Recraft's latest text-to-image generation model with high-quality output, supporting various aspect ratios and custom color palettes
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
Recraft V4
#11 of 62 in Text-to-Image
Seedream 4.5
#9 of 62 in Text-to-Image
Where the votes landed
Recraft V4
40.0%
win rate
Ties
20.0%
Seedream 4.5
40.0%
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
Recraft V4
- + Excellent photorealistic texture on the wooden table and book edges.
- + Accurate glass refraction and reflections on the sphere.
- + Strong adherence to the lighting direction from the left.
- − The sphere appears to be floating mid-air, which is physically impossible without support.
- − The plant is somewhat cluttered and dominates the upper background.
Seedream 4.5
- + Perfect compliance with spatial instructions, placing the sphere on the floor of the cube.
- + Very clear visibility of the plant through the glass panels.
- + Clean and elegant composition with realistic shadows.
- − The lighting on the cube's top edge and book feels a bit flat compared to the table.
Verdict: While Recraft V4 produces a more high-fidelity, gritty photograph, it fails the logic of physics by making the sphere float. Seedream 4.5 captures all prompt elements perfectly, including the spatial relationships and the visibility of the plant through the glass, making it the superior choice for prompt adherence.
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
Recraft V4
- + Excellent adherence to the 'imperfect framing' and 'candid' aspects of the prompt.
- + Highly realistic light rain and pavement reflection quality.
- + Accurate 50mm-like shallow depth of field and authentic motion blur on passing cars.
- − The red bicycle is partially obscured and its structure is slightly anatomically incorrect at the fork.
- − The man's face is hidden, making it harder to judge natural skin texture.
Seedream 4.5
- + Superb natural skin texture and facial detail on the subject.
- + Clearer focus on the specific action of repairing the bike with a tool.
- + Strong cinematic lighting and color contrast.
- − The motion blur on the car looks more like a digital filter than a long exposure.
- − The composition feels more like a posed portrait rather than a 'candid street photo' with 'imperfect framing'.
- − The bicycle chain and wrench interaction is physically non-functional.
Verdict: Recraft V4 better captures the specific technical photography requests like 'imperfect framing' and 'candid street photo,' resulting in an image that feels like a real snapshot. Seedream 4.5 produces a more beautiful and detailed subject portrait with excellent skin textures, but it feels more staged and artistic rather than the raw, realistic style requested.
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
Recraft V4
- + Excellent depiction of ornate engraving on the armor
- + Very clear and detailed skin texture showing dirt and grime
- + Strong composition with a direct, intense gaze
- − Lighting feels slightly flat compared to the prompt's 'warm torchlight' description
- − Braids look a bit stiff and artificial
- − Skin scars look more like skin growths or bumps than flat battle scars
Seedream 4.5
- + Sublime lighting and atmospheric bokeh that perfectly captures warm torchlight
- + Incredibly lifelike eye reflections and skin detailing
- + Braids with metal beads are rendered with high realism and better integration
- − The leather strap stitching/texture is a bit messy upon close inspection
- − Armor engraving is less crisp than in Model A
Verdict: Both models followed the prompt exceptionally well, but Seedream 4.5 captures the cinematic atmosphere much better with its superior lighting and skin shaders. While Recraft V4 has more defined armor engravings, Seedream 4.5's lifelike eyes and realistic hair integration make for a more high-quality portrait.
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
Recraft V4
- + Excellent text rendering with almost zero spelling errors in item names and descriptions.
- + Perfect adherence to the grid layout requirement with clean, professional item isolation.
- + Accurately includes all requested sections: Appetizers, Pizza, and Mains.
- − The layout is quite sparse with large amounts of white space.
- − One spelling error is present in the word 'solmon' in the salmon description.
Seedream 4.5
- + High-quality, appetizing photography for the food featured in the layout.
- + Uses vibrant color-coded borders to create a visually interesting accent theme.
- − Major text failures with gibberish words and repetitive placeholder text.
- − Does not follow the 'grid' layout requested, instead using a single vertical column.
- − Prices are unrealistic for a casual dining menu ($79 - $99).
Verdict: Recraft V4 is the clear winner as it produces a functional, professional menu with accurate text, correct logical sections, and a clean grid layout that perfectly matches the prompt. Seedream 4.5 fails significantly on the textual requirements, producing gibberish words and failing to implement the requested grid structure.
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
Recraft V4
- + Excellent photographic texture on the bun and meat patty
- + Accurate rendering of the price in the requested euro currency
- + Clean typography for all three text requirements
- − The 'fiery background' is mostly dark smoke with very few embers
- − The burger composition feels static rather than dynamic or in motion
- − The price starburst is very thin and lacks the requested 'glowing' effect
Seedream 4.5
- + Dynamic composition with strong motion blur and diagonal orientation
- + Superior background that perfectly matches the 'fiery' and 'glowing' prompt
- + Strong adherence to the 'glowing effect' for all text elements
- − The bottom half of the burger is still mostly assembled, losing some 'exploded' effect
- − The price starburst is slightly cluttered with overlapping fire graphics
Verdict: Seedream 4.5 captures the intended energy of the ad much better, with a truly fiery background and a more dynamic sense of motion in the burger's suspension. While Recraft V4 has slightly more realistic food textures, it feels like a static studio shot compared to the explosive, glowing aesthetic achieved by Seedream 4.5.
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
Recraft V4
- + Excellent chalk texture with realistic smudges and dust on the board.
- + Perfect spelling and punctuation for all requested text.
- + Clean, readable layout with a strong sense of depth in the background.
- − The title is not in the requested 'elegant cursive' style.
- − The board feels slightly flat compared to the wooden frame.
Seedream 4.5
- + Features a more vertical, chalkboard-accurate composition.
- + Includes convincing chalk dust and texture variations.
- + Handwriting has a more natural 'human' slant and variation in line weight.
- − Missed the 'elegant cursive' requirement for the title.
- − The text layout is a bit crowded at the bottom edge.
Verdict: Recraft V4 produced a technically superior image with perfect spelling and a highly detailed chalk texture that feels authentic to a café setting. While Seedream 4.5 also handled the text well, Recraft V4's rendering of the font characters was more consistent and readable, making it the more effective menu mockup.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
Recraft V4
- + Excellent cinematic lighting and texture on the horse's coat.
- + High level of detail in the asteroid environment and space background.
- + Clearer, more realistic aesthetic.
- − Failed the negative prompt constraint: the astronaut is on top of the horse, not vice versa.
Seedream 4.5
- + Successfully followed the difficult spatial reasoning prompt 'horse on top'.
- + Vibrant, colorful nebula background.
- + Creative interpretation of the 'surreal' aspect of the prompt.
- − Anatomy of the horse and astronaut is distorted and poorly rendered.
- − Lower image clarity with several visual artifacts and messy bridles.
Verdict: Recraft V4 produced a much higher quality image in terms of detail and cinematic appeal, but it completely ignored the specific spatial instruction to put the horse on top of the astronaut. Seedream 4.5 successfully interpreted the 'horse on top' request, creating a truly surreal scene, even though its technical execution of anatomy and detail is significantly weaker.
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
Recraft V4
- + Excellent photorealism with convincing skin textures and material rendering
- + Captures a very authentic New York street atmosphere through the windows
- + The passenger's outfit and disinterested expression perfectly match the 'bored businesswoman' prompt
- − The capybara's anatomy is slightly awkward, appearing more like a large head on a human-like body
- − The yellow cap is simple and lacks the text often found on driver caps
Seedream 4.5
- + Perfect adherence to the instruction for the capybara to have both front paws on the steering wheel
- + Very clear 'TAXI' text on the driver's cap
- + Strong frontal composition that makes the capybara's expression the focal point
- − The passenger's face is slightly soft and lacks the realistic detail found in the capybara's fur
- − The lighting on the capybara's fur seems a bit too studio-like compared to the background
Verdict: Recraft V4 produces a more cinematic and photorealistic outcome with superior texture work on the human character and car interior. However, Seedream 4.5 captures the specific technical requirements of the prompt better, specifically the 'paws on the wheel' and the 'TAXI' cap, while maintaining a very expressive central character.
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
Recraft V4
- + Excellent typography style that feels very gothic
- + Highly detailed background with weather effects like rain
- + Sophisticated spider web border design
- − The 'small scroll banner' is just a flat white rectangle
- − Background is a bit cluttered with many small bat shapes
Seedream 4.5
- + Perfect implementation of the scroll banner request
- + Clearer 'parchment' texture on the border
- + Excellent legibility for the event details at the bottom
- − The thorns in the border look more like modern barbed wire
- − Composition is a bit more generic than Model A
Verdict: Both models followed the prompt's complex text requirements perfectly. Seedream 4.5 captures the vintage aesthetic better with its parchment texture and actual scroll banner, while Recraft V4 offers a more atmospheric, cinematic illustration but fails the 'scroll' element by using a flat white shape.
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
Recraft V4
- + Excellent typography rendering with clean, professional spacing
- + High-fidelity textures on the fish and ice-block base
- + Perfectly centered and balanced composition
- − Lean more toward photo-realistic than the requested '3D cartoon scene' style
- − The ice base is a bit busy and distracting compared to a clean diorama
Seedream 4.5
- + Perfectly captures the '3D cartoon' and 'miniature' aesthetic requested
- + Warm, gentle lighting that enhances the stylized PBR materials
- + Very clean diorama base that fits the isometric theme well
- − Text alignment is slightly off-center and the flag icon is placed awkwardly
- − Lower resolution/clarity on the text edges compared to Model A
Verdict: Recraft V4 delivers a more polished, high-resolution graphic design that looks like a professional advertisement, though it misses the 'cartoon' stylistic prompt. Seedream 4.5 much more accurately follows the stylistic prompt of a 3D isometric cartoon miniature, though its text layout is less refined. Seedream 4.5 is the winner for better adherence to the specific 'look and feel' requested.
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
Recraft V4
- + Naturalistic and realistic lighting and fur textures.
- + Excellent field of depth with butterflies in the fore, mid, and background.
- + Highly detailed meadow with realistic dew and wildflowers.
- − The fox's front right leg has an awkward anatomical structure.
- − The god rays are a bit subtle compared to the request for 'masterpiece' lighting.
Seedream 4.5
- + Stronger adherence to 'god rays' and 'golden sunrise' light.
- + Dynamic posing that captures the 'tumbling/chasing' aspect well.
- + Vibrant colors that enhance the joyful vibe.
- − The fox looks significantly more stylized and less 'photorealistic' than the other animals.
- − The kitten's front right paw is anatomically messy and blends into the grass.
Verdict: Recraft V4 produces a more grounded and realistic image with superior texture work, especially on the fur and meadow. While Seedream 4.5 captures the magical lighting and dynamic motion better, it suffers from some anatomical distortions and a slight shift toward a CGI/illustration style rather than the requested hyper-photorealism.
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
Recraft V4
- + Excellent custom typography with stylish ligatures
- + Clean vector aesthetic and balanced composition
- + Accurate spelling including the grave accent on Caffè
- − Missed the 'banner' requirement for the Est. 1720 text
- − Texture on background is extremely subtle to the point of being almost absent
Seedream 4.5
- + Successfully included the 'Est. 1720' banner as requested
- + Good use of warm cream and brown tones
- + Subtle paper-like texture on the background
- − Typography is less sophisticated and a bit generic
- − The 'a' in Florian is slightly distorted
- − Steam lines are a bit disconnected from the cloche
Verdict: Both models followed the prompt well, but Model B is the technical winner for including the 'banner' element and the requested background texture which Model A ignored. While Recraft V4 produced a more professional and stylish logo, Seedream 4.5 adhered more closely to every specific detail of the written instructions.
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
Recraft V4
- + Excellent typography with a professional, modern aesthetic
- + Accurate and creative icons that follow the requested narrative flow
- + Clean layout that uses white space effectively
- − The lunar orbit icon uses a globe that looks identical to Earth instead of the moon
Seedream 4.5
- + Strict adherence to the 1-6 numbering requested in the prompt
- + Accurate icon for the moon in the orbit phase
- + Clear, linear information hierarchy
- − The 'Descent' icon is a generic satellite rather than the lunar module
- − Overall composition feels more cluttered and less 'modern' than the competitor
Verdict: Recraft V4 produces a much more polished and professional design that looks like a real infographic poster, despite a small error where it repeats the Earth icon for the Moon's orbit. Seedream 4.5 captures the step-by-step numbering better but fails on the visual quality of the 'Descent' icon and generally has a more dated graphic style.
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