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
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
Stable Diffusion 3.5 Large
22.9
arena score
#29 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
Stable Diffusion 3.5 Large
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.”
Stable Diffusion 3.5 Large
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent photorealism with realistic glass reflections and surface imperfections.
- + Matches the soft window lighting request perfectly with natural shadows.
- + Creative interpretation showing parts of the background through the glass.
- − Failed the spatial instruction 'red book on top of the cube', placing it underneath instead.
- − The blue sphere appears to be sitting on the book rather than just 'inside the cube'.
Stable Diffusion 3.5 Large Turbo
- + High contrast colors and clean graphic style.
- + Correctly places the plant behind the cube as requested.
- + Good clarity on the wooden table texture.
- − Failed the spatial instruction 'red book on top of the cube', placing it inside/underneath.
- − The rendering is more CGI-like and less photorealistic compared to Model A.
- − Lighting feels a bit artificial despite the window shadows.
Verdict: Both models struggled with the complex spatial relationship of placing the book on top of the cube, as both opted to place the cube over the book. Stable Diffusion 3.5 Large is the preferred choice due to its superior photorealistic rendering, realistic glass transparency, and more sophisticated lighting, whereas Stable Diffusion 3.5 Large Turbo produces a flatter, more digital look.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent depiction of rain and wet surface reflections
- + Realistic skin texture and lifelike anatomical details
- + Successfully captures motion blur and a cinematic street photography aesthetic
- − The bike's red paint is a bit desaturated compared to the prompt's focus
- − Slightly messy hair rendering where it meets the background
Stable Diffusion 3.5 Large Turbo
- + Strong, vibrant red on the bicycle
- + Smooth background bokeh
- − Lacks the requested 'light rain' and 'motion blur' effects almost entirely
- − Anatomical errors in the hands and simplified plastic-like textures
- − Poor bicycle geometry with missing spokes and disconnected parts
Verdict: Stable Diffusion 3.5 Large is the clear winner as it captures the atmospheric conditions of rain and motion blur requested in the prompt. In contrast, Stable Diffusion 3.5 Large Turbo produces a more plastic, static image with significant anatomical errors and fails to render the rain or realistic reflections.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent representation of ornate engraved plate armor with high material realism.
- + Very lifelike skin texture with realistic dirt and subtle scarring.
- + Superior depth of field and bokeh effects that fit the cinematic theme.
- − The beads in the hair are extremely subtle and difficult to distinguish.
- − The lighting feels a bit more like natural daylight than warm torchlight.
Stable Diffusion 3.5 Large Turbo
- + Includes visible beads and bands in the braids as requested.
- + Stronger contrast in lighting with a vibrant warm glow on the side of the face.
- − Skin textures look overly smooth and plastic, lacking the 'lifelike' quality requested.
- − The blood/dirt on the face looks like digital paint rather than realistic wounds.
- − The armor engraving is less refined and looks more like a stylized graphic than metal.
Verdict: Stable Diffusion 3.5 Large is the clear winner due to its significantly higher level of detail and realism in both the plate armor and facial textures. While Stable Diffusion 3.5 Large Turbo followed the specific detail of 'beads' more closely, it suffered from a plastic-like skin finish and artificial-looking blood that undermined the 'battle-worn' prompt.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent photographic quality of food items
- + Strong adherence to the grid layout request
- + Professional editorial design feel
- − Significant spelling errors in the headings
- − The central menu block is slightly overcrowded with gibberish text
Stable Diffusion 3.5 Large Turbo
- + Legible category headers like Pizza
- + Captures the vibrant accents requested in the prompt
- + Clean separation of menu sections
- − Food items look somewhat artificial or plastic-like
- − Composition is a bit cluttered with floating elements
- − Physics of the shadows are inconsistent
Verdict: Stable Diffusion 3.5 Large (Image A) produces a much more realistic and professional-looking menu that fits the 'casual dining' vibe perfectly, despite having more spelling errors. Stable Diffusion 3.5 Large Turbo (Image B) has cleaner headline text and more vibrant colors, but the food photography feels like 3D renders rather than real meals, making it less appetizing and less professional.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent photorealistic texture on the meat and bun
- + Realistic lighting interaction between the fire and the food
- + Good sense of vertical motion with dripping cheese and flames
- − Completely failed to render any of the requested text ('MAGIC BURGER', etc.)
- − The burger is mostly intact rather than 'exploded' into separate floating layers
Stable Diffusion 3.5 Large Turbo
- + Dynamic composition with embers and smoke
- + Clean, stylized lighting
- − Failed to include all requested text elements
- − Food textures look plastic or illustrative rather than photorealistic
- − The burger is held together by a stick, failing the 'exploded' and 'suspended' part of the prompt
Verdict: Both models failed significantly on the text integration and the 'exploded' layout requested in the prompt. Stable Diffusion 3.5 Large is the better image due to its superior photorealistic textures and lighting, whereas Stable Diffusion 3.5 Large Turbo produced a more artificial, waxy-looking burger.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent realistic café atmosphere and composition
- + Text looks like genuine chalk with appropriate texture
- + Includes the correct date of 2024 (though prompt asked for 2026)
- − Significant spelling errors throughout the menu text
- − Text layout is cluttered and difficult to read
- − Did not use cursive for the title as requested
Stable Diffusion 3.5 Large Turbo
- + Layout is very clean and structured
- + Font style is closer to the requested handwriting/cursive style
- + Better lighting and vibrant colors in the frame
- − Text consists of nonsensical 'gibberish' words and symbols
- − The 'chalk' looks like a digital white brush rather than real chalk
- − Failed to include the specific year or complete prices
Verdict: Stable Diffusion 3.5 Large wins because it creates a much more believable and realistic scene, where the chalk texture actually looks authentic to a physical board. While Stable Diffusion 3.5 Large Turbo has a cleaner layout, its text is completely illegible and it lacks the realistic environmental depth of the non-turbo version.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent cinematic lighting and atmospheric dust effects.
- + Highly detailed textures on the astronaut's suit and the horse's coat.
- + Strong sense of scale and depth with the planet background.
- − Failed the negative constraint; the astronaut is riding the horse instead of vice versa.
Stable Diffusion 3.5 Large Turbo
- + Clean, high-contrast visual style.
- + Good anatomical rendering of the horse's mane and tail.
- − Failed the crucial negative constraint regarding position.
- − Floating moon element looks disconnected and lacks detail.
- − Lower level of texture detail compared to Model A.
Verdict: Both Stable Diffusion 3.5 Large and Stable Diffusion 3.5 Large Turbo failed the logic test inherent in the prompt, placing the astronaut on top of the horse despite explicit instructions otherwise. Stable Diffusion 3.5 Large is the preferred image as it significantly outperforms the Turbo version in terms of cinematic quality, realistic textures, and overall atmospheric depth.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent texture on the capybara fur and whiskers
- + Vibrant colors and high resolution
- + Captures the 'professional expression' very well
- − Completely fails to include the human businesswoman in the back seat
- − Only one paw is actually on the steering wheel
Stable Diffusion 3.5 Large Turbo
- + Includes the human passenger in the back seat as requested
- + Both front paws are positioned on the steering wheel
- + Realistic night lighting and bokeh effect in the background
- − Lower level of detail on the fur and facial textures compared to the other model
- − The capybara's head is slightly oddly shaped
Verdict: Stable Diffusion 3.5 Large Turbo provides a more complete interpretation of the prompt by including the passenger and the correct 'both paws on the wheel' instruction. While Stable Diffusion 3.5 Large has superior texture for the fur and clothing, it fails to include the primary secondary element (the woman in the back seat), making it less successful as a text-to-image response.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent atmospheric lighting with a moody, cinematic night sky and a large glowing moon.
- + Higher level of detail in the border illustrations, featuring intricate spiderwebs and cemetery elements.
- + Successfully included the requested scroll banner and captured most of the text.
- − Completely failed to include the specific event details (Date, Time, Location) at the bottom.
- − The text in the scroll banner contains minor garbled characters at the very bottom.
Stable Diffusion 3.5 Large Turbo
- + Clearer rendering of the central jack-o-lantern with sharp vector-like features.
- + Highly legible 'Halloween Party' title text at the top.
- − Failed almost all specific text instructions, missing the scroll banner, the subtitle, and the event details.
- − The composition feels more like a generic clipart frame rather than the 'dark parchment' vintage poster requested.
- − Lacks the 'cinematic lighting' and 'moody sky' requested, opting for a flat beige background instead.
Verdict: Stable Diffusion 3.5 Large is the clear winner for its superior ability to capture the gothic atmosphere and cinematic lighting requested. While both models failed to include the specific event details (Date, Time, Location), Stable Diffusion 3.5 Large managed to render the invitation title and the scroll banner with much better adherence to the prompt's aesthetic compared to the overly simplistic and flat output from Stable Diffusion 3.5 Large Turbo.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent texture rendering for the rice and fish, showing high visual clarity.
- + Accurate text rendering for both 'JAPAN' and 'SUSHI'.
- + Sophisticated lighting and material details that give a premium 3D miniature feel.
- − The camera angle is less isometric and more of a standard high-angle perspective.
- − Composition is slightly cluttered with many elements, straying from the 'minimal' request.
Stable Diffusion 3.5 Large Turbo
- + Strict adherence to the 45-degree isometric perspective requested in the prompt.
- + Clean, stylized 'cartoon' aesthetic that matches the miniature 3D theme.
- + Distinct diorama base that effectively separates the scene from the background.
- − Text contains a misspelling ('SIIHI' instead of 'SUSHI').
- − The flag icon is incorrect, appearing as a red and white rectangle rather than the Japanese flag.
Verdict: Stable Diffusion 3.5 Large wins due to its superior text accuracy and much higher rendering quality, particularly in the realistic PBR textures of the sushi. While Stable Diffusion 3.5 Large Turbo followed the isometric layout more strictly, it failed significantly on the text spelling and the specific identity of the Japanese flag.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Successfully included all four requested animals (dog, cat, rabbit, fox).
- + Beautifully captures the warm 'god rays' and bokeh effects requested in the prompt.
- + Superior textures on fur and environment, giving a high-quality 8K feel.
- − The fox kit in the background is slightly less detailed than the foreground animals.
Stable Diffusion 3.5 Large Turbo
- + Bright, vibrant colors and sharp focus on the primary subjects.
- + Clean, stylized rendering that feels very 'wholesome' and cute.
- − Failed to include the baby bunny and the red fox kit from the prompt.
- − The 'tabby kitten' looks more like a caracal or exotic wild cat hybrid.
- − Lacks the atmospheric depth and realistic lighting (god rays/dew) requested.
Verdict: Stable Diffusion 3.5 Large is the clear winner as it fulfilled all aspects of the complex prompt, including all four specific animal species. Stable Diffusion 3.5 Large Turbo only generated two animals and failed to capture the 'hyper-photorealistic' and atmospheric lighting requirements, resulting in a more plastic, simplified image.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Excellent adherence to the 'minimalist' and 'vector emblem' style instructions.
- + Clear and legible typography for 'Caffè Florian'.
- + Sophisticated use of subtle texture and a balanced color palette.
- − Includes an extra 'e' in 'Cafféé', though this is a common AI spelling hurdle.
Stable Diffusion 3.5 Large Turbo
- + Strong 'vintage' aesthetic with high-contrast shading.
- + Creative integration of a coffee mug shape beneath the cloche dome.
- − Text rendering is messy with garbled characters and overlapping lines in 'Florin'.
- − Less 'minimalist' than requested, appearing more like a complex illustration than a logo.
Verdict: Stable Diffusion 3.5 Large is the clear winner as it successfully interprets the 'minimalist' and 'vector' requirements of the prompt, creating a clean, professional logo. Stable Diffusion 3.5 Large Turbo produces a more cluttered illustration with significant text errors and artifacts that make the logo unusable.
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
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
Stable Diffusion 3.5 Large
- + Captures a more complex technical diagram feel.
- + Includes a variety of orbital and planetary imagery.
- − Failed to follow the Saturn V launch vehicle description, showing a space shuttle instead.
- − The layout is cluttered and lacks the requested clean vector infographic structure.
- − Text is completely illegible gibberish.
Stable Diffusion 3.5 Large Turbo
- + Successfully follows the modern vector infographic style with clean columns and icons.
- + Better adherence to the NASA-inspired color palette and flat design aesthetic.
- + Layout is logical and easy to read as a poster.
- − Missed several of the requested steps (Launch, Translunar, etc.) in the icon list.
- − Text rendering is poor with several misspellings like 'Apoll.o'.
- − Includes an astronaut figure which was not requested in the prompt.
Verdict: Stable Diffusion 3.5 Large Turbo is the clear winner for its superior adherence to the 'clean, modern vector infographic' style requested. While Stable Diffusion 3.5 Large produced a messy collection of space-themed assets with a Space Shuttle (incorrect for Apollo 11), the Turbo model successfully organized the information into a structured layout that feels like a real poster.
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