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 4.5
26.1
arena score
#9 of 62 in Text-to-Image
Vote tally
Where the votes landed
Qwen Image 2512
40.0%
win rate
Ties
0.0%
Seedream 4.5
60.0%
win rate
40.0%
0.0% ties
60.0%
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent photographic quality and lighting
- + Realistic glass physics and plant visibility
- + Perfect adherence to the spatial arrangement requested
- − The glass has a slight cyan tint not explicitly requested, though it looks natural
Seedream 4.5
- + Accurate adherence to all objects in the prompt
- + Good use of high-contrast window lighting
- − The glass cube has broken geometry, appearing as a solid block that the sphere is clipping through
- − The sphere is heart-shaped or deformed rather than a simple sphere
- − Perspective on the table and cube is slightly skewed
Verdict: Qwen Image 2512 is the clear winner as it produces a realistic, coherent photograph with believable glass reflections and a correctly rendered sphere inside the cube. Seedream 4.5 struggles with the spatial logic, making the cube look like a solid chunk of glass that the blue object is magically stuck inside, and the object itself is not a perfect sphere.
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 4.5
67% wins
0% ties
33% wins
AI Judge Analysis
Qwen Image 2512
- + Natural and highly detailed skin texture
- + Excellent composition with realistic background traffic
- + Accurate 50mm lens feel with a shallow depth of field
- − The subject is posing rather than 'repairing' the bicycle
- − Rain is very subtle, making the scene feel more post-rain than in light rain
Seedream 4.5
- + More dynamic motion blur from passing cars
- + Shows the active repair motion with a tool
- + Captures the atmosphere of light rain and wet pavement reflections more vividly
- − Anatomical issues with the left hand merging into the wrench/bike
- − Skin texture on the face is slightly smoothed compared to the prompt's request
- − Distorted bicycle wheel spokes
Verdict: Qwen Image 2512 produces a superior photographic portrait with impressive skin realism and a lens-accurate background, though the man is simply sitting with the bike rather than repairing it. Seedream 4.5 captures the 'action' and atmospheric rain much better, but suffers from significant AI artifacts in the hands and bicycle geometry. Qwen Image 2512 is the winner due to its photographic coherence and adherence to the 'no stylization' request.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent facial textures with realistic scars and skin grain
- + Strong adherence to the 'braided hair with small beads' requirement
- + Great balance of torchlight reflection on the metallic surfaces
- − The bokeh sparks appear a bit static and noisy in the top right
- − The armor engraving is slightly shallower compared to Model B
Seedream 4.5
- + Ornate engraving is highly detailed and three-dimensional
- + Superior shallow depth of field and soft bokeh effects
- + Life-like eyes with complex reflections and colors
- − The beads in the hair are less prominent and more integrated into the braids
- − The dirt/grime on the skin looks slightly speckled rather than naturally smeared
Verdict: Both models performed exceptionally well, capturing the atmosphere of a battle-worn knight perfectly. Qwen Image 2512 is preferred for its superior depiction of the braided hair and beads, whereas Seedream 4.5 offers a more cinematic depth of field and highly intricate armor engravings.
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 4.5
0% wins
0% ties
100% wins
AI Judge Analysis
Qwen Image 2512
- + Excellent photographic density with a nice 2x5 grid of food items.
- + Captures the 'bold sans-serif' and 'modern minimalist' aesthetic perfectly.
- + Realistic food photography with high detail and appetizing colors.
- − Internal body text is garbled/illegible.
- − Combines 'Pizza' and 'Mains' into one section instead of separating them as requested.
Seedream 4.5
- + Text rendering is significantly clearer and legible for headings.
- + Distinct colorful borders around photos provide 'vibrant accents'.
- + Strictly adheres to the three specific sections: Appetizers, Pizza, and Mains.
- − Composition is more like a slide or digital menu than a printed layout.
- − Smaller variety of food photos compared to Model A.
Verdict: Qwen Image 2512 produces a better overall layout that feels like a professional physical menu, but features illegible text and merges categories. Seedream 4.5 offers much better text clarity and strict adherence to the requested category structure, though the layout is simpler and less dense. Seedream 4.5 is the winner for its superior text legibility and categorical accuracy.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent text rendering with clean, professional typography.
- + Highly detailed burger textures with sharp focus on all ingredients.
- + Includes all requested elements including embers, starburst, and accurate 'exploded' debris.
- − The 'exploded' effect feels a bit more like debris scattered around rather than the core burger layers being separated.
Seedream 4.5
- + Strong sense of motion with motion blur on flying ingredients.
- + Good fiery glow effect on the main title text.
- + Atmospheric background with visible coals and fire.
- − The secondary 'LIMITED TIME ONLY' text is slightly warped.
- − The starburst graphic for the price is less integrated and looks a bit like a generic sticker.
- − Lower overall sharpness compared to the other model.
Verdict: Qwen Image 2512 produces a superior advertisement with professional-grade typography and exceptionally sharp, appetizing food photography. While Seedream 4.5 captures the 'motion' aspect well with blur effects, it lacks the overall polish and text clarity found in Qwen Image 2512.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent chalk texture and realistic handwriting variations.
- + Correct spelling of complex words like 'Mushroom' and 'Octopus'.
- + Beautiful elegant cursive title as requested.
- − Spelling error in 'Risitto' instead of 'Risotto'.
Seedream 4.5
- + Natural-looking chalkboard dust and smudges.
- + Clear, legible handwriting style.
- + Correctly completed the truncated prompt for 'Brown Butter Chocolate Chip Cookies'.
- − Repeated the price for the first item on two lines.
- − The handwriting looks slightly more uniform and digital compared to the other model.
Verdict: Qwen Image 2512 produces a much more realistic chalk texture and an artistic, elegant handwriting style that fits the 'cozy café' prompt perfectly, despite a minor typo in 'Risotto'. Seedream 4.5 is very legible and follows the content instructions well, but the layout is slightly redundant with the price repetition. Overall, Qwen Image 2512 feels more like an authentic, hand-drawn chalkboard.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + High resolution and crisp textures on the spacesuit and horse fur.
- + Cinematic lighting with a clear view of the astronaut's face through the visor.
- − Completely failed the spatial reasoning instruction to have the horse on top of the astronaut.
- − The horse's back legs are anatomyically confused, appearing to blend into a tail-like structure.
Seedream 4.5
- + Highly atmospheric and surreal background with vibrant nebulas.
- + Good composition with a dynamic horse pose and stardust effects.
- − Completely failed the negation/spatial instruction by placing the astronaut on top of the horse.
- − The astronaut's left leg is disconnected and floating behind the horse's back.
Verdict: Both Qwen Image 2512 and Seedream 4.5 failed the specific spatial logic prompt to have the horse riding the astronaut, opting instead for the standard trope. Seedream 4.5 produced a more visually artistic and surreal composition, whereas Qwen Image 2512 reached for a higher level of photorealistic detail despite technical anatomical errors.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent photorealistic texture on the capybara fur and jacket.
- + Strong cinematography with realistic depth of field and bokeh lights.
- + Captures a very expressive and slightly humorous 'bored' face on the passenger.
- − The capybara's paws look somewhat mangled and humanoid/claw-like.
- − The hat looks more like a police or captain's hat than a standard taxi cap.
Seedream 4.5
- + Successfully renders text 'TAXI' on the cap which adds to the theme.
- + Great composition that clearly shows both the driver and the passenger as requested.
- + The capybara has a very authentic 'calm' and professional expression.
- − The passenger's phone screen is blank/dark when it should likely be on.
- − The lighting on the passenger is slightly flat compared to the driver.
Verdict: Both models followed the complex prompt very well, but Seedream 4.5 is the winner due to better adherence to details like the 'TAXI' text on the cap and more natural-looking paws on the steering wheel. While Qwen Image 2512 has slightly better lighting and fur texture, the anatomy of the hands and the ambiguous hat choice make it less effective than Seedream 4.5.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent gothic typography that matches the requested aesthetic
- + Highly detailed thorny border design that encircles the image
- + Accurate rendering of the jack-o-lantern and scroll banner
- − Spelling error in the main title: 'Hallowern' instead of 'Halloween'
Seedream 4.5
- + Perfect text rendering with no spelling errors
- + Clean cinematic lighting and atmospheric background
- + Good adherence to the scroll banner and event detail requirements
- − The thorns are limited to the bottom corners rather than a full border
- − The font choice for the event details is a bit modern and plain for a gothic theme
Verdict: Both models followed the prompt instructions very closely, capturing the spooky atmosphere and specific text requirements. Seedream 4.5 is the winner because it successfully spelled all text correctly, whereas Qwen Image 2512 had a typo in the main title ('Hallowern'), which is a significant flaw for an invitation design.
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 4.5
0% wins
0% ties
100% wins
AI Judge Analysis
Qwen Image 2512
- + Excellent miniature 3D aesthetic with soft, clay-like textures.
- + Text rendering is stylized and fits the cartoon theme perfectly.
- + Rich details on the sushi materials, particularly the rice and fish grain.
- − The text is slightly off-center compared to the diorama base.
- − The flag icon is a bit simplified/stylized compared to a real flag.
Seedream 4.5
- + Perfectly centered composition following the prompt's layout instructions.
- + Clean, professional font choice for 'JAPAN' and 'SUSHI'.
- + Realistic material rendering on the salmon and the diorama base texture.
- − The flag icon is positioned awkwardly to the side rather than 'below' the main text header.
- − Slightly less variety in the sushi types presented compared to Model A.
Verdict: Qwen Image 2512 wins by a narrow margin due to its superior artistic cohesion; while both models followed the prompt well, Qwen's stylized text and richer diorama details better capture the requested '3D cartoon scene' aesthetic. Seedream 4.5 produced a very clean and centered image but the text and flag placement felt less integrated into the overall design.
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 4.5
AI judge analysis unavailable for this challenge.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent illustration style with vintage woodcut-like shading.
- + Perfect text rendering for both the main title and the banner.
- + Accurate and artistic representation of steam as part of the logo.
- − The design is more of a complex illustration than a 'minimalist' logo.
- − The banner is slightly asymmetrical at the ends.
Seedream 4.5
- + Follows the 'minimalist' and 'vector emblem' part of the prompt much better.
- + Clean, professional layout suitable for actual branding usage.
- + Very accurate text rendering and color palette adherence.
- − The steam icon is a bit generic compared to the rest of the logo.
- − Minor kerning/spacing issues in the word 'Caffè'.
Verdict: Qwen Image 2512 produces a stunning vintage illustration with superior artistic detail and texture, though it leans more towards an ornate badge than 'minimalist'. Seedream 4.5 delivers a true minimalist vector logo that is much more practical for branding, although it lacks the atmospheric charm of the other. Qwen Image 2512 wins slightly due to the high quality of its classic typography and the more sophisticated execution of the cloche and steam.
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 4.5
AI Judge Analysis
Qwen Image 2512
- + Excellent flat-vector illustration style for the Saturn V and Lunar Module.
- + Includes the names of all three astronauts correctly.
- + Strong aesthetic depth with a consistent celestial theme.
- − Significant text hallucinations and spelling errors in the step labels.
- − The numbered steps are disorganized and repetitive.
- − Includes a meta prompt snippet 'Steps stop at landing' inside the visual.
Seedream 4.5
- + Perfect text rendering for all mission labels and names.
- + Logical, clean infographic layout with a clear timeline progression.
- + Strict adherence to the NASA-inspired color palette and flat-vector style.
- − The icon for 'Descent' appears to be a generic satellite rather than a lunar module.
- − The icons for steps 5 and 6 are visually disconnected from the timeline line.
Verdict: Seedream 4.5 is the clear winner as it successfully produces a functional infographic with perfect spelling and a logical layout. Qwen Image 2512 has superior artistic illustrations of the spacecraft but fails as an infographic due to severe text hallucinations and a disorganized numbering system.
Next steps
Explore each model
ByteDance's latest image generation model unifying text-to-image and image editing in a single architecture, with improved text rendering and 30-40% faster generation than v4.0