Distilled version of SD 3.5 Large that generates high-quality images in just 4 steps, offering faster inference and reduced costs
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
Stable Diffusion 3.5 Large Turbo
10.7
arena score
#61 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Stable Diffusion 3.5 Medium
16.8
arena score
#58 of 62 in Text-to-Image
Vote tally
Where the votes landed
Stable Diffusion 3.5 Large Turbo
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent crispness and high resolution
- + Very accurate lighting and shadows
- + Clean geometric rendering
- − Failed the spatial prompt by putting the book inside the cube
- − Failed the spatial prompt by putting the sphere on top of the book
Stable Diffusion 3.5 Medium
- + Perfectly followed the spatial relationships defined in the prompt
- + Good glass refraction effects
- + Natural, soft lighting
- − The sphere appears to be floating unnaturally without support
- − Slightly lower resolution and more blur compared to model A
- − Book texture is a bit fuzzy
Verdict: While Stable Diffusion 3.5 Large Turbo produces a more polished and visually striking image, it failed significantly on the spatial logic of the prompt, placing the book inside and the sphere on top. Stable Diffusion 3.5 Medium followed every single instruction perfectly, correctly positioning the sphere inside, the book on top, and the plant behind, making it the winner for prompt adherence.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean representation of the red bicycle
- + Accurate depiction of an elderly man with realistic age spots and posture
- − Static composition fails to show the requested motion blur of passing cars
- − Image looks like it was taken in a studio with a backdrop rather than a candid street setting
- − The man is floating/disconnected from the pavement shadow-wise
Stable Diffusion 3.5 Medium
- + Excellent adherence to cinematic atmosphere and rain effects
- + Captures the 'candid' street photography aesthetic with realistic pavement reflections
- + Includes the motion blur and environmental context requested in the prompt
- − Anatomical issues with the man's hands and fingers
- − Bicycle geometry is inconsistent with some warped spokes and frame elements
Verdict: Stable Diffusion 3.5 Medium far more effectively captures the mood and technical requirements of the prompt, successfully translating the 'candid street photo' and 'wet pavement reflection' instructions into a believable scene. While Stable Diffusion 3.5 Large Turbo has cleaner subject geometry, it fails to incorporate the background motion blur and rain atmosphere, resulting in a flat, artificial look.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent clarity on the ornate engraving of the plate armor.
- + Clean, sharp details on the facial features and hair braids.
- + Good composition with a clear focus on the subject and a creamy bokeh background.
- − Skin texture appears overly smooth and 'plastic' like a digital painting rather than lifelike.
- − The 'dirt and scars' look more like stylistic face paint than realistic battle wear.
- − Missed the request for small beads in the hair braids.
Stable Diffusion 3.5 Medium
- + Highly realistic skin texture with lifelike pores, sweat, and genuine-looking dirt/scars.
- + Strong adherence to the 'warm torchlight' and 'bokeh sparks' atmosphere.
- + The hair braids include the requested beads and a much more complex, battle-worn appearance.
- − The eyes, while intense, have slight artifacts in the iris symmetry.
- − The overall image has a slightly grainier texture compared to the smoothness of Image A.
Verdict: Stable Diffusion 3.5 Medium is the winner as it much more accurately captures the 'battle-worn' and 'lifelike' aspects of the prompt, including specific details like the beads in the hair and the grit on the skin. While Stable Diffusion 3.5 Large Turbo produced a cleaner digital art look, it failed to provide the realistic texture and atmospheric lighting that Medium mastered.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent typography with legible headers for 'Pizza' and 'Mains'.
- + Very clean, bright white background and professional layout.
- + High-resolution dish photography with vibrant colors.
- − The food photos look slightly artificial/CGI rather than realistic culinary photography.
- − Composition feels more like a flat-lay collage than a functional single-page menu.
Stable Diffusion 3.5 Medium
- + Follows the requested grid layout for food photos very effectively.
- + Captures a more realistic 'restaurant menu' aesthetic with price columns and item descriptions.
- + Food photography looks more authentic and less processed.
- − The text is largely illegible gibberish.
- − The font choice looks a bit 'bubbly' or hand-drawn rather than the requested bold sans-serif.
- − The bottom half of the menu looks cluttered and messy.
Verdict: Stable Diffusion 3.5 Large Turbo produces a much cleaner and more professional-looking graphic design with legible typography, though the food photos look somewhat plastic. Stable Diffusion 3.5 Medium creates a more realistic food menu layout with better grid integration, but suffers from severe text corruption and a lack of polish. Stable Diffusion 3.5 Large Turbo is the winner for its clarity, professional finish, and adherence to the 'minimalist' aesthetic.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Strong composition with a vibrant, fiery atmosphere.
- + Good execution of the 'glowing embers' and smoke effects.
- − Completely failed to include any of the requested text.
- − The burger is not 'exploded' with components suspended in mid-air as requested.
- − The burger looks like an illustration rather than photorealistic.
Stable Diffusion 3.5 Medium
- + Excellent adherence to the text requirements, including the price starburst.
- + Better photorealistic texture on the meat and bun.
- + Successful capture of the fiery background aesthetic.
- − Failed to create an 'exploded' view; the burger is mostly intact and just floating.
- − The text does not have the 'fiery, glowing effect' requested, appearing as a standard overlay.
Verdict: Stable Diffusion 3.5 Medium is the winner because it successfully integrated all the requested text elements with high accuracy, whereas Stable Diffusion 3.5 Large Turbo omitted them entirely. While neither model fully achieved the 'exploded' component look, Medium provided a more realistic texture for a food advertisement.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean and modern café aesthetic
- + Text is generally legible and clearly separated
- + Good lighting and depth of field in the environment
- − Text is full of typos (e.g., 'trulale', 'ocopus', 'risctto')
- − Failed to include the specific date 2026
- − Handwriting looks more like a digital font than real chalk
Stable Diffusion 3.5 Medium
- + Texture looks much more like authentic chalk on a slate board
- + Better adherence to the 'elegant cursive' request for parts of the text
- + Handwriting has realistic variations and smudging
- − Text is highly garbled and mostly illegible
- − The layout is cluttered and contains non-sensical symbols
- − Failed to render the specific menu item names correctly
Verdict: Stable Diffusion 3.5 Large Turbo produces a much cleaner and more aesthetically pleasing image, though it fails significantly on spelling. Stable Diffusion 3.5 Medium approximates the texture of chalk and handwriting much better but fails entirely on legibility. Stable Diffusion 3.5 Large Turbo is preferred for overall coherence despite the spelling errors.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean, high-resolution textures on the spacesuit and horse.
- + Excellent lighting that creates a cinematic feeling with the planetary backdrop.
- − Completely failed to follow the 'horse on top' spatial instruction.
- − The horse's front legs have significant anatomy errors/merging issues.
Stable Diffusion 3.5 Medium
- + Strong composition with a clear starfield and Earth perspective.
- + High level of detail on the patches and equipment of the astronaut suit.
- − Completely failed to follow the 'horse on top' spatial instruction.
- − Anatomical issues with the horse's legs and the astronaut's leg placement.
Verdict: Both Stable Diffusion 3.5 Large Turbo and Stable Diffusion 3.5 Medium failed the negative constraint of having the horse positioned on top of the astronaut, instead defaulting to the standard horse-riding trope. Stable Diffusion 3.5 Large Turbo is slightly better due to its cleaner lighting and more cohesive cinematic style, despite both models struggling with horse anatomy.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Strong composition from the driver's perspective
- + The capybara's hands are realistically placed on the steering wheel
- + Excellent rendering of the taxi's exterior and interior textures
- − The passenger in the back is very blurry and lacks clear detail
- − The passenger is not visibly looking at a phone as requested
Stable Diffusion 3.5 Medium
- + Front-facing perspective creates a more engaging portrait of the capybara
- + Includes the passenger in a more centered, visible position
- + The capybara's expression is very calm and fits the 'professional' prompt
- − The capybara's paws are morphed into the steering wheel/clothes and look like bird talons
- − The passenger is not looking at a phone
- − Severe anatomical errors where the driver's jacket meets the seat
Verdict: Stable Diffusion 3.5 Large Turbo produces a much more coherent image with believable lighting and anatomical structure, even though it misses the specific 'looking at phone' detail. Stable Diffusion 3.5 Medium struggles significantly with the capybara's paws and the overall structure of the vehicle interior, making the scene feel less grounded.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean and polished graphic design style
- + Strong central jack-o-lantern illustration
- + Effective use of negative space for invitation layout
- − Failed to include most of the requested text including the specific details and scroll banner
- − The parchment look is very subtle compared to the request for dark parchment
- − Text that is present is cut off at the top
Stable Diffusion 3.5 Medium
- + Successfully attempted all requested text fields including date, time, and location
- + Strong atmosphere with the moody night sky and twisted trees
- + Captured the 'torn parchment' aesthetic well
- − Numerous spelling errors in the text like 'Halloweeen' and 'Invilloween'
- − The composition is a bit cluttered with the pumpkins overlapping the trees
- − The text alignment and formatting is inconsistent
Verdict: Stable Diffusion 3.5 Medium is the winner for following the complex prompt instructions, including attempting the specific date, time, and location details which Stable Diffusion 3.5 Large Turbo ignored. While Stable Diffusion 3.5 Large Turbo produced a cleaner graphic, it failed as an invitation by omitting the core event information.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent 3D miniature/diorama aesthetic following the isometric prompt.
- + High clarity with realistic PBR material textures for the rice and seaweed.
- + Accurate text rendering for 'JAPAN' in a bold 3D style.
- − Spelling error in the word 'SUSHI' (rendered as 'SIIHI').
- − The flag icon does not look like the Japanese flag.
Stable Diffusion 3.5 Medium
- + Clean layout with a solid light blue background as requested.
- + Accurate spelling for both 'JAPAN' and 'SUSHI'.
- + Gentle lighting and soft textures that match the cartoon/miniature vibe.
- − Missing the raised diorama base requested in the prompt.
- − Lacks the requested flag icon.
- − The text is missing a letter stroke on the 'S' of 'SUSHI'.
Verdict: Stable Diffusion 3.5 Large Turbo followed the stylistic cues for an 'isometric miniature 3D scene' much better than the Medium model, providing a true diorama feel. While Stable Diffusion 3.5 Medium had better spelling accuracy for the main text, it failed to include the diorama base and flag icon, resulting in a simpler composition that didn't fully realize the prompt.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Strong backlighting creates a beautiful glow on the animals' fur.
- + The central puppy has a soft, high-quality rendered appearance.
- + Vibrant high-contrast colors in the meadow.
- − Failed to include all requested animals, missing the baby bunny and red fox.
- − The eyes and anatomy of the kitten on the left are slightly distorted.
- − Anatomical merging between the animals' limbs.
Stable Diffusion 3.5 Medium
- + Successfully included the golden retriever, red fox kit, and a kitten.
- + Excellent warmth and golden sunrise lighting with visible god rays as requested.
- + The fur textures are very detailed and distinct across the different species.
- − Failed to include the baby bunny, only showing three of the four requested animals.
- − The kitten's ears are unusually large and fox-like, making it less distinct from the fox kit.
- − The grass in the foreground is somewhat blurry compared to the subject focus.
Verdict: Stable Diffusion 3.5 Medium is the winner as it adhered closer to the complex list of animals by including the fox kit, whereas Stable Diffusion 3.5 Large Turbo only rendered generic domestic-looking pets. Stable Diffusion 3.5 Medium also captured the specific 'god rays' lighting request much more effectively, creating a more cohesive and wholesome atmosphere.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent typography with only minor character artifacts
- + Clear depiction of the cloche dome requested
- + Strong adherence to the 'Est. 1720' banner prompt
- − The 'n' in Florian has some slight graphical glitches
- − The background texture is a bit blotchy rather than a subtle paper grain
Stable Diffusion 3.5 Medium
- + Beautiful hand-drawn engraving style
- + High aesthetic appeal for a vintage logo
- − Significant spelling errors including 'Florrian' and 'Est 170'
- − The 'Est 170' is not in a banner as requested
- − The cloche dome is less distinct and looks more like a rounded frame
Verdict: Stable Diffusion 3.5 Large Turbo followed the prompt instructions much more accurately, correctly including the banner and getting much closer to the requested text. Stable Diffusion 3.5 Medium has a more sophisticated artistic style, but fails the text and layout requirements of the prompt.
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.”
Stable Diffusion 3.5 Large Turbo
Stable Diffusion 3.5 Medium
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Strong aesthetic appeal with a clean, retro-modern vector style.
- + Effective use of the NASA-inspired color palette.
- + Good balance of negative space and infographic layout.
- − Failed to provide the specific 6-step sequence requested.
- − Significant text errors and gibberish in larger titles.
- − The iconography lacks direct correlation to the specific phases like 'Translunar' or 'Saturn V'.
Stable Diffusion 3.5 Medium
- + Successfully followed the requested 6-step vertical sequence.
- + Icons are more clearly aligned with the specific text entries.
- + Text rendering on labels like 'Launch' and 'Landing' is more legible.
- − The design is cluttered with excessive star-field dots.
- − Visual quality of the icons is inconsistent and slightly messy.
- − Technical text below headers is entirely illegible gibberish.
Verdict: Stable Diffusion 3.5 Medium is the winner because it successfully followed the complex instructional logic of the prompt, providing 6 distinct steps in chronological order. While Stable Diffusion 3.5 Large Turbo produced a more visually polished and artistic 'poster,' it failed to include the primary requirement of a 6-step infographic, instead condensing the information into four vague sections.
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