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
Wan 2.6
23.3
arena score
#28 of 62 in Text-to-Image
Top 2 in Image-to-Video
Vote tally
Where the votes landed
Stable Diffusion 3.5 Large Turbo
0%
win rate
Ties
0%
Wan 2.6
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Crisp and clean modern aesthetic with high resolution.
- + Correct lighting direction from the left as requested.
- − Failed the spatial reasoning for the book, placing the red book inside the cube instead of on top.
- − The plant is to the side rather than strictly behind the cube.
- − The blue sphere is sitting on the book rather than inside the glass cube directly.
Wan 2.6
- + Perfect adherence to spatial instructions with the red book on top and blue sphere inside.
- + Highly realistic textures, especially on the weathered book and wooden table.
- + Successfully placed the plant behind the glass cube with realistic refraction/transparency.
- − The plant is slightly more to the right than 'directly' behind, though still partially visible through the glass.
Verdict: Wan 2.6 followed every complex spatial instruction perfectly, placing the red book on top and the sphere inside the cube. Stable Diffusion 3.5 Large Turbo failed the logic of the prompt, placing the red book inside the cube and the sphere on top of the book, although it maintained higher clarity and modern sharpness.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Simple, clean silhouette of the bicycle
- + Follows the red color requirement perfectly
- − Anatomy is severely broken with the head merged into the shoulder
- − Lacks the requested rain atmosphere and motion blur from cars
- − The bicycle geometry is illogical with missing components
Wan 2.6
- + Excellent adherence to atmosphere, including rain droplets and pavement reflections
- + Highly realistic skin textures and believable elderly Japanese features
- + Captures the technical camera requirements like shallow depth of field and motion blur perfectly
- − The rear wheel of the bicycle is slightly distorted where it meets the frame
- − The hands and tools are a bit muddy in detail upon close inspection
Verdict: Stable Diffusion 3.5 Large Turbo failed significantly on human anatomy and the 'realistic' requirement, producing a strangely merged character with no environmental effects. Wan 2.6 followed every detail of the prompt, delivering a cinematic, atmospheric, and highly realistic street photo that perfectly captured the lighting, rain, and cultural context.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Strong decorative detail on the plate armor
- + Clear, vibrant colors with high contrast
- − Texture looks excessively smooth and 'plastic' like CGI
- − Failed to include beads in the hair braids
- − Blood/dirt effects look painted on rather than integrated
Wan 2.6
- + Excellent adherence to the 'beads in braids' and 'bokeh sparks' prompt
- + Highly realistic skin texture including believable dirt and scarring
- + Authentic material rendering on worn leather straps and frayed cloth underlayers
- − The lighting on the face is a bit flat compared to the bright torches in the background
Verdict: Wan 2.6 is the clear winner as it masterfully interpreted every detail of the prompt, specifically the beads in the hair and the texture of the cloth and leather. Stable Diffusion 3.5 Large Turbo produced a much more generic, 'airbrushed' fantasy look that missed several key descriptors and lacked the gritty realism 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.”
Stable Diffusion 3.5 Large Turbo
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Features a very clean white background
- + Food photography is vibrant and high-quality
- + Organized layout that resembles a branding flat-lay
- − Includes a significant spelling error in 'Mians'
- − The layout is more of a collage than a cohesive, readable menu page
- − Food items look slightly artificial and plastic-like
Wan 2.6
- + Excellent adherence to the 'grid' request for photos
- + Clean and readable typography following professional menu standards
- + Functional layout with clear price points and section headers
- − Includes some minor gibberish text in subheaders
- − Small artifact in the photo grid where a photo is labeled 'Pizza' but doesn't align with the text category below
- − The top 'Restaur Menu' text has slight spelling artifacts
Verdict: Wan 2.6 provides a much more realistic and functional menu design that correctly organizes sections for appetizers, pizza, and mains in a usable format. Stable Diffusion 3.5 Large Turbo produces a visually striking flat-lay composition, but it fails as a 'menu design' due to poor information hierarchy and a major typo in a primary heading.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent depiction of the fiery background and glowing embers.
- + The burger image is vibrant and appetizing with good ingredient detail.
- − Completely failed to generate any of the requested text.
- − The burger is mostly stacked rather than 'exploded' or suspended in mid-air as requested.
Wan 2.6
- + Perfect adherence to all text requirements including 'MAGIC BURGER', 'LIMITED TIME ONLY', and the price in a starburst.
- + Accurately represents the 'exploded' view with components floating dynamically.
- + High level of photorealistic detail on the burger patty and vegetable textures.
- − The 'LIMITED TIME ONLY' text is slightly warped along the bottom edge.
- − The price starburst has a flatter, more 2D graphic feel compared to the realistic background.
Verdict: Wan 2.6 is the clear winner as it followed every instruction in the prompt, including the complex text integration and the 'exploded' burger layout. Stable Diffusion 3.5 Large Turbo produced a visually pleasing image but failed entirely to include any text or the exploded suspension effect.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean aesthetic and well-lit background composition.
- − Numerous spelling errors including 'risctto', 'apri', and 'ocopas'.
- − Text looks like a digital font rather than natural chalk handwriting.
- − Fails to follow the specific menu item list correctly.
Wan 2.6
- + Perfect adherence to text content with zero spelling errors.
- + Extremely realistic chalk texture with dusty smudges and authentic handwriting variations.
- + Matches the atmospheric 'cozy café' request with depth of field.
- − The perspective is slightly angled rather than straight-on, though it fits the café vibe.
Verdict: Wan 2.6 is the clear winner as it followed every instruction, including the specific date and menu items with perfect spelling and realistic chalk textures. Stable Diffusion 3.5 Large Turbo failed significantly on text rendering, producing gibberish words and a clean digital look that ignored the 'handwritten-style' and 'chalk texture' requirements.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean aesthetic with sharp, high-contrast lighting.
- + High resolution with smooth textures and clear anatomical lines.
- − Failed the negative constraint: the astronaut is on top, not the horse.
- − The composition is somewhat generic, resembling a common stock image concept.
Wan 2.6
- + Beautiful cinematic lighting and intricate details in the space background.
- + Excellent textures on the horse's coat and the astronaut's suit.
- − Failed the negative constraint: the astronaut is on top, not the horse.
- − The horse's front legs have slightly awkward positioning/extra joints upon close inspection.
Verdict: Both models failed the specific spatial instruction to place the horse on top of the astronaut, instead defaulting to the common 'astronaut riding a horse' trope. Stable Diffusion 3.5 Large Turbo offers a cleaner, more graphic look, while Wan 2.6 provides a much more detailed and visually stunning cinematic environment, making it the preferred choice despite the shared prompt adherence failure.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + High resolution with very sharp textures on the capybara's fur and jacket.
- + Clean cinematic lighting and bokeh in the background.
- − Missed the core prompt requirement of the passenger looking at her phone.
- − Passenger is extremely blurry and lacks detail.
- − The capybara's paws are merged awkwardly with the steering wheel.
Wan 2.6
- + Excellent prompt adherence including the businesswoman looking at her phone.
- + Complex rain and interior lighting effects add to the realism of a New York night.
- + Both subjects are in clear focus and correctly positioned according to the description.
- − The passenger's hand holding the phone has some anatomical distortion.
- − The taxi's roof light 'BOY' text is nonsensical.
Verdict: Wan 2.6 is the clear winner as it followed every instruction in the prompt, particularly the specific interaction of the passenger looking at her phone. Stable Diffusion 3.5 Large Turbo produced a high-quality close-up of the capybara but failed to include the primary action of the human character.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Clean graphic design aesthetic
- + Clear and legible title text
- + Good use of space for a border concept
- − Failed to include most of the requested text including event details
- − The lighting is flat rather than cinematic
- − Missing the parchment texture and scroll element
Wan 2.6
- + Full adherence to all requested text including specific event details
- + Highly detailed gothic aesthetic with cinematic lighting and textures
- + Excellent execution of the scroll banner and twisted trees
- − The thorn border is a bit busy and overlaps some texture elements
- − Slightly less 'clean' than a digital vector-style invitation
Verdict: Wan 2.6 followed the prompt instructions near perfectly, including all requested text, the scroll banner, and the specific event details which Stable Diffusion 3.5 Large Turbo omitted entirely. Wan 2.6 also captured the 'vintage gothic' mood much more effectively with its rich textures and moody lighting versus the more generic vector look of Stable Diffusion.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent 3D toy-like texture on the sushi rice and fish components.
- + Strong isometric perspective and clean rendering.
- − Text is misspelled as 'SIIHI'.
- − The flag icon is generic and does not represent the Japanese flag requested.
- − The 'JAPAN' text is on a physical board rather than being part of the graphic design overlay.
Wan 2.6
- + Followed all text instructions perfectly, including 'JAPAN', 'SUSHI', and a correct Japanese flag icon.
- + Achieved much better framing for a diorama with a clean graphic design feel.
- + Correctly placed text at top-center as requested.
- − The sushi details are slightly less crisp than Model A.
- − The background is a bit more gradient-like than a 'solid' color.
Verdict: Wan 2.6 is the clear winner as it followed every complex instruction, including multiple strings of text and a specific flag icon, with perfect spelling and placement. Stable Diffusion 3.5 Large Turbo failed on the spelling of 'SUSHI' and generated a generic red/white flag instead 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 Turbo
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Strong backlighting and vibrant colors
- + Clean character rendering with expressive eyes
- − Failed to include the rabbit and fox kit accurately
- − Look is more like an illustrative 3D animation than hyper-photorealistic
- − Anatomy of the smaller animals is merged and confusing
Wan 2.6
- + Includes all four requested animals correctly: puppy, kitten, bunny, and fox
- + Achieves a high level of photorealism with natural fur textures and lighting
- + Captures the 'tumbling' and 'playful' action requested in the prompt
- − Sun rays are a bit digitally artificial in their convergence
- − Small amount of floating debris or distorted petals in the air
Verdict: Wan 2.6 is the clear winner as it successfully rendered all four distinct animals (golden retriever, tabby kitten, bunny, and fox) whereas Stable Diffusion 3.5 Large Turbo omitted the bunny and fox, replacing them with generic cat/dog hybrids. Wan 2.6 also adhered much better to the request for photorealism and a dynamic scene, while Stable Diffusion 3.5 Large Turbo produced a static, stylized 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 Turbo
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent banner design with high-quality shading
- + Good adherence to the requested warm brown and cream color palette
- + Includes all requested elements like steam and the 'Est. 1720' text
- − Confused iconography blending a cloche with a coffee mug
- − Minor spelling error in the name ('Caffeé Florin')
- − Large shadows on the background reduce the 'minimalist' vector feel
Wan 2.6
- + Perfect text rendering for the main name 'Caffè Florian'
- + Clean, minimalist vector aesthetic that is very professional
- + Accurate cloche dome representation
- − The 'Est. 1720' banner is small and integrated awkwardly into the side of the dome
- − Slightly less 'classic' typography on the banner compared to the main text
Verdict: Wan 2.6 is the winner as it perfectly captures the minimalist vector style and correctly spells the name 'Caffè Florian' with the proper accent. While Stable Diffusion 3.5 Large Turbo has a more elaborate banner, it fails on spelling and creates a confusing hybrid object that is half-cloche and half-mug.
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
Wan 2.6
AI Judge Analysis
Stable Diffusion 3.5 Large Turbo
- + Excellent adherence to the infographic layout requested with multiple steps.
- + Strong visual style that captures the NASA-inspired color palette and vector aesthetic.
- + Includes complex illustrative elements like planetary bodies and orbiting trajectories.
- − Text rendering is mostly gibberish or misspelled ('Apoll.o', 'Laurch').
- − The astronaut appears to be standing on a ladder in a strange, non-physical way.
Wan 2.6
- + Perfect text rendering for the names and title.
- + Clean, minimalist composition that uses the requested color palette well.
- − Completely failed to provide the requested 6-step infographic content.
- − The background has a textured, fabric-like appearance instead of a crisp vector style.
- − Lacks all requested icons (Saturn V, orbit rings, lunar module).
Verdict: Stable Diffusion 3.5 Large Turbo followed the complex structural requirements of the prompt, creating a true multi-step infographic even though the text within it is garbled. Wan 2.6 rendered perfectly legible text but completely missed the requested content, providing only a title and names instead of the 6-step mission progression. Stable Diffusion is much more useful as a starting point for a design task like this.
Next steps
Explore each model
Alibaba's multimodal generation model from the Wan AI suite, supporting text-to-video, image-to-video, reference-to-video with audio, and text-to-image, in both Chinese and English