Imagen 3.0 Generate 002 Google Stable Diffusion 3.5 Large Turbo Stability AI

Settled by community votes across 6 shared challenges, with an AI judge weighing in on each.

Imagen 3.0 Generate 002

21.0 arena score

#36 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

Imagen 3.0 Generate 002

win rate

Ties

Stable Diffusion 3.5 Large Turbo

win rate

Shared challenges 6

Challenge by challenge

The strongest take from each model on every shared challenge, with the AI judge's read.

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.”

Imagen 3.0 Generate 002

Stable Diffusion 3.5 Large Turbo

AI Judge Analysis

Imagen 3.0 Generate 002

+ Excellent adherence to the 'grid' layout mentioned in the prompt
+ Clean, professional typography that actually looks like a menu design
+ Realistic and high-quality food photography

− Text contains gibberish/nonsense characters despite looking like words
− Repetitive use of pizza images in almost every photo slot

Stable Diffusion 3.5 Large Turbo

+ Vibrant color palette and creative food presentation
+ Readable section headings like 'Pizza' and 'Mians'

− Layout is disparate and looks more like a collage than a professional menu
− Food items have a rubbery, artificial, or 'plastic' appearance
− Fails to create a cohesive single-page design with a white background

Verdict: Imagen 3.0 successfully captured the requested aesthetic of a modern, minimalist menu with a professional grid layout, whereas Stable Diffusion 3.5 Large Turbo produced a disorganized collection of assets. While Imagen 3.0 has issues with text legibility, its composition and realistic food visuals make it much more useful for a design challenge.

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.”

Imagen 3.0 Generate 002

Stable Diffusion 3.5 Large Turbo

AI Judge Analysis

Imagen 3.0 Generate 002

+ Excellent adherence to the exploded view requested in the prompt.
+ Renders all requested text accurately, including the specific price in a starburst.
+ High level of photorealism and professional commercial lighting.

− Includes some nonsensical gibberish text inside the starburst above the price.

Stable Diffusion 3.5 Large Turbo

+ Dynamic lighting and vibrant fire effects create a strong sense of mood.
+ Good rendering of food textures like the toasted bun and melting cheese.

− Completely ignores all requested text elements (MAGIC BURGER, LIMITED TIME, Price).
− Fails to provide the 'exploded' view, showing a mostly assembled burger instead.
− Includes a random stick or skewer poking out of the top bun.

Verdict: Imagen 3.0 Generate 002 is the clear winner as it followed all complex prompt instructions, including the specific exploded layout and all multi-part text requirements. Stable Diffusion 3.5 Large Turbo failed to include any of the text and did not deliver an exploded view, resulting in a standard burger image that does not meet the ad campaign criteria.

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.”

Imagen 3.0 Generate 002

Stable Diffusion 3.5 Large Turbo

AI Judge Analysis

Imagen 3.0 Generate 002

+ Excellent adherence to the specific text and menu items requested.
+ Very realistic chalk texture and legitimate handwritten appearance as per the prompt.
+ Accurately rendered date and pricing symbols.

− Contains minor spelling repetitions like 'Risottto' and 'Octpsuin'.
− Text at the bottom becomes slightly redundant ('Ask about daily, - gluten free options').

Stable Diffusion 3.5 Large Turbo

+ Successfully captures a 'cozy café' atmosphere with lighting and decor.
+ Follows the general layout of a chalkboard menu.

− Severe spelling errors throughout almost every word.
− Failed to render the specific date requested (April 31 instead of April 30).
− The font looks digital and smoothed out rather than like authentic hand-drawn chalk.

Verdict: Imagen 3.0 provides a significantly superior response by actually rendering the specific text and pricing requested in the prompt with high accuracy. While Stable Diffusion 3.5 Large Turbo creates a nice environmental scene, it fails on almost every text-based instruction, displaying garbled characters and the wrong date.

The Reversed Rodeo

Text-to-Image

“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”

Imagen 3.0 Generate 002

Stable Diffusion 3.5 Large Turbo

AI Judge Analysis

Imagen 3.0 Generate 002

+ Excellent cinematic lighting and atmospheric nebula details.
+ Anatomically correct horse with high-quality texture on hair and space suit.
+ Great use of depth and stellar lighting to create a surreal yet cohesive image.

− Followed the standard 'Astronaut on a horse' interpretation, ignoring the 'horse on top' spatial instruction.

Stable Diffusion 3.5 Large Turbo

+ Strong contrast and clean, graphic visual style.
+ Dynamic composition with the inclusion of a planet and orbital curve.
+ Good rendering of the horse's mane and tail physics.

− Failed the spatial reasoning prompt 'horse on top, not vice versa' just as Model A did.
− Noticeable anatomy issues with the horse's front legs and hooves.

Verdict: Both models failed the negative constraint and spatial reasoning challenge, providing a standard 'astronaut riding a horse' instead of the requested 'horse on top' inversion. However, Imagen 3.0 is the clear winner due to its superior artistic rendering, highly detailed textures, and more realistic anatomy compared to the distorted legs in the Stable Diffusion 3.5 output.

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.”

Imagen 3.0 Generate 002

Stable Diffusion 3.5 Large Turbo

AI Judge Analysis

Imagen 3.0 Generate 002

+ Excellent text rendering with no spelling errors
+ Accurate 45-degree isometric perspective
+ Clean, refined 3D cartoon textures that perfectly match the prompt

− The text is top-left rather than top-center as requested

Stable Diffusion 3.5 Large Turbo

+ Creative use of 3D signage within the diorama
+ Good lighting and material shaders on the rice and fish

− Significant spelling error with 'SIIHI' instead of 'SUSHI'
− The flag is generic and does not represent Japan
− Composition is a bit cluttered compared to the 'ultra-clean' request

Verdict: Imagen 3.0 Generate 002 follows the prompt much more effectively, delivering remarkably clean 3D assets and perfect typography. In contrast, Stable Diffusion 3.5 Large Turbo fails on the text rendering (spelling 'SIIHI') and includes a non-Japanese flag, which misses a key thematic element of 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.”

Imagen 3.0 Generate 002

Stable Diffusion 3.5 Large Turbo

AI Judge Analysis

Imagen 3.0 Generate 002

+ Follows all prompt requirements including all four specific animal species.
+ Excellent photorealistic textures on the fur, grass, and wildflowers.
+ Natural lighting with soft god rays and beautiful dew-like bokeh that enhances the atmosphere.

− The bunny's anatomy on its back is slightly awkward, though it fits the tumbling theme.
− The fox kit's ears are a bit small compared to a real fox.

Stable Diffusion 3.5 Large Turbo

+ Very expressive, large eyes that convey a cute and wholesome vibe.
+ Strong backlighting creates a nice halo effect on the fur.
+ High contrast and vibrant colors make the image pop.

− Failed to include all requested animals, missing the bunny and the distinct red fox kit.
− The fur texture looks airbrushed and 'plasticky' rather than hyper-photorealistic.
− Anatomy issues occur where the kitten and dog blend together in a confusing way.

Verdict: Imagen 3.0 Generate 002 is the clear winner as it successfully included all four requested animal species with a high degree of photorealism and natural lighting. Stable Diffusion 3.5 Large Turbo failed to follow the prompt's subject list, only showing three ambiguous animals, and resulted in a more stylized, digital art appearance rather than the requested hyper-photorealistic masterpiece.