Improved version of Alibaba's Qwen image model with better text rendering, finer natural textures, and more realistic human generation.
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
Qwen Image 2512
22.9
arena score
#30 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Seedream 4.0
24.7
arena score
#15 of 62 in Text-to-Image
Vote tally
Where the votes landed
Qwen Image 2512
40.0%
win rate
Ties
20.0%
Seedream 4.0
40.0%
win rate
40.0%
20.0% ties
40.0%
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent photographic realism and texture on the book and table.
- + Accurate soft lighting from the left following the prompt.
- + Superior glass rendering with convincing thickness and reflections.
- − The glass box appears more like a rectangular prism than a perfect cube.
Seedream 4.0
- + Perfect geometric cube shape.
- + Good representation of the blue sphere's translucency.
- + Clever use of caustic light patterns on the table.
- − The plant appears to be growing inside the cube rather than behind it, violating the spatial logic of the prompt.
- − The book has a slight floating effect on the left edge.
Verdict: Both models followed the prompt's instructions for the specific objects and colors. Qwen Image 2512 is the winner because it correctly placed the plant behind the cube, whereas Seedream 4.0 made the plant appear as if it were inside the glass. Qwen Image 2512 also provided a much more realistic texture for the wooden table and red book.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent skin texture and elderly facial features
- + Strong cinematic lighting and wet pavement reflections
- + High resolution with great detail on the bicycle frame
- − The subject is posing/staring at the camera rather than repairing the bike
- − The background cars lack the motion blur requested
- − Bicycle anatomy is slightly warped toward the rear
Seedream 4.0
- + Captured the 'repairing' action perfectly
- + Excellent application of motion blur on the passing vehicle
- + Natural, 'candid' composition with tools on the ground
- − Physical artifacts on the bicycle (disconnected spokes and frame parts)
- − Character's face and hair are slightly blurry/low detail
- − Hands are poorly defined while interacting with the bike
Verdict: Seedream 4.0 followed the prompt's narrative and technical requirements much more closely, successfully depicting the act of repairing, horizontal motion blur on cars, and an imperfect candid framing. Qwen Image 2512 has superior textures and facial detail, but it feels like a staged portrait rather than the candid action scene requested, failing the 'repairing' and 'motion blur' instructions. Seedream 4.0 is the winner for better prompt adherence and atmosphere, despite some structural issues with the bicycle's spokes.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Extremely high detail in the metal engraving and skin texture
- + Excellent rendering of light and shadow from the torch
- + Lifelike iris detail and realistic sweat/dirt particles
- − The torch in the background is a bit distracting in its shape
- − The symmetry of the braids feels slightly artificial
Seedream 4.0
- + Great sense of movement with the bokeh sparks
- + Good composition with a more dynamic pose
- + Accurate depiction of the cloth underlayer and leather straps mentioned in the prompt
- − Slightly less realistic skin texture compared to Model A
- − The armor engraving is not as crisp or intricate
Verdict: Qwen Image 2512 is the superior image due to its incredible textural detail, particularly in the weathered skin and the ornate filigree on the breastplate. While Seedream 4.0 does a great job with the atmosphere and meeting all the prompt's clothing requirements, it lacks the raw photographic realism and sharpness found in Qwen Image 2512.
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.”
Qwen Image 2512
Seedream 4.0
0% wins
0% ties
100% wins
AI Judge Analysis
Qwen Image 2512
- + Excellent layout that actually looks like a professional restaurant menu
- + Good adherence to the 'grid' request for food photos
- + Includes placeholders for prices and descriptions, creating a realistic document
- − Text is mostly gibberish or illegible upon closer inspection
- − Food photos are somewhat repetitive in style
Seedream 4.0
- + Text rendering for section headers is clear and legible
- + High-quality, vibrant food photography
- − Fails the 'menu design' prompt by providing a collage rather than a functional document
- − Poor composition with awkward white spaces and overlapping crops
- − Does not include menu items or sections for appetizers/pizza/mains as requested beyond simple headers
Verdict: Qwen Image 2512 successfully interprets the prompt as a design task, creating a realistic and structured menu layout with defined sections and a grid of photos. Seedream 4.0 produces a disjointed collage of images that lacks the professional formatting of a menu. Despite the illegible small text, Qwen Image 2512 is the superior choice for its adherence to the composition and design requirements.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Perfect adherence to all text requirements including the specific €6.99 price.
- + High photorealistic detail in food textures, particularly the patty and tomatoes.
- + Excellent professional layout that feels like a real advertisement.
- − The burger is slightly less 'exploded' than Model B, appearing more like a tall stack with some debris.
Seedream 4.0
- + Great sense of motion with circular motion blur trails.
- + Dynamic 'exploded' composition with components widely separated.
- − Incorrect price displayed as €5.99 instead of the requested €6.99.
- − The burger composition is slightly confusing, appearing to have multiple patties and buns in different states of assembly.
- − The text 'LIMITED TIME ONLY' is slightly warped.
Verdict: Qwen Image 2512 is the clear winner as it followed all instructions perfectly, including the specific price point and starburst rendering. While Seedream 4.0 had a very dynamic sense of motion, it failed on the text accuracy and the burger assembly looked more like two messy halves rather than one exploded whole.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent calligraphy style that looks very elegant and professional
- + Perfectly legible text including the very long bottom disclaimer
- − Includes a spelling error ('Risitto')
- − The chalk texture feels slightly like a digital overlay rather than physical chalk dust
Seedream 4.0
- + Superb realism regarding 'chalk texture' with smudges and dust buildup
- + Correct spelling of 'Risotto'
- + Very natural variation in letter size and slant as requested
- − The cursive is slightly less 'elegant' compared to Model A
- − A few letters in the bottom disclaimer are slightly messy
Verdict: Seedream 4.0 is the winner because it successfully captured the 'chalk texture' and natural messiness of a real chalkboard while adhering to the spelling of the prompt items. Qwen Image 2512 has a beautiful font style, but the spelling error in the first item and the overly clean texture make it feel less authentic than Seedream 4.0.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent anatomical details on the horse's musculature and skin.
- + Highly realistic textures on the space suit and leather saddle.
- + Clean, cinematic composition with a clear horizon line.
- − Failed the specific negative constraint; the astronaut is riding the horse instead of the horse being on top.
Seedream 4.0
- + Beautiful celestial background with nebulae and star effects.
- + Good dynamic posing of the horse's mane and tail.
- + Consistent lighting across the horse and astronaut.
- − Failed the the specific negative constraint; the astronaut is riding the horse.
- − Significant anatomical errors with the horse's front legs and hooves.
Verdict: Both models failed the specific prompt instruction to place the 'horse on top' of the astronaut, providing a standard 'astronaut riding a horse' image instead. Qwen Image 2512 is the superior image as it features much higher anatomical accuracy and more realistic technical details compared to Seedream 4.0, which has visible glitches in the horse's legs.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent photorealistic texture on the capybara's fur.
- + Highly cinematic lighting that captures the NY night atmosphere.
- + Subject expressions perfectly match the 'calm' and 'bored' descriptors.
- − The paws on the steering wheel look slightly anthropomorphized and unnatural.
Seedream 4.0
- + Strong composition showing both the internal and external street elements.
- + Accurate representation of a capybara's paw anatomy on the wheel.
- + Good use of depth of field on the background city lights.
- − The capybara's hat looks a bit like a modern baseball cap rather than a traditional driver cap.
- − Slightly less 'photorealistic' skin texture on the human passenger.
Verdict: Qwen Image 2512 produces a more high-end cinematic image with superior textures and lighting, making it feel more like a professional film still. Seedream 4.0 follows the composition instructions well and handles the anatomy of the animal paws more realistically, but it lacks the visual polish and realism found in Qwen's output.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent typography rendering with almost perfect spelling
- + Very clean and balanced composition following the request for a border
- + Crisp, high-resolution details on the jack-o-lantern and thorn border
- − Has a slight misspelling of 'Halloween' as 'Hallowern' in the main title
- − The background trees feel a bit more generic compared to the foreground
Seedream 4.0
- + Adds a very effective torn parchment effect that enhances the vintage feel
- + Moody atmospheric lighting and more organic, twisted tree shapes
- + Correct spelling in the primary title text
- − Text in the scroll banner is messy and difficult to read
- − The overall composition feels a bit more cluttered with the overlapping thorns
Verdict: Qwen Image 2512 produces a cleaner, more professional-looking invitation with superior font rendering and layout, despite a minor typo in the title. Seedream 4.0 captures the 'vintage parchment' texture more effectively, but fails on the legibility of the banner text and overall clarity.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Excellent text rendering with a stylized 3D outline
- + Strong diorama aesthetic with a distinct raised base
- + High-quality textures on the sushi and garnish
- − The sushi and plate are slightly off-center compared to the text
Seedream 4.0
- + Precise 45° isometric angle
- + Clean and simple text layout
- + Vibrant colors and realistic lighting on the salmon and ikura
- − The diorama base is a simple footed tray rather than a raised platform piece
- − Text is a bit more generic in style compared to the 3D art style of the prompt
Verdict: Qwen Image 2512 better captures the 'miniature diorama' and '3D cartoon' feel requested, using stylized text that complements the scene perfectly. Seedream 4.0 provides a very clean, high-quality isometric render with excellent sushi textures, but falls slightly behind on the creative interpretation of a diorama base.
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.”
Qwen Image 2512
Seedream 4.0
33% wins
33% ties
33% wins
AI Judge Analysis
Qwen Image 2512
- + Excellent fur texture and individual strand detail
- + Features all four requested animals clearly
- + Very clean, high-resolution rendering with no major anatomical errors
- − The animals are posing for a portrait rather than 'playfully chasing' or 'tumbling'
- − The lighting feels a bit more artificial/studio-like despite the god rays
Seedream 4.0
- + Stronger adherence to the 'playfully chasing' and 'tumbling' part of the prompt
- + Beautiful atmospheric lighting with realistic dew sparkles and backlighting
- + Dynamic composition that conveys a 'joyful wholesome vibe' effectively
- − The kitten has some anatomical blurring/muddiness in its paws
- − The fox's face lacks the high-fidelity realism found in Model A
Verdict: Qwen Image 2512 produces a much higher quality '8K' image with incredible fur detail and perfect anatomy, but it treats the prompt as a static group portrait. Seedream 4.0 captures the spirit of the prompt much better, showing the animals in motion and interacting with the environment, though it suffers from slightly lower technical clarity and some minor artifacts in the animals' limbs.
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.”
Qwen Image 2512
Seedream 4.0
100% wins
0% ties
0% wins
AI Judge Analysis
Qwen Image 2512
- + Excellent typography with professional swashes
- + Higher level of illustrative detail on the cloche and steam
- + Sophisticated use of texture and shading
- − The steam is a bit heavy, slightly clashing with the 'minimalist' prompt
Seedream 4.0
- + Follows the 'minimalist' aspect of the prompt more closely
- + Accurate typography and banner placement
- + Clean vector style
- − The steam effect is very basic and small
- − Cloche illustration lacks the depth and premium feel of the other model
Verdict: Both models followed the prompt instructions perfectly, including the specific name and founding date. Qwen Image 2512 is the winner because it provides a much more professional and aesthetically pleasing illustration with superior typography and shading, whereas Seedream 4.0 looks a bit like a generic clip-art style logo.
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.”
Qwen Image 2512
Seedream 4.0
AI Judge Analysis
Qwen Image 2512
- + Strong adherence to the flat-vector aesthetic with a clean navy background.
- + Excellent Saturn V and Lunar Module illustrations that maintain professional technical detail.
- + Successfully included the NASA-inspired color palette and astronaut names.
- − Significant text hallucinations and spelling errors in the step descriptions.
- − Step numbering and sequence are chaotic and overlapping.
Seedream 4.0
- + Clear, logical infographic flow from left to right with a dedicated landing section.
- + Accurate spelling of the mission steps and names compared to the other model.
- + Follows the specific icon requests for the trajectory arc and orbit rings effectively.
- − The 'Descent Surface' text on step 5 is slightly crowded.
- − The rocket illustration is less detailed than in the competing model.
Verdict: While Qwen Image 2512 produces more beautiful individual illustrations, it fails as an infographic because the text is garbled and the sequence is confusingly numbered. Seedream 4.0 is the clear winner for this task as it provides a legible, well-structured layout with accurate spelling and a clear mission progression that follows the prompt's logical constraints.
Next steps
Explore each model
ByteDance's image generation model with integrated text-to-image and image editing capabilities in a unified architecture, supporting up to 4K resolution