable Diffusion 3.5 Medium
- + Dynamic composition with the curve of the Earth and the horse's flowing mane.
- + Clean rendering of the astronaut's helmet and gear.
- − Failed the negative constraint; the astronaut is riding the horse instead of the reverse.
- − Significant anatomical errors on the horse's legs, which appear jointless and elongated.
Verdict: Both FLUX.2 [klein] 4B and Stable Diffusion 3.5 Medium failed the complex spatial instruction to put the horse on top of the astronaut. However, FLUX.2 produced a much higher quality image with realistic anatomy and cinematic lighting, whereas Stable Diffusion 3.5 Medium suffered from severe distortions 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 [klein] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent photorealism in lighting and textures
- + Perfect adherence to prompt details like the phone and bored expression
- + Very coherent interior perspective and scale
- − The capybara's paws look slightly artificial/flat in their grip
Stable Diffusion 3.5 Medium
- + Great lighting on the capybara's fur
- + Good character expression on the capybara
- − Failed to include the phone in the passenger's hand
- − The passenger is out of focus and looks confused rather than bored
- − Composition is a bit cramped with the dashboard taking up too much foreground
Verdict: FLUX.2 [klein] 4B is the clear winner as it followed every instruction in the prompt, including the specific behavior of the businesswoman and the placement of her phone. Stable Diffusion 3.5 Medium missed several key details and created a less convincing photorealistic scene with a crowded composition.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Atmospheric cinematic lighting with a very high visual polish.
- + Includes the requested thorn and spiderweb border in a cohesive way.
- + Excellent layout balance with a beautiful central jack-o-lantern.
- − Significant spelling errors in the main title ('Hallbwom Party niisation').
- − Missing the 'Time' detail and minor year typo in the date.
Stable Diffusion 3.5 Medium
- + Distinct parchment scroll style as requested in the prompt.
- + Contains all required text fields including Location and Date.
- + Clear, bold gothic font choice.
- − Very poor spelling throughout ('Halloweeen Inviloween', 'Aches', 'Timme').
- − Composition feels less 'cinematic' and more like clip art.
- − The jack-o-lanterns are not centrally located as requested.
Verdict: Both models struggled significantly with the text rendering, but FLUX.2 [klein] 4B is the far superior image in terms of aesthetic quality and cinematic lighting. While Stable Diffusion 3.5 Medium included the parchment look, its execution was cluttered and the spelling errors were more distracting compared to the professional layout of FLUX.2.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent texture on the salmon and rice grains.
- + Proper 45-degree isometric perspective and clean presentation.
- + Rich, realistic PBR materials that stay true to the '3D cartoon' aesthetic.
- − Text is slightly misspelled as 'SUSH'.
- − The flag icon is stylized incorrectly (looks more like the Austrian flag).
Stable Diffusion 3.5 Medium
- + Text 'JAPAN' and 'SUSHI' are both spelled correctly.
- + Nice glossy materials on the fish roe.
- − The word 'SUSHI' is excessively large and overrides the requested 'bold' hierarchy for 'JAPAN'.
- − Missing the requested flag icon.
- − The perspective is slightly flatter than the requested 45-degree isometric view.
Verdict: FLUX.2 [klein] 4B followed the stylistic and composition instructions much more closely, delivering a high-quality 3D diorama despite a minor misspelling in the text. Stable Diffusion 3.5 Medium managed the text spelling better but failed to include the flag icon and had a less refined isometric composition. FLUX.2 is preferred for its superior texture work and adherence to the layout requested.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Excellent adherence to lighting requests with clear god rays and dew sparkles.
- + Correctly generated most of the animals including a fox and a golden retriever puppy.
- + Fine fur texture and realistic eye reflections.
- − Missed the baby bunny completely, substituting it with a second kitten.
- − Large butterfly wings appear to be floating or disconnected in some areas.
Stable Diffusion 3.5 Medium
- + Successfully captured the warm golden sunrise vibe.
- + High color saturation that enhances the joyful feel.
- + Good rendering of the red fox kit features.
- − Failed to include the baby bunny and the tabby kitten is morphologically ambiguous.
- − The fur texture looks more like a digital painting than the requested hyper-photorealistic style.
- − Anatomical issues with the puppy's paws merging into the grass.
Verdict: FLUX.2 [klein] 4B produced a much more realistic image with convincing lighting and textures, though it failed to include the bunny. Stable Diffusion 3.5 Medium struggled with the photorealistic requirement, producing a more stylized, illustrative output with fewer of the requested animals. FLUX.2 is preferred for its superior clarity and better adherence to the environment and lighting descriptions.
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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Clean vector emblem style that matches the minimalist request.
- + Correct inclusion of all prompt elements like the steam and the banner.
- + Solid rendering of the established year 'Est. 1720'.
- − Spelling error in the main text rendering it as 'FLAXTION'.
- − Redundant repetition of text below the banner.
Stable Diffusion 3.5 Medium
- + Intricate vintage illustration style with good texture.
- + Creative integration of the cloche dome as a frame.
- − Failed to spell 'Florian' correctly, adding an extra 'r'.
- − Incorrect date 'Est 170' instead of '1720'.
- − The composition is overly busy and lacks the requested minimalism.
Verdict: FLUX.2 [klein] 4B followed the stylistic 'minimalist vector' instructions much better than Stable Diffusion 3.5 Medium, which produced a very dense and complex illustration. Although FLUX.2 failed the primary name spelling by turning 'Florian' into 'Flaxtion', it remained more faithful to the layout and date requested, whereas Stable Diffusion 3.5 Medium failed both the spelling and 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] 4B
Stable Diffusion 3.5 Medium
AI Judge Analysis
FLUX.2 [klein] 4B
- + Successfully includes the specific requested steps like launch, trajectory, and landing.
- + Adheres well to the requested flat-vector style with crisp icons.
- + Follows the NASA-inspired color palette effectively.
- − Text rendering is very poor with significant misspellings throughout.
- − Layout of the steps is disorganized and lacks a logical flow for an infographic.
Stable Diffusion 3.5 Medium
- + Strong composition with a sense of movement across the frame.
- + The color palette is cohesive and visually appealing.
- + Text for section headings like 'Landing' is more legible than the competitor's.
- − Failed to follow the specific iconography requested (e.g., Saturn V icon, Lunar Orbit ring).
- − Included several redundant or abstract circular icons that do not depict the requested mission phases.
- − The infographic structure is confusing and doesn't clearly show a progression of steps.
Verdict: FLUX.2 [klein] 4B is the likely winner because it actually attempted to follow the sequential instructions and iconography requested in the prompt, such as the Saturn V rocket and the lunar module. Stable Diffusion 3.5 Medium created a more artistically pleasing composition but failed on the specific symbols and logical layout needed for an infographic, essentially ignoring the technical requirements of the steps.
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