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
Settled by community votes across 17 shared challenges, with an AI judge weighing in on each.
Seedream 4.5
#9 of 62 in Text-to-Image
Z-Image Turbo
#12 of 62 in Text-to-Image
Where the votes landed
Seedream 4.5
54.5%
win rate
Ties
36.4%
Z-Image Turbo
9.1%
win rate
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent photographic lighting with realistic caustic shadows on the table.
- + High tactile detail on the book's canvas texture and the glass's edges.
- + Successfully places the plant behind the glass with accurate refraction.
- − The blue 'sphere' has a heart-like indentation, failing to be a perfect sphere.
- − The perspective of the glass cube is slightly distorted at the bottom edge.
Z-Image Turbo
- + Perfectly rendered blue sphere with a clear reflection on the bottom glass.
- + Accurate interpretation of all spatial requirements including 'behind' and 'on top'.
- + Very clean, modern composition with a natural depth of field.
- − The plant is significantly blurred, making it less clearly 'visible through the glass' as requested.
- − The lighting is a bit flat compared to the dramatic shadows in the other image.
Verdict: Both models followed the prompt instructions near-perfectly. Seedream 4.5 captures much more realistic lighting and materials, especially with the caustic reflections on the table, but failed to make the sphere perfectly round. Z-Image Turbo produced a cleaner, more accurate geometric sphere and better overall scene composition, making it the more reliable interpretation despite the softer lighting.
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent adherence to the 'motion blur from passing cars' prompt with light trails.
- + Highly realistic skin textures and wet surfaces with cinematic lighting.
- + Captures the 'repairing' action effectively with a visible tool and engagement with the chain.
- − The bicycle mechanics are slightly surreal, with the chain appearing to pass through the frame/back of the hub.
- − The subject's eyes are looking directly at the camera, which feels slightly less 'candid'.
Z-Image Turbo
- + The man's posture and downward gaze feel very natural and candid.
- + Good depiction of a rainy environment and wet asphalt.
- − Completely missed the prompt for 'motion blur from passing cars', as the background vehicles are static.
- − The action is more 'walking' or 'standing with' the bike rather than 'repairing' it.
- − Lower overall lighting complexity and shallower depth of field than requested.
Verdict: Seedream 4.5 is the clear winner as it followed almost every specific detail of the prompt, including the difficult-to-execute motion blur of passing cars and the specific 'repairing' action. Z-Image Turbo produced a high-quality but generic image that ignored the motion blur and cinematic lighting requests, resulting in a less dynamic scene.
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.”
AI Judge Analysis
Seedream 4.5
- + Exquisite macro detail on the skin, including pores and subtle scars.
- + The engraved plate armor exhibits realistic light diffusion and high-quality textures.
- + Excellent depth of field with cinematic lighting that feels integrated into the scene.
- − The braiding is less physically distinct than in model b, feeling more like surface ornamentation.
Z-Image Turbo
- + Strong adherence to the braiding and beads requested in the prompt.
- + Good overall composition showing more of the armor and the torch source.
- + Clear depiction of the ‘battle-worn’ aspect through facial grime and bloody scars.
- − Lower resolution textures on the armor and skin compared to the competitor.
- − The character's expression is somewhat stiff and the lighting on the face feels slightly flat.
- − The fire sparks appear a bit digital and lack the soft bokeh quality requested.
Verdict: Seedream 4.5 delivers a superior 'close portrait' with incredible photographic detail in the skin and metal textures, creating a much more lifelike image. While Z-Image Turbo does a better job of including all prompt elements like the full braids and the physical torch, its overall visual fidelity and lighting are less sophisticated than Seedream 4.5.
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.”
AI Judge Analysis
Seedream 4.5
- + Perfect adherence to section headings (Appetizers, Pizza, Mains)
- + Clean, minimalist layout that actually looks like a functional menu
- + Excellent photo quality for the food items
- − Alignment of text and colored borders is slightly uneven
- − Contains minor spelling errors in menu item names
Z-Image Turbo
- + Stronger 'grid' layout as requested in the prompt
- + Uses bold, modern typography effectively for headers
- + Vibrant colors and high-quality food photography
- − Failed to include 'Mains' correctly, misspelling it as 'MANS'
- − Layout is cluttered and lacks logical flow for a restaurant menu
- − Repeating pizza photos and generic list items make it less practical
Verdict: Seedream 4.5 is the winner because it successfully organized the menu into the three requested sections with high-quality images and a professional layout. While Z-Image Turbo followed the 'grid' instruction more literally, it suffered from significant spelling errors and a confusing hierarchy that failed to function as a readable menu.
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.”
AI Judge Analysis
Seedream 4.5
- + Perfect execution of the 'exploded' burger request with dynamic motion blur.
- + Accurate and high-quality rendering of all requested text elements.
- + Superior photorealistic detail in the food textures like the cheese strings and patty.
- − None notable.
Z-Image Turbo
- + Clean and vibrant lighting effect on the text and starburst.
- + Solid food photography aesthetic.
- − Failed to create an 'exploded' burger as requested, showing a mostly assembled stack.
- − Included redundant text 'MAGIC BURGER BURGER'.
- − The starburst graphic is slightly less integrated into the environment.
Verdict: Seedream 4.5 followed the complex prompt instructions perfectly, specifically capturing the 'exploded' motion of the ingredients and rendering the text accurately. Z-Image Turbo failed to separate the burger components and added redundant words to the title, making it less effective as an advertisement.
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent chalk texture with natural-looking smudges and dust
- + Perfect spelling for all requested menu items
- + Very realistic background atmosphere and lighting
- − Redundant line for the first item repeating the price
Z-Image Turbo
- + Strong handwriting consistency across the entire board
- + Clean layout with good spacing between sections
- + Legible and neat presentation
- − Several spelling errors including 'Mustroom' for 'Mushroom'
- − The chalk texture looks a bit more like a digital brush than real chalk smudges
Verdict: Seedream 4.5 captures a much more authentic café atmosphere with incredibly realistic chalk textures and perfect spelling, despite a minor repetition error on the first line. Z-Image Turbo has a cleaner layout but suffers from spelling mistakes like 'Mustroom' and lacks the natural grit and smudging that makes the handwriting in Seedream 4.5 feel truly physical.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
Seedream 4.5
- + Excellent cinematic lighting with a vibrant, colorful nebula background.
- + High level of detail on the space suit and the horse's fur textures.
- + Dynamic composition that feels truly 'in space' rather than floating.
- − Internal anatomy errors, specifically a random human foot appearing out of the horse's stomach.
Z-Image Turbo
- + Very clean rendering of the horse and astronaut equipment.
- + Good clarity on the stirrups and saddle details.
- + Anatomically correct horse silhouette compared to the competitor.
- − The background is relatively flat and less 'cinematic' than requested.
- − Lighting is a bit sterile and doesn't match the space environment as well as model_a.
Verdict: Seedream 4.5 captures the 'cinematic' and 'surreal' aspects of the prompt much more effectively with its rich colors and lighting, despite a significant anatomical glitch (an extra foot). Z-Image Turbo provides a technically cleaner subject, but the overall image feels less like 'space' and more like a cutout on a dark background. Seedream 4.5 is preferred for its artistic adherence to the cinematic requirement.
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent adherence to the 'two front paws on the steering wheel' instruction.
- + Clear, legible text on the driver's cap.
- + Superior photorealistic quality in the textures of the capybara fur and the lighting of the city background.
- − The scale of the capybara relative to the dashboard is slightly exaggerated.
Z-Image Turbo
- + Features a more traditional, formal chauffeur-style cap.
- + Accurate passenger pose and 'bored' expression.
- − Failed the instruction to have both paws on the steering wheel (only one paw is visible).
- − The steering wheel position and rotation look anatomically incorrect for the capybara's reach.
- − The passenger is out of focus compared to the passenger in model A.
Verdict: Seedream 4.5 is the clear winner as it followed all specific prompt constraints, particularly the requirement for the capybara to have both paws on the steering wheel. Seedream 4.5 also produced a more convincing night-time atmosphere with better text rendering on the hat, whereas Z-Image Turbo struggled with the paw placement and overall image sharpness.
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.”
AI Judge Analysis
Seedream 4.5
- + Perfect text rendering for both titles and numerical event details.
- + Sophisticated cinematic lighting and atmospheric depth in the background.
- + Integrated the scroll banner and border elements naturally into the composition.
- − The thorns on the border look more like barbed wire than organic thorns.
- − Composition places the event details heavily at the bottom, overlapping some background textures.
Z-Image Turbo
- + Excellent vintage parchment texture with realistic torn edges.
- + Artistic thorny vines and spider web details that match the gothic theme perfectly.
- + Strong central composition that feels like a physical poster.
- − Contains a spelling error in 'The Archves' (instead of Arches).
- − Missing the small scroll for the 'invited' text, instead using scrolls for the main title and footer.
- − The lighting on the pumpkin feels slightly flatter compared to the environment.
Verdict: Seedream 4.5 is the winner because it followed the text instructions perfectly, including the specific phrasing on the scroll and accurate spelling of the location. While Z-Image Turbo had a more authentic vintage 'paper' feel, its spelling error and failure to place the specific text on a banner make it less effective as a functional invitation.
Bald man challenge
Image Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
Seedream 4.5
- + Perfect preservation of original facial features and clothing
- + Highly realistic hair texture and natural hairline integration
- + Consistent lighting and shadows on the new hair
- − None
Z-Image Turbo
- + Follows the general composition of the source image
- − Fails to add a 'full, thick head of hair' as requested, only adding a light stubble
- − Significantly alters the person's facial features and bone structure
- − Removes the glasses from the subject
- − Changes the background and overall lighting
Verdict: Seedream 4.5 successfully performed a complex localized edit by adding realistic hair while perfectly preserving every other detail of the source image. In contrast, Z-Image Turbo failed the prompt by only adding stubble, while simultaneously changing the subject's face, removing his glasses, and altering the background.
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.”
AI Judge Analysis
Seedream 4.5
- + Perfect adherence to the requested flag (Japan)
- + Excellent PBR materials with realistic textures on the salmon and rice
- + Superior diorama base with organic textures and depth
- − The 'J' and 'A' in the text are slightly overlapping
- − Slight depth-of-field blur on the rear sushi piece
Z-Image Turbo
- + Very clean, rounded cartoon aesthetics
- + Clear and legible typography
- + Solid adherence to the isometric perspective
- − Significant factual error: used the flag of China instead of Japan
- − Textures look more like plastic than realistic PBR materials
- − The diorama base is very simple and lacks the requested 'refined texture'
Verdict: Seedream 4.5 is the clear winner as it correctly rendered the Japanese flag and applied much higher quality PBR textures to the sushi, making it look both appetizing and artistic. Z-Image Turbo failed a key part of the prompt by displaying the flag of China for a 'Japan' themed image, and its materials appear overly simplistic.
Over-the-top cartoon caricature
Editing“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
AI Judge Analysis
Seedream 4.5
- + Excellent caricature style with exaggerated head and features while maintaining likeness
- + Perfect adherence to all prompt elements: TV anchor desk, microphone, hockey equipment, and a dog
- + High visual quality with sharp details and balanced composition
- − The transition from the 3D-style caricature head to the slightly more realistic desk and background is a bit sharp
Z-Image Turbo
- + Successfully added a small dog in the background while preserving the original image's style
- − Completely failed the 'caricature' instruction, keeping a realistic face
- − Ignored the 'TV show anchor' and 'hockey' elements entirely
- − The dog in the background is blurry and lacks detail
Verdict: Seedream 4.5 followed the complex editing instructions perfectly, creating a fun, high-quality caricature that integrated all requested elements (dog, hockey, news desk) while clearly maintaining the subject's likeness. Z-Image Turbo failed most of the prompt requirements, only adding a small, blurry dog in the background and neglecting the caricature and professional elements entirely.
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent depiction of dawn lighting with god rays and dew sparkles as requested.
- + Dynamic composition showing movement and energy with the 'tumbling' action.
- + Rich, vibrant colors and high level of detail in the fur texture.
- − The fox's eyes appear slightly oversized and 'cartoonish' rather than hyper-photorealistic.
- − Small anatomical glitch on the cat's front right paw.
Z-Image Turbo
- + Great expressions on the animals, particularly the happy puppy and fox.
- + Clean, soft lighting that creates a wholesome atmosphere.
- + Clearer visibility of all four requested animals in the center of the frame.
- − The lighting lacks the requested 'god rays' and specific golden sunrise intensity.
- − Static composition that feels more like a posed photo than 'playfully chasing and tumbling'.
- − The kitten's mouth and eyes have slight AI artifacts upon close inspection.
Verdict: Seedream 4.5 captures the cinematic lighting and dynamic movement described in the prompt significantly better, using god rays and a sense of action. While Z-Image Turbo produces a very cute and bright image, it is more of a static group portrait and misses the atmospheric 'masterpiece' lighting effects requested. Seedream 4.5 is the winner for its superior composition, lighting, and adherence to the 'tumbling' action.
Studio Ghibli Anime Style
Editing“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
AI Judge Analysis
Seedream 4.5
- + Perfectly captures the Studio Ghibli art style with characteristic line work and facial features.
- + Excellent use of soft watercolor textures and pastel backgrounds as requested.
- + Maintains the exact composition and identifiable subject matter of the source meme while transforming the medium.
- − The man's expression is slightly more neutral compared to the original puckered lips in the photo.
Z-Image Turbo
- + Preserves the photo-realistic details of the original image with very high fidelity.
- + Successfully keeps the exact facial expressions and clothing of the subjects.
- − Completely failed the styling instruction; there are no Ghibli, hand-painted, or illustration elements present.
- − Colors remain realistic rather than being transformed into the requested pastel/dreamy palette.
Verdict: Seedream 4.5 followed the complex stylistic instructions perfectly, recreating the meme as a classic Studio Ghibli anime cel with beautiful watercolor textures. Z-Image Turbo failed to apply any of the requested edits, returning a result that is effectively identical to the source photo with minor variations.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
Seedream 4.5
- + Excellent adherence to the 'windy' instruction with highly dynamic hair movement.
- + Vivid and prominent leaves flying throughout the scene.
- + Preserved the identity and pose of the subject and dog very well.
- − The leaf rendering is a bit large and sharp, lacking some motion blur for realism.
- − The background lighting was slightly altered towards a golden hour glow.
Z-Image Turbo
- + Successfully added subtle motion to the hair.
- + Maintained the original lighting and color palette of the source image.
- + Integrated the flying leaves with more natural sizes.
- − Failed to preserve the leash from the original image.
- − The overall sense of energy and motion is much lower than requested.
- − Subtle changes to facial features make the subject look slightly different.
Verdict: Seedream 4.5 is the clear winner as it fully embraced the 'energetic and lively' request, providing dramatic hair movement and plenty of flying leaves while keeping the core elements of the original photo intact. Z-Image Turbo was too conservative with the requested edits and failed a key part of source preservation by removing the dog's leash entirely.
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent typography with elegant arched placement.
- + Superior vector shading and hatch marks on the cloche and banner.
- + Accurate rendering of the accent mark in 'Caffè'.
- − The 'f's in 'Caffè' are slightly inconsistent in their descenders.
Z-Image Turbo
- + Clean, minimalist design that fits the vector emblem request.
- + Good contrast between the brown and cream tones.
- + Correct spelling and accent placement.
- − The 'Est. 1720' text is slightly off-center within the banner.
- − The cloche design is a bit generic compared to the first model.
- − The typography is less sophisticated than the arched style in Model A.
Verdict: Seedream 4.5 is the clear winner as it provides a much more professional and aesthetically pleasing logo design. Its use of arched typography, detailed vector shading, and a well-composed banner creates a more authentic vintage feel than the flatter, slightly off-center design of Z-Image Turbo.
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.”
AI Judge Analysis
Seedream 4.5
- + Excellent text rendering with correct spelling for mission members and steps.
- + Clean professional infographic layout with a logical 1-6 timeline.
- + Strong adherence to the NASA-inspired color palette and vector style.
- − The 'Descent' icon shows a satellite-like object rather than a lunar module.
- − The Saturn V rocket is missing its characteristic black and white roll patterns.
Z-Image Turbo
- + Features a bold, clear title and large iconography.
- + Includes icons for astronauts and landing site as requested in supporting details.
- − Numerous spelling errors including 'Apolio', 'Translurian', and 'Descenty'.
- − Composition is disorganized and loses the sequential 'steps' format requested.
- − Yellow and orange accents deviate from the requested muted red palette.
Verdict: Seedream 4.5 creates a high-quality, professional infographic that perfectly follows the requested 6-step structure and maintains perfect spelling. Z-Image Turbo suffers from significant typographical errors and fails to organize the information into a cohesive timeline, making it much less functional as an infographic.
Explore each model
Tongyi-MAI's 6-billion parameter distilled text-to-image model optimized for speed, achieving high-quality generation in 8 steps or fewer with support for bilingual text rendering