OpenAI's legacy image generation model supporting generations, edits with masks (inpainting), and variations
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
DALL-E 2
15.8
arena score
#59 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
DALL-E 2
100.0%
win rate
Ties
0.0%
Stable Diffusion 3.5 Large Turbo
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Features a small glass cube on a wooden surface
- + The color palette reasonably includes red, blue, and green
- − Fails significantly on spatial arrangement and logic
- − Objects are merged or confused, with the red book appearing inside the cube rather than on top
- − Low image resolution and extreme depth-of-field blurring the background
Stable Diffusion 3.5 Large Turbo
- + Excellent adherence to the complex spatial prompt
- + High visual clarity and realistic lighting with shadows on the wooden table
- + Accurately places the red book at the base while keeping the blue sphere clearly inside the glass structure
- − Reversed the instruction for the red book and the cube, placing the cube on top of the book
- − The plant is beside the cube rather than partially behind it
Verdict: Stable Diffusion 3.5 Large Turbo produces a much higher quality image with clear, distinct objects and realistic lighting, although it swapped the positions of the red book and the glass cube. DALL-E 2 struggled significantly with coherence, resulting in a blurry image where the objects are small, poorly defined, and fused together incorrectly.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Successfully captures the imperfect framing and shallow depth of field requested.
- + Conveys a strong atmospheric and cinematic mood with realistic wet pavement reflections.
- − The subject is heavily obscured and out of focus, failing to show the 'elderly Japanese man' or 'natural skin texture'.
- − The composition makes it difficult to distinguish the specific actions being performed.
Stable Diffusion 3.5 Large Turbo
- + Clearly depicts an elderly Japanese man repairing a red bicycle as requested.
- + Matches the specific subject requirements and shows visible rain and wet pavement.
- − Fails to include the 'motion blur from passing cars' as the background vehicles are static.
- − The image has a flat, digital appearance with poor 'shallow depth of field' and no cinematic quality.
Verdict: Stable Diffusion 3.5 Large Turbo followed the subject description much better by showing the man's face and the full red bicycle, whereas DALL-E 2 completely missed the subject details. However, DALL-E 2 captured the requested 'cinematic' and 'imperfect framing' mood more effectively than the sterile, static output of Stable Diffusion 3.5 Large Turbo.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Successfully captures a weathered, gritty texture
- + Effective use of shallow depth of field
- − Extremely low resolution and clarity
- − Anatomically incoherent face and armor structure
- − Fails to show specific details like braided hair or beads
Stable Diffusion 3.5 Large Turbo
- + Excellent adherence to technical details like braided hair and ornate engravings
- + Clear, high-resolution rendering with lifelike eyes
- + Good lighting and skin textures with requested scars/dirt
- − The blood/dirt on the face looks slightly like a decal or paint
- − Composition is a bit generic for a 'battle-worn' character
Verdict: Stable Diffusion 3.5 Large Turbo accurately follows all prompts, delivering a high-quality portrait with clear details on the armor and hair. In contrast, DALL-E 2 produces a blurry, distorted image where the features are unrecognizable and the specific details of the prompt are lost in noise.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Strong bold sans-serif typography that feels impactful
- + High-contrast visual style matching the minimalist label
- − The food photos are fragmented into messy, unrecognizable shapes
- − Fails to create distinct menu sections like 'apps' or 'pizza'
- − Text is completely illegible gibberish
Stable Diffusion 3.5 Large Turbo
- + Successfully creates a grid layout for food photography
- + Includes specific sections for 'Pizza' and 'Mians' (Mains) as requested
- + Clean, high-resolution food assets with a professional look
- − Food looks somewhat artificial and plastic-like
- − Typography contains minor spelling errors ('Mians')
- − Layout is more like a top-down collage than a singular menu design
Verdict: Stable Diffusion 3.5 Large Turbo significantly outperforms DALL-E 2 by following the structural prompts for a grid and specific menu sections. DALL-E 2 produced an abstract design with fragmented imagery that did not function as a legible menu, whereas Stable Diffusion 3.5 Large Turbo created a clear, professional layout despite some artificiality in the food renders.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
100% wins
0% ties
0% wins
AI Judge Analysis
DALL-E 2
- + Successfully captures the 'exploded' and 'suspended' effect for most components.
- + Creates a cohesive fiery atmosphere around the food.
- − Nonsensical text rendering with misspellings like 'MARGIC'.
- − Low image resolution and lack of photorealistic detail.
- − Failed to include the price starburst and secondary messaging.
Stable Diffusion 3.5 Large Turbo
- + High visual clarity and sharp, crisp rendering.
- + Excellent fire and smoke effects that match the requested background.
- − Failed to follow the 'exploded' instruction, showing a stacked burger instead.
- − Completely missed all text requirements including the title and price.
- − Image feels more illustrative/CG than photorealistic.
Verdict: Both models failed significantly on the specific text and layout requirements. DALL-E 2 attempted the 'exploded' view and some text, but the results are low-quality and riddled with spelling errors. Stable Diffusion 3.5 Large Turbo produced a much cleaner and more professional image, but it ignored the 'exploded' instruction and failed to include any text at all.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + The text has a convincing chalk-on-board texture.
- + It mimics organic handwriting better than the other model.
- − The text is completely illegible gibberish.
- − The composition is a messy close-up that ignores the 'cozy café' setting.
- − It failed to include the specific requested dates and prices.
Stable Diffusion 3.5 Large Turbo
- + Sets the scene effectively with a cozy café environment and lighting.
- + The text is much more legible and attempts almost all the specific words in the prompt.
- + Follows the requested layout and composition.
- − The 'handwriting' looks more like a digital brush font than authentic chalk.
- − Contains several spelling errors and incomplete dates (e.g., 'apri 3l 20-').
- − The three-column layout was not requested in the prompt.
Verdict: Stable Diffusion 3.5 Large Turbo is the clear winner as it successfully rendered a cozy café scene and legible text that closely follows the prompt items. DALL-E 2 failed significantly, producing garbled text that is entirely unreadable and lacks the requested context. While Stable Diffusion 3.5 has some spelling errors and the text looks a bit too 'clean' for real chalk, it is much closer to the user's requirements.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Successfully captures a cinematic sense of space scale
- + Artistic and gritty texture fits a surreal theme
- − Fails the specific prompt instruction for the horse to be on top of the astronaut
- − Low resolution and grainy details compared to modern standards
- − Anatomical issues with the horse's legs and tail
Stable Diffusion 3.5 Large Turbo
- + High resolution with clean, modern digital art style
- + Good composition with a clear focal point and interesting shadows
- − Fails the specific prompt instruction for the horse to be on top of the astronaut
- − Physical proportions of the astronaut relative to the horse are slightly awkward
- − The 'surreal' aspect is minimal, opting for a literal interpretation
Verdict: Both DALL-E 2 and Stable Diffusion 3.5 Large Turbo failed the negative constraint/spatial reasoning test of placing the horse on top of the astronaut. However, Stable Diffusion 3.5 Large Turbo is the superior image due to its significantly higher resolution, cleaner details, and better lighting, whereas DALL-E 2 produces a muddy and pixelated result.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Attempts to show both the driver and the passenger as requested.
- − Very low visual quality with severe anatomical distortions and artifacts.
- − The capybara does not look like a capybara and lacks the requested clothing.
- − The woman's face and hands are mangled, and she appears shocked rather than bored.
Stable Diffusion 3.5 Large Turbo
- + Excellent photorealistic texture and lighting.
- + High prompt adherence for the capybara's appearance, clothing, and posture.
- + Great composition that realistically captures the interior of a vehicle at night.
- − The passenger in the back seat is out of focus and not clearly interacting with a phone.
- − The steering wheel is positioned strangely relative to the capybara's body.
Verdict: Stable Diffusion 3.5 Large Turbo produces a high-quality, professional-looking image that captures the requested atmosphere and character details extremely well. DALL-E 2 fails significantly, producing an image filled with grotesque artifacts and poor anatomical structures. Stable Diffusion 3.5 Large Turbo is the clear winner for its superior realism and overall adherence to the prompt.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Strong vintage parchment aesthetic
- + The border feels organic and detailed
- + Moody lighting captures the requested cinematic atmosphere
- − Text is largely unintelligible and poorly rendered
- − None of the specific date or location details are visible
- − Composition is messy and lacks focus
Stable Diffusion 3.5 Large Turbo
- + Clear and legible main heading
- + Excellent inclusion of specific elements like webs, thorns, and jack-o-lanterns
- + Clean, high-resolution graphic design style
- − Missed the secondary scroll banner and event details at the bottom
- − The 'vintage' feel is replaced by a modern clean-cut vector look
- − The composition feels slightly empty in the center-bottom
Verdict: Stable Diffusion 3.5 Large Turbo wins because it adheres much more closely to the specific elements requested, including legible text and specific imagery like webs. DALL-E 2 captures a better 'vintage' vibe but fails completely on text rendering and clarity, resulting in an unusable invitation.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Successfully uses a solid light blue background
- + Captures an isometric 45-degree angle correctly
- − Fails to include the required text 'JAPAN'
- − Spells 'SUSHI' incorrectly as 'Sush'
- − Lacks the 3D diorama/miniature aesthetic requested
- − Poor visual quality with incoherent shapes representing the food
Stable Diffusion 3.5 Large Turbo
- + Excellent 3D miniature diorama aesthetic with high-quality PBR textures
- + Includes almost all requested text elements and a flag icon
- + Rich, detailed modeling of sushi rolls and rice textures
- − Spells 'SUSHI' incorrectly as 'SIIHI'
- − Background is a grayish-blue rather than a solid light blue
- − Flag icon is generic rather than explicitly the Japanese flag
Verdict: Stable Diffusion 3.5 Large Turbo significantly outperforms DALL-E 2 by providing a high-fidelity 3D miniature rendered style that matches the prompt's aesthetic requirements. While it has a minor typo in the secondary text, the overall composition, lighting, and detail are professional, whereas the DALL-E 2 image is sparse, incorrectly spelled, and lacks technical clarity.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Attempts to include all animals requested in the prompt
- + Better sense of spatial depth in the meadow
- − Severely distorted anatomy on the kitten and butterflies
- − Poor image quality with significant artifacts and smearing
- − Muddled textures on the fur and grass
Stable Diffusion 3.5 Large Turbo
- + Excellent visual clarity and sharp focus
- + Beautiful backlighting with high-quality fur textures
- + Clean, artistic composition with expressive features
- − Missing the bunny and the red fox kit from the prompt
- − The fox-like animal looks more like a stylized kitten than a fox kit
- − Leans toward an illustrative/3D render style rather than photorealistic
Verdict: Stable Diffusion 3.5 Large Turbo produces a significantly higher quality image with superior lighting and textures, though it fails to include all the animals requested in the complex prompt. DALL-E 2 attempts the full scene but fails on a technical level, resulting in heavily distorted creatures and poor resolution. Stable Diffusion 3.5 Large Turbo is the clear winner for its aesthetic polish and coherence.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Simple vector style
- + Follows the brown and cream color palette
- − Text is complete gibberish and unreadable
- − Lacks the requested banner and steam elements
- − Poor visual coherence with messy, fragmented characters
Stable Diffusion 3.5 Large Turbo
- + Excellent typography rendering with almost perfect spelling
- + Incorporates all prompt elements including the banner and steam
- + Strong vintage aesthetic with high-quality shading and texture
- − Minor typo in 'Caffé' (added an extra 'e')
- − Includes a handle that makes it look like a hybrid between a cloche and a coffee pot
Verdict: Stable Diffusion 3.5 Large Turbo vastly outperforms DALL-E 2 by providing readable text and adhering to all complex prompt elements like the banner and steam. While DALL-E 2 produced an abstract and unreadable mess, Stable Diffusion 3.5 created a polished, professional logo that captures the vintage minimalist aesthetic perfectly.
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.”
DALL-E 2
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
DALL-E 2
- + Adheres better to the requested high-contrast navy/white palette.
- + Includes a complex network of lines that suggests a data-driven infographic.
- + Uses a very clean, flat-vector style for the primary font.
- − Text is nonsensical and includes large, distracting typos like 'ALLPOO'.
- − The layout is chaotic and fails to clearly visualize the sequential 6-step mission timeline.
- − Icons are abstract and unrecognizable as the Saturn V or Lunar Module.
Stable Diffusion 3.5 Large Turbo
- + Clearly attempts a sequential layout with better typographic hierarchy.
- + The illustration style is much more atmospheric and aesthetically pleasing as a poster.
- + Follows the NASA-inspired color palette more effectively with muted tones.
- − Fails to include all 6 specific steps requested in the prompt, collapsing them into fewer items.
- − Typography contains several misspellings like 'Laurch' and 'Traneluna'.
- − Uses more of an illustrative style rather than the requested flat-vector infographic style.
Verdict: Stable Diffusion 3.5 Large Turbo is the clear winner as it produces a coherent poster layout that actually resembles the intended subject matter. While both models struggle with spelling, Stable Diffusion 3.5 Large Turbo provides recognizable illustrations and a sequential structure, whereas DALL-E 2 generates a cluttered, abstract design with an unfortunate and distracting typo in the main header.
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