Unified multimodal model for text-to-image generation, instruction-guided image editing, personalized generation, and virtual try-on
Settled by community votes across 15 shared challenges, with an AI judge weighing in on each.
OmniGen v2
#57 of 62 in Text-to-Image
Seedream 5.0 Lite
#22 of 62 in Text-to-Image
Where the votes landed
OmniGen v2
0.0%
win rate
Ties
0.0%
Seedream 5.0 Lite
100.0%
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
OmniGen v2
- + Excellent refraction and reflection within the glass cube.
- + Superior image resolution and sharp textures on the book and sphere.
- + Very accurate adherence to the plant being visible through the glass.
- − The glass cube appears to have an open top or lacks a defined top surface under the book.
- − The lighting, while soft, feels slightly more clinical than Image B.
Seedream 5.0 Lite
- + Beautifully captures the 'soft window light' with warm tones and realistic shadows.
- + The glass cube has a clearly defined top surface supporting the book.
- + Natural composition with realistic depth of field.
- − The plant's visibility through the glass is less distinct compared to Model A.
- − Slightly lower sharpness on the fine details of the book's texture.
Verdict: Both models followed the prompt perfectly, but OmniGen v2 stands out for its technical rendering of glass physics and local detail. However, Seedream 5.0 Lite captured the requested lighting mood and atmosphere more effectively, creating a more convincing photographic feel.
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
OmniGen v2
- + Excellent full-body reflection on the wet pavement
- + Good adherence to the 'light rain' prompt with visible raindrops
- − The man is just holding the bike, not repairing it
- − The image feels somewhat staged rather than candid
- − The bicycle has structural issues, specifically the drivetrain and chain positioning
Seedream 5.0 Lite
- + Successfully captures the act of repairing the bike
- + Excellent implementation of motion blur from a passing car
- + Provides a truly candid feel with realistic skin textures and imperfect framing
- − The hands have some anatomical merging with the bike chain area
- − Less emphasis on the light rain visual compared to Model A
Verdict: Seedream 5.0 Lite followed the specific technical aspects of the prompt much better, including the requested motion blur and the specific action of 'repairing' the bicycle. While OmniGen v2 produced a clean image with beautiful reflections, it felt more like a posed portrait rather than the candid street photography 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
OmniGen v2
- + Excellent engraving detail on the pauldrons and chestplate
- + Clean, high-contrast lighting from the torch
- + Clear focal point on the eyes and face
- − The character looks too clean and youthful for the 'battle-worn' description
- − Beads in the hair are few and look like simple dots
- − Dirt on the skin looks more like intentional freckles or makeup
Seedream 5.0 Lite
- + Successfully captures a 'battle-worn' appearance with visible facial scars and stubble
- + Superior execution of the braided hair with numerous small beads
- + Detailed texture on the cloth underlayer and leather straps
- − Lighting is slightly more diffused and less dramatic than image A
- − Some minor blending issues where the braids meet the shoulders
Verdict: Seedream 5.0 Lite followed the prompt much more closely, particularly regarding the 'battle-worn' aspect and the specific request for braided hair with beads. While OmniGen v2 produced a very clean and aesthetically pleasing portrait, it failed to convey the grit, scars, and complex hair details requested by the user.
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
OmniGen v2
- + Features a bold, vibrant use of color blocks for a modern aesthetic.
- + Includes diverse food photography following the grid layout.
- − Poor text rendering with several misspellings like 'RESTAURATED MENTS'.
- − The food photos lack consistency in lighting and style across the grid.
Seedream 5.0 Lite
- + Excellent text rendering with clear, legible English and logical pricing.
- + Perfectly adheres to the requested sections: appetizers, pizzas, and mains.
- + Clean professional layout that looks like a real-world functional menu.
- − The composition is a bit standard/conservative compared to more avant-garde designs.
Verdict: Seedream 5.0 Lite is the clear winner as it produces a fully functional, professional-grade menu with legible text, accurate categories, and high-quality food photography. OmniGen v2 struggles with typography and logical layout, resulting in a design that looks like a rough template with jibberish text.
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
OmniGen v2
- + Clean and legible typography for the main title
- + High-contrast, vibrant colors that fit an advertisement style
- + Good execution of the 'fire' background concept
- − Failed the core 'exploded'/suspended instruction by showing a fully stacked burger
- − Missing the Euro symbol (€) in the price tag
- − The 'Limited Time Only' text is partially obscured by the background glow
Seedream 5.0 Lite
- + Excellent adherence to the 'exploded' and suspended burger components instruction
- + Perfectly rendered all requested text including the Euro symbol and starburst
- + Very photorealistic textures on the tomato, lettuce, and bun
- − The 'Magic Burger' text is slightly less prominent than requested compared to the burger size
Verdict: Seedream 5.0 Lite followed every instruction perfectly, including the difficult 'exploded' layout and all specific text strings and symbols. OmniGen v2 failed to explode the burger, delivering a standard stacked burger and missing the Euro symbol, making it a poor fit for the specific prompt requirements.
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
OmniGen v2
- + Includes a nice wooden frame around the blackboard.
- + The text is very legible and follows a consistent chalk-like style.
- − Numerous spelling errors including 'SPECALS', 'Musonhom', and 'Lemont'.
- − The layout of the menu items is disorganized and overlapping, making it hard to read the prices.
- − Failed to provide the elegant cursive title requested.
Seedream 5.0 Lite
- + Excellent spelling accuracy for the primary menu items like 'Truffle Mushroom Risotto'.
- + Realistic chalkboard texture with dust, smudges, and authentic chalk line quality.
- + Clean, professional layout that matches a real cafe environment.
- − Small minor spelling errors in the footer text 'frese opstoons' and 'Heriss'.
- − Missing the 'elegant cursive' requirement for the title line.
Verdict: Seedream 5.0 Lite is the clear winner as it provides a realistic, high-quality image of a chalkboard with mostly accurate spelling and a professional layout. In contrast, OmniGen v2 suffers from significant spelling distortions and a messy, incoherent arrangement of the menu text.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
OmniGen v2
- + Clean, vector-like aesthetic with clear subjects
- + Accurate depiction of an astronaut's suit details
- − Failed the spatial prompt regarding the 'horse on top'
- − Background is overly simplistic and lacks 'cinematic' depth
- − Lighting feels flat and digital rather than surreal
Seedream 5.0 Lite
- + Beautiful cinematic background with rich nebulae and lighting
- + More dynamic pose suggesting movement through space
- + High level of texture detail in the horse's coat and mane
- − Failed the primary prompt instruction for the horse to be on top
- − The horse's hind legs are poorly formed and blend into the tail weirdly
Verdict: Both OmniGen v2 and Seedream 5.0 Lite failed to follow the specific spatial instruction to place the 'horse on top' of the astronaut, choosing instead to render the traditional image of an astronaut riding a horse. Seedream 5.0 Lite is the preferred image because its lighting, background, and overall composition are significantly more cinematic and visually impressive compared to the simplistic rendering of OmniGen v2.
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
OmniGen v2
- + The capybara's head has realistic fur texture and a clear expression.
- + The businessman's facial expression conveys the requested boredom/normalcy.
- − Serious anatomical error with human hands attached to the capybara's body.
- − The capybara's head is poorly integrated onto the torso, looking like an overlay.
- − The passenger appears to be in the front seat rather than the back.
Seedream 5.0 Lite
- + Successfully rendered capybara paws on the steering wheel.
- + Excellent composition with clear depth between the driver and the backseat passenger.
- + Much more detailed taxi interior including dashboard and roof signage.
- − The passenger's face is slightly blurry compared to the driver.
- − The hands of the passenger are somewhat claw-like and lack fine detail.
Verdict: Seedream 5.0 Lite is the clear winner because it correctly placed the capybara's paws on the wheel, whereas OmniGen v2 gave the capybara human hands and arms. Seedream 5.0 Lite also followed the spatial instructions accurately, placing the businesswoman in the back seat, while OmniGen v2 placed her in the front next to the driver.
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
OmniGen v2
- + Strong, clean typography for the main title
- + Good use of color contrast with the golden glow
- + Follows the square format request correctly
- − Significant spelling errors in the subtitle and event details
- − The art style feels more like a modern vector illustration than a vintage gothic parchment
- − The layout of the bottom text is cluttered and repetitive
Seedream 5.0 Lite
- + Perfect text rendering for all requested details including date, time, and location
- + Excellent adherence to the 'vintage gothic' and 'parchment' aesthetic requirements
- + Includes all thematic elements like thorns, webs, and twisted trees with high visual quality
- − The composition is a bit crowded around the central jack-o-lantern
- − The jack-o-lantern 3D render style slightly clashes with the 2D parchment texture
Verdict: Seedream 5.0 Lite performed significantly better on this task, accurately rendering every piece of requested text with perfect spelling and fulfilling the 'vintage' aesthetic. OmniGen v2 struggled with the specific text details, producing several typos and a more cartoonish illustration style that lacked the requested gothic atmosphere.
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
OmniGen v2
- + Excellent typography with clean drop shadows
- + High-quality PBR material rendering on the salmon textures
- + Strong 3D pop and vibrant colors
- − The flag icon is incorrect for Japan
- − Anatomically strange sushi (shrimp tail on a salmon roll piece)
Seedream 5.0 Lite
- + Accurate Japanese flag icon
- + Correct variety of sushi types (Nigiri and Maki)
- + Better adherence to the 'light blue background' prompt
- − Text is not perfectly top-centered as requested
- − Text rendering is slightly less crisp than Model A
Verdict: OmniGen v2 produces a more visually striking 3D render with superior typography, but fails on the cultural details like the flag. Seedream 5.0 Lite adheres better to the specific content requests, providing an accurate flag and a more realistic sushi variety, despite slightly weaker text layout.
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
OmniGen v2
- + Strong god ray effects and vibrant lighting
- + Clear, well-defined butterflies
- − Failed to include the requested baby bunny
- − Style is overly 'cartoonish' rather than 'hyper-photorealistic'
- − The animal on the right is a strange fox-rabbit hybrid with a cotton tail
Seedream 5.0 Lite
- + Included all four requested animals: golden retriever, tabby kitten, baby bunny, and fox kit
- + Captures the 'playfully chasing' and 'tumbling' interaction much better
- + Successfully rendered dew sparkles on the grass as requested
- − Anatomy on the fox's legs is slightly awkward while tumbling
- − The god rays are a bit soft/blurry compared to the rest of the scene
Verdict: Seedream 5.0 Lite followed the prompt significantly better than OmniGen v2, successfully including all four specific animals while OmniGen v2 missed the bunny. Seedream 5.0 Lite also achieved a much more realistic yet whimsical aesthetic, whereas OmniGen v2 produced a flat, illustrative style that ignored the 'hyper-photorealistic' instruction.
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
OmniGen v2
- + Strongly recreates the 'distracted boyfriend' pose while applying a clean anime style
- + Good use of vibrant colors and crisp line art
- − Loses the facial expressions of the original, with the girlfriend looking happy rather than angry
- − The background is simplified into generic shapes rather than Ghibli-style scenery
Seedream 5.0 Lite
- + Excellent preservation of the original facial expressions, especially the girlfriend's anger
- + Closer stylistic match to Studio Ghibli's hand-drawn character design and facial features
- + Preserves the original background architecture and depth better
- − The line work on the man's arm and the girl's hand is slightly disconnected
- − Colors are a bit muted compared to typical 'vibrant' Ghibli landscapes
Verdict: Seedream 5.0 Lite is the winner because it successfully transforms the image into a Ghibli aesthetic while perfectly preserving the character expressions and intent of the 'distracted boyfriend' meme. OmniGen v2 provides a clean style but fails the prompt by making the girlfriend look happy/smiling, which fundamentally changes the context of the source image.
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
OmniGen v2
- + Successfully added wind motion to hair
- + High vibrance and clean visual quality
- + Added leaves as requested
- − Failed to preserve the source image's identity, completely regenerating the subject and dog
- − Leaves look more like flat orange graphics than natural elements
Seedream 5.0 Lite
- + Excellent source preservation, keeping the original woman, dog, and background almost perfectly intact
- + Dynamic hair motion looks natural and maintains hair texture
- + Leaves are integrated with lighting and variety
- − The leaf placement can feel slightly like a static overlay in some areas
Verdict: This is a clear comparison between a model that regenerates a scene from scratch versus a model that performs a true edit. OmniGen v2 failed to preserve the source image, replacing the people and animals with entirely different versions, whereas Seedream 5.0 Lite successfully applied the requested dynamic motion (hair and leaves) while keeping the original image pixel-perfectly similar in all other aspects.
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
OmniGen v2
- + Clean vector-style execution
- + Strong visual balance in the circular emblem
- + Good use of high-contrast chocolate brown and cream tones
- − Major spelling error in main text ('CAFFFLORIN')
- − Failed to include the accent mark on the 'e'
- − Minimal subtle texture on the background
Seedream 5.0 Lite
- + Accurate spelling of 'Caffè Florian' including the accent mark
- + Pleasant subtle texture on the background
- + Very clean typography that matches the vintage aesthetic
- − The 'E' in 'Caffè' uses a circumflex (Caffê) instead of an accent grave
- − Asymmetrical banner tails on the 'Est. 1720' ribbon
- − The cloche is a bit flat compared to the header text
Verdict: Seedream 5.0 Lite is the clear winner for its superior text rendering, correctly spelling the brand name while OmniGen v2 produced a significant typo ('CAFFFLORIN'). Seedream 5.0 Lite also better captured the requested 'subtle texture' on the background and established a more elegant vintage feel.
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
OmniGen v2
- + Successfully uses the requested color palette.
- + Achieves a clean, flat-vector aesthetic.
- − Significant text errors including 'APOLO 17' instead of Apollo 11 and 'NSA' instead of NASA.
- − Failed to include all 6 requested steps, showing only a confusing layout of icons.
- − Iconography is abstract and does not clearly relate to the specific mission phases.
Seedream 5.0 Lite
- + Excellent prompt adherence, including all 6 specific mission stages with correct labels.
- + High-quality text rendering and accurate names for the crew and landing site.
- + Clear and consistent iconography that matches the flat-vector style requested.
- − The 'Descent' and 'Landing' icons are nearly identical with minor additions.
- − Small inclusion of blue in the earth/module icons departs slightly from the strict navy/red/gray palette.
Verdict: Seedream 5.0 Lite performed significantly better, accurately following all instructions including the specific 6 steps of the mission and correct text for Apollo 11. OmniGen v2 failed on almost every text requirement, erroneously labeling the mission as Apollo 17 and misspelling NASA, while only providing a partial and confusing infographic layout.
Explore each model
ByteDance's image generation model with built-in reasoning, example-based editing, and deep domain knowledge, supporting up to 3K resolution