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
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
Stable Diffusion 3.5 Medium
16.9
arena score
#57 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Vidu Q2
19.8
arena score
#42 of 62 in Text-to-Image
Vote tally
Where the votes landed
Stable Diffusion 3.5 Medium
100.0%
win rate
Ties
0.0%
Vidu Q2
0.0%
win rate
100.0%
0.0% ties
0.0%
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Strong photographic rendering of the glass material and blue sphere.
- + Maintains a clean, minimalist aesthetic.
- − Failed the spatial logical request by placing the red object inside the cube rather than on top.
- − The plant is in front of/above the cube rather than behind it as requested.
Vidu Q2
- + Perfect adherence to all spatial instructions, including the book on top and plant behind.
- + Excellent rendering of light and shadows, including the reflection of the sphere on the bottom of the glass.
- + High detail in the wood grain and book texture.
- − None notable; the image follows all aspects of the complex spatial prompt.
Verdict: Vidu Q2 followed every detail of the prompt, correctly positioning the red book on top of the cube and the plant behind it, while Stable Diffusion 3.5 Medium failed the spatial logic by placing the red object inside the cube. Vidu Q2 also demonstrated superior rendering of lighting, reflections, and textures.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Captures a great atmospheric mood with cinematic color grading
- + Effectively shows the rainy environment and street reflections
- + Includes the requested cars in the background for context
- − The man's anatomy and posture are slightly warped and unrealistic
- − The bicycle geometry is nonsensical, particularly the handlebars and frame connection
- − Lacks the specific 'repairing' action requested, looking more like he is just holding it
Vidu Q2
- + Excellent skin texture and realistic hand details
- + Successfully captures the specific action of repairing a bicycle chain
- + Good close-up composition that emphasizes the 'imperfect framing' prompt
- − The background car is perfectly sharp despite the request for motion blur
- − The bicycle has structural issues, including a double front fork/wheel assembly
- − The lighting feels a bit more studio-like than 'natural cinematic' light rain
Verdict: Stable Diffusion 3.5 Medium captures a much better street atmosphere and mood, but the central subject's anatomy and bicycle are poorly formed. Vidu Q2 excels at fine details like skin texture and the specific action of repairing the bike, though it fails to include the requested motion blur on the car and suffers from its own mechanical glitches with the bicycle structure. Vidu Q2 is the preferred choice for its superior rendering of the human subject and more convincing 'candid' feel.
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 Medium
Vidu Q2
100% wins
0% ties
0% wins
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Excellent facial detail and skin texture including pores and subtle scars
- + Strong adherence to the close-up portrait framing
- + Intricate engraving details on the armor plates
- − Missed the request for beads in the hair braids
- − The bokeh sparks appear a bit flat and superimposed
Vidu Q2
- + Successfully included beads in the hair braids
- + Better depiction of torchlight reflections and atmospheric orange glow
- + Very sharp and detailed texture on the leather straps and underlayer
- − Skin texture appears slightly more airbrushed/smooth compared to model A
- − The framing is more of a medium shot rather than a close-up portrait
Verdict: Stable Diffusion 3.5 Medium produces a more intimate and realistic facial portrait with superior skin texture, while Vidu Q2 captures more of the specific prompt details like the beads in the braids and the lighting atmosphere. Vidu Q2 is the preferred choice for its superior rendering of materials like leather and metal under torchlight, despite being a wider shot than 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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Features a clear grid layout of food photos as requested
- + Includes specific sections for pizza and mains
- + The text alignment and dot leaders for prices are very professional
- − The font is slightly stylized and blocky rather than a clean modern sans-serif
- − Text rendering is nonsensical gibberish
- − The layout feels slightly cramped with too many images
Vidu Q2
- + Higher visual quality with vibrant, appealing food photography
- + Cleaner use of white space and modern sans-serif fonts
- + Much better adherence to the 'modern minimalist' aesthetic
- − Does not follow the grid layout for photos as strictly as requested
- − Text is still mostly unreadable, though it looks more like real words at a glance
Verdict: Stable Diffusion 3.5 Medium followed the structural grid prompt more closely, but the overall design is cluttered. Vidu Q2 produced a significantly more professional and realistic 'modern minimalist' design with better typography and more vibrant colors, making it the preferred choice for a casual dining asset.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + The text is sharp and legible
- + The burger texture is very photorealistic and high-detail
- − Failed to provide an 'exploded' view; the burger is mostly assembled
- − The background fire looks somewhat flat and lacks depth
Vidu Q2
- + Successfully captured the 'exploded' view with suspended components
- + Text perfectly adheres to the 'fiery, glowing' requirement and brand names
- + Composition feels more dynamic and energetic
- − The currency symbol in the starburst is malformed
- − Some sauce highlights look slightly plastic compared to the bread
Verdict: Vidu Q2 is the clear winner for its superior prompt adherence regarding the 'exploded' layout and fiery text effects. Stable Diffusion 3.5 Medium failed the primary structural requirement by presenting a mostly contiguous burger despite the request for mid-air suspension of all parts.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + The chalk texture on the 'TODAY SPECIALS' text is highly realistic.
- + The wooden frame composition is clean and centered.
- − Failed significantly on text legibility for the specific menu items.
- − The date 'APRIL 30, 2026' is garbled and incomplete.
- − Contains numerous artifacts in the lower left text section.
Vidu Q2
- + Excellent adherence to the requested text content, including the specific date and prices.
- + The font style captures natural handwriting variations and realistic chalk smudging.
- + Successfully interpreted the 'Brown But...' prompt to mean 'Brown Butter Choco Chip Cookies'.
- − Minor spelling errors in longer words like 'Musshoom' and 'Octopd'.
- − The prices ($34) don't perfectly match the prompt ($24) for the first item.
Verdict: Vidu Q2 is the clear winner as it successfully rendered the specific date and complex menu items requested, despite minor spelling hiccups. Stable Diffusion 3.5 Medium failed to follow the text requirements, producing mostly illegible gibberish for the specific dishes listed in the prompt.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Clean photographic style with realistic lighting from the planet below.
- + Correctly depicts a white horse and astronaut in space.
- + High resolution with sharp details on the astronaut's suit.
- − Failed the spatial reasoning prompt 'horse on top' of the astronaut.
- − Anatomy error with an extra set of legs appearing behind the horse.
Vidu Q2
- + Beautiful surrealist interpretation with a nebula-effect horse hide.
- + Dynamic composition with a more cinematic feel.
- + Excellent detail in the background galaxies and stars.
- − Completely ignored the counter-intuitive instruction to place the horse on top of the astronaut.
- − Some messy artifacts where the reigns meet the horse's neck.
Verdict: Both models failed the specific logic test of placing the 'horse on top' of the astronaut, instead defaulting to the standard horse-and-rider configuration. Stable Diffusion 3.5 Medium offers a more traditional photographic look, but Vidu Q2 is preferred for its superior creativity and adherence to the 'surreal' and 'cinematic' style descriptors.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Excellent fur texture and lighting details on the capybara
- + Cinematic depth of field with realistic bokeh
- − The woman in the back is not looking at a phone
- − The capybara's paws are positioned awkwardly and not clearly gripping the wheel
Vidu Q2
- + Perfect adherence to all prompt instructions including the phone and bored expression
- + Clear, professional composition showing the full interior and city character
- + Accurate depiction of the capybara's paws on the steering wheel
- − The capybara's hands look slightly more humanoid/primate-like than natural capybara paws
- − Slightly less 'photorealistic' texture compared to Model A
Verdict: Vidu Q2 is the clear winner as it followed every detail of the prompt, specifically the businesswoman being bored and looking at her phone. While Stable Diffusion 3.5 Medium has slightly more realistic lighting and fur textures, it failed to include the phone and the businessman's interaction with it, making the scene less cohesive.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Strong composition with a decorative outer border
- + Includes all requested elements like trees and jack-o-lanterns
- − Numerous spelling errors including 'Halloweeen Inviloween'
- − Failed to include the specific 'scroll banner' requested
- − The main text layout is cluttered and poorly aligned
Vidu Q2
- + Successfully included the scroll banner for the sub-text
- + Includes the thorns and webs in a more integrated, 3D-looking border
- + The 'central jack-o-lantern' placement aligns better with the prompt
- − Several spelling errors in the main title and sub-text
- − Incorrect date and missing time details
- − Background trees are less distinct compared to the other model
Verdict: Both models struggled significantly with spelling the requested text details. However, Vidu Q2 followed the layout instructions more closely by including the requested central jack-o-lantern and the scroll banner, whereas Stable Diffusion 3.5 Medium misplaced the pumpkins and failed the banner requirement. Vidu Q2 also achieved a more professional 'cinematic' lighting effect on the parchment.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Excellent photorealistic textures on the roe and nori wraps.
- + Clean, modern typography that integrates well with the background.
- − Failed to include the requested flag icon.
- − Missed the 'raised diorama base' instruction, opting for a simple flat plate instead.
- − The text 'SUSHI' has a slight transparency artifact on the 3D-effect 'S'.
Vidu Q2
- + Perfectly followed all prompt instructions including the diorama base and flag icon.
- + Captured the 'miniature 3D cartoon' aesthetic much more accurately than Model A.
- + Highly legible and well-ordered bold text.
- − The textures look more like plastic/clay than realistic PBR materials.
- − The flag icon is a bit simple compared to the 3D styling of the sushi.
Verdict: Vidu Q2 is the clear winner as it adhered to every specific detail of the prompt, including the 'raised diorama base' and the flag icon which Stable Diffusion 3.5 Medium missed. While Stable Diffusion 3.5 Medium produced more realistic food textures, Vidu Q2 much better captured the requested isometric cartoon scene composition.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Strong foreground detail on the animal faces and eyes.
- + Vibrant color palette with clear, saturated lighting.
- + Large, expressive features that enhance the 'cute' aesthetic.
- − Failed to include a bunny/rabbit in the image.
- − Missing the tabby kitten (image shows a generic orange kitten instead).
- − The 'chasing' and 'tumbling' action is missing; animals are mostly static.
Vidu Q2
- + Excellent adherence to all requested animals including the bunny and tabby kitten.
- + Dynamic composition with more realistic movement/tumbling in the meadow.
- + Superior lighting effects with visible god rays and dew sparkles as requested.
- − Anatomy on the leaping fox is slightly distorted.
- − Includes an extra puppy not specifically requested in the prompt.
- − Some butterfly shapes appear slightly repetitive.
Verdict: Vidu Q2 is the clear winner as it successfully incorporated all four requested animal types and followed the 'tumbling/chasing' action instructions, whereas Stable Diffusion 3.5 Medium failed to include the bunny and replaced the tabby kitten with another orange cat. Vidu Q2 also captured the atmosphere better with visible god rays and a more sprawling, lifelike meadow scene.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Refined vintage woodcut aesthetic that feels authentic to the 'vintage' prompt.
- + Strong, clear typography for the primary text.
- + Consistent color palette with nice texture.
- − Misspelled 'Florian' as 'Florrian'.
- − Truncated the date to 'Est 170' instead of '1720'.
- − The cloche is stylized into a circle, losing some of the traditional dome shape.
Vidu Q2
- + Accurately depicts the steam and the cloche dome shape requested.
- + Cleaner vector-style lines consistent with modern logo design.
- − Significant text hallucinations and spelling errors like 'Farmiin' and 'Fopli20'.
- − Repetitive text elements that clutter the composition.
- − Lacks the 'vintage minimalist' feel, appearing more like a generic clip-art style.
Verdict: Stable Diffusion 3.5 Medium is the preferred choice as it captures the 'vintage' and 'emblem' aesthetic far more effectively than Vidu Q2, despite failing on the specific year. While both models struggled significantly with spelling, Stable Diffusion 3.5 Medium produced a cohesive, high-quality illustration, whereas Vidu Q2 suffered from incoherent text repetitions and a less sophisticated design.
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 Medium
Vidu Q2
AI Judge Analysis
Stable Diffusion 3.5 Medium
- + Strong composition with a dynamic planetary arc.
- + Adheres well to the navy, white, and muted red color palette.
- + Includes a variety of icons corresponding to mission steps.
- − Text is highly garbled and mostly illegible.
- − Visual sequence of icons is logical but messy in execution.
- − Iconographic style varies significantly between elements.
Vidu Q2
- + Excellent flat-vector style with consistent iconography.
- + Layout is clear and easy to read as an infographic.
- + Text is cleaner and closer to recognizable English words despite errors.
- − Failed the specific step instructions (skipped launch etc. in favor of general steps).
- − Color palette is overly light/gray, lacking the requested navy contrast.
- − Logic of the steps (e.g., Earth in step 1, Moon in step 2) is slightly simplified compared to the prompt.
Verdict: Stable Diffusion 3.5 Medium created a more visually interesting and color-accurate composition, but the text and icons were messy. Vidu Q2 produced a much cleaner, professional-looking vector graphic that fits the 'modern infographic' aesthetic perfectly, despite some deviations from the specific mission steps requested. Vidu Q2 is preferred for its superior clarity and consistency.
Next steps
Explore each model
ShengShu Technology's text-to-image and reference-to-image model with support for character consistency and multi-reference image processing