GPT Image 1.5 vs Seedream 5.0 Lite
Head-to-head across 17 challenges
GPT Image 1.5
60.0%
win rate
Ties
0.0%
Seedream 5.0 Lite
40.0%
win rate
Challenge Results
Man and Car in California
Editing“Make a photo of the man driving the car down the California coastline”
AI Judge Analysis
GPT Image 1.5
- + Successfully places the specific man and car into a California coastline setting.
- + The lighting on the man's face matches the outdoor environment well.
- + Captures a high-quality, cinematic perspective of the road ahead.
- − The car is a right-hand drive model, which is incorrect for driving in California.
- − The car's exterior design is significantly altered from the source image (the hood and grille are mostly cropped out).
Seedream 5.0 Lite
- + Excellent preservation of the source car, maintaining the grille, wheels, and logo details.
- + Correctly places the man as a left-hand driver for a US/California setting.
- + Very convincing motion blur on the road and wheels, enhancing the 'driving' feel.
- − The man's scale relative to the car is slightly off; he appears a bit small in the cockpit.
Verdict: Seedream 5.0 Lite followed the instructions much better by maintaining the original car's identity and correctly placing the driver on the left side for a California setting. GPT Image 1.5 changed the car to a right-hand drive model and cropped out most of the car's defining features, resulting in a less accurate edit.
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
GPT Image 1.5
- + Excellent integration of textures, showing a highly detailed and appetizing patty.
- + Strong adherence to the 'fiery' text effect and 'dark, fiery background' description.
- + Includes more complex visual elements like flying sauce splashes and embers that enhance the sense of motion.
- − The burger is less 'exploded' vertically, making some layers harder to distinguish clearly.
- − Small amounts of messy AI artifacts around the flying crumbs and lettuce edges.
Seedream 5.0 Lite
- + Perfect 'exploded' composition with clear separation of every single ingredient.
- + Clean, professional graphic design aesthetic with high-contrast neon-style text.
- + Very sharp, clean image with minimal artifacts.
- − The 'starburst' for the price is a simplified vector-style outline rather than a fiery effect.
- − The background is less dynamic and intense than requested by the prompt.
Verdict: GPT Image 1.5 excels at creating a cinematic, gritty, and highly detailed food advertisement that feels cohesive with the fiery theme. Seedream 5.0 Lite delivers a much clearer 'exploded' view of the ingredients with cleaner text, but it misses the intensity and specifically the fiery rendering of the price starburst. GPT Image 1.5 is the winner for its superior texture and atmosphere that better matches the 'Magic' and 'Fiery' keywords.
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
GPT Image 1.5
- + Excellent photorealism with cinematic lighting and realistic textures.
- + Perfect adherence to 'bored' expression for the passenger and 'professional' for the capybara.
- + Logical composition with the capybara in the foreground and passenger in the background.
- − The capybara's paws look slightly more like hands/fingers than natural capybara feet.
Seedream 5.0 Lite
- + High resolution with very clear details in the background city lights.
- + Captures the requested outfit and hat accurately.
- + Unique wide-angle composition showing more of the car exterior.
- − Confused spatial logic where the passenger is sitting in the front passenger seat instead of the back seat.
- − The capybara's face looks slightly static and overlaid onto the body.
- − The background lighting is a bit too bright and clean, losing the gritty NYC night atmosphere.
Verdict: GPT Image 1.5 is the clear winner as it correctly follows the spatial instructions of the prompt, placing the passenger in the back seat to create a humorous contrast. While Seedream 5.0 Lite has impressive clarity, it fails the basic composition by putting the passenger in the front seat, and the cinematic atmosphere in GPT Image 1.5 is much more convincing and photorealistic.
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
GPT Image 1.5
- + Excellent depiction of a realistic, wet environment with believable puddles and reflections.
- + The elderly man's posture and skin texture feel very authentic and less 'perfected'.
- + Captures a gritty, unstylized atmosphere that fits the 'candid street photo' request.
- − The car in the background lacks the requested motion blur, appearing mostly static.
- − The bicycle's rear wheel and frame geometry are physically inconsistent.
Seedream 5.0 Lite
- + Successfully incorporates visible motion blur on the passing car as requested.
- + Very high quality skin texture and hair detail on the subject.
- + Stronger 50mm shallow depth of field effect with cinematic lighting.
- − The man's hands have structural issues, specifically a strange protrusion on the right hand's wrist/thumb area.
- − The bicycle chain and derailleur are floating and not properly connected to the wheel.
Verdict: Both models followed the prompt well, but GPT Image 1.5 captured the 'imperfect' and 'unstylized' requested aesthetic more naturally, despite failing the motion blur requirement. Seedream 5.0 Lite produced a more polished, cinematic image with correct motion blur, but it suffered from noticeable anatomical errors in the hands and mechanical errors in the bicycle.
Pose & Character Mashup
Editing“Use Image 1 as the exact pose reference and Image 2 as the character reference. Recreate the person/character from Image 2 in the exact dynamic pose and body position from Image 1. Keep the exact face, hair, clothing style/details, and expression from Image 2. Match the lighting and environment of Image 1. The final image must show the character from Image 2 performing the precise action/pose from Image 1 with perfect anatomy and natural integration.”
AI Judge Analysis
GPT Image 1.5
- + Excellent character preservation of the face and accessories
- + High visual clarity and resolution
- + Successful integration of all clothing elements including the scarf
- − Failed to match the extreme torso lean and head tilt of the orientation
- − Feet are positioned flat rather than in the crossed, dynamic balance of the reference
Seedream 5.0 Lite
- + Successfully captured the extreme head tilt and torso angle from the pose reference
- + Accurately recreated the complex leg/foot positioning on the stool
- + Maintained the character's signature look and accessories
- − Lower resolution and slight blurriness compared to Model A
- − Occasional artifacts around the hair and hand edges
Verdict: Both models did an impressive job of character transfer, but Seedream 5.0 Lite is the clear winner for pose accuracy, correctly replicating the precarious balance and head tilt of the reference image. GPT Image 1.5 produced a higher-quality render with cleaner details, but it defaulted to a much more upright and conventional standing pose, failing the core request of mirroring the dynamic position.
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
GPT Image 1.5
- + Excellent depiction of glass refraction and reflections on the sphere.
- + High level of detail in the textures of the table wood and the red book cover.
- + Sharp resolution and realistic lighting interaction with the surrounding environment.
- − The sphere is quite large relative to the cube, pushing the definition of 'small' sphere.
Seedream 5.0 Lite
- + Follows the relative scale of 'small sphere' better than the competitor.
- + Authentic soft window lighting consistent with the prompt's direction.
- + Clean, minimalist composition with a natural-looking plant.
- − The glass cube is missing its bottom surface/base, making the sphere appear to sit directly on the wood.
- − The rendering of the glass is slightly more artificial than Model A.
Verdict: Both models followed the prompt instructions perfectly regarding object placement. GPT Image 1.5 is the winner due to its superior rendering of glass and reflections, though Seedream 5.0 Lite followed the spatial 'small' sphere instruction slightly better. However, Seedream failed to realistically render the bottom of the glass cube, whereas GPT Image 1.5 handled the complex physics of glass and light with much higher fidelity.
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
GPT Image 1.5
- + Perfect spelling and text rendering for all requested items.
- + Very realistic chalk texture with smudges and natural variations in stroke pressure.
- + Excellent adherence to the 'elegant cursive' requirement for the title.
- − The framing is a bit tight, showing only the surface of the board rather than the 'cozy café' atmosphere.
Seedream 5.0 Lite
- + Good composition that suggests a physical object in a real environment.
- + Clean, legible handwriting that captures a different but valid chalkboard style.
- − Several spelling errors in the menu items, such as 'Heriss', 'Beliter', and 'frese'.
- − Failed the handwriting style requirement for the title, which was requested to be in 'elegant cursive'.
- − The text looks slightly like a digital overlay rather than integrated chalk texture.
Verdict: GPT Image 1.5 is the clear winner as it followed every instruction perfectly, including complex text spelling and the specific 'elegant cursive' style for the header. Seedream 5.0 Lite struggled with accuracy, producing multiple spelling errors and ignoring the cursive requirement.
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
GPT Image 1.5
- + Exceptional photographic realism with lifelike skin texture and eyes.
- + Highly detailed engraving and weathering on the plate armor.
- + Perfect execution of bokeh sparks and warm lighting reflections.
- − The braids are a bit messy, though fitting for a battle-worn look.
Seedream 5.0 Lite
- + Clear interpretation of the beaded braids requested in the prompt.
- + Ornate and clean engraving on the armor.
- + Good use of warm torchlight on the character's side.
- − The image has a CGI/video game cinematic look rather than a lifelike photo.
- − Armor looks too clean and shiny for a 'battle-worn' character.
Verdict: GPT Image 1.5 significantly outperforms Seedream 5.0 Lite in terms of realism, texture, and atmospheric lighting. While Seedream 5.0 Lite followed the specific instruction for beads more literally, GPT Image 1.5 captured the 'battle-worn' aesthetic with far more convincing skin details, grimy armor, and professional-grade photographic depth.
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
GPT Image 1.5
- + Excellent typography with clear hierarchies and actual dish descriptions.
- + Flawless text rendering with zero spelling errors or artifacts.
- + Very high-quality food photography that fits the professional restaurant aesthetic.
- − The layout uses a split-screen approach rather than a full grid interspersing photos with text.
- − One pizza photo shows a thick vegetable topping that looks slightly cluttered.
Seedream 5.0 Lite
- + Strictly follows the 'grid' instruction by pairing each item with its own photo.
- + Includes a restaurant name which adds to the realism of a menu design.
- + Clean, minimalist use of bold sans-serif fonts for a casual dining vibe.
- − Food photography is slightly lower in resolution compared to Model A.
- − Less detail provided for the menu items (no ingredients or descriptions).
- − Overall layout has a lot of empty white space between text and borders.
Verdict: GPT Image 1.5 produces a much more professional and realistic menu with complete descriptions and high-fidelity food imagery, although it places photos in a side-column rather than a tight grid. Seedream 5.0 Lite adheres better to the grid layout request and includes a header, but the overall design feels more like a template or a digital kiosk than a polished printed menu. GPT Image 1.5 is the winner for its exceptional text clarity and superior visual quality.
Outfit Transfer Challenge
Editing“Use Image 1 as the base person. Dress them in the exact elaborate outfit from Image 2 (including all layers, accessories, jewelry, and shoes). Carefully adapt the clothing to the body shape and pose in Image 1 while maintaining realistic fabric behavior, correct proportions, and perfect lighting/shadow matching. Keep the person’s exact face, hair, and background completely unchanged.”
AI Judge Analysis
GPT Image 1.5
- + Excellent transfer of the specific plaid scarf and navy peacoat textures.
- + Maintains the subject's vitiligo patterns with high fidelity on the face and hand.
- + Includes almost all accessories like the gold watch and ring.
- − The person's face shape and features have been slightly altered compared to Image 1.
- − The background has been simplified and slightly altered during the generation process.
Seedream 5.0 Lite
- + Near-perfect preservation of the background and the original wood structure from Image 1.
- + Subject's face and unique vitiligo markings remain exactly like the source image.
- + Captures the accessories from Image 2, including the sunglasses and jewelry.
- − The scarf pattern, while close, is simplified compared to the source image.
- − The hand with the ring has some slight anatomical blurring.
Verdict: Both models followed the complex instructions very well, effectively 'dressing' the subject from the first image in the clothes of the second. Seedream 5.0 Lite is the clear winner for its superior preservation of the source image's identity and background, whereas GPT Image 1.5 subtly changed the subject's facial features and reconstructed the background details.
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
GPT Image 1.5
- + Excellent texture matching the beard
- + Highly realistic, natural hairline
- + Flawless preservation of the original face, clothes, and background
- − None notable
Seedream 5.0 Lite
- + Natural-looking hair style
- + Good preservation of most original elements
- − Slightly altered facial features, particularly around the eyes and nose
- − The glasses frames look slightly more modern/different than the source
Verdict: GPT Image 1.5 performed a near-perfect edit, seamlessly integrating the new hair while preserving every detail of the original face and environment. Seedream 5.0 Lite also provided a good result, but subtly altered the person's features, resulting in a slightly different likeness compared to the source image.
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
GPT Image 1.5
- + Captures the subject's likeness effectively within a stylized caricature format.
- + Comprehensive integration of all prompt elements including studio equipment, hockey on screens, and multiple dogs.
- + Excellent text rendering with 'BREAKING NEWS' and 'NEWS' on the microphone.
- − The composition is a bit cluttered with many overlapping elements.
- − The subject's hands have some anatomical inconsistencies typical of AI.
Seedream 5.0 Lite
- + Clean, vibrant 2D illustration style that feels very 'caricature-like'.
- + Preserves the original outfit (denim jacket over black top) perfectly.
- + High level of clarity and very clean lines throughout the image.
- − The facial likeness is generic and less recognizable as the woman in the source image compared to Model A.
- − The hockey stick is being held in a slightly awkward, stiff manner.
Verdict: GPT Image 1.5 does a significantly better job of maintaining the subject's specific facial features in the caricature, while also creatively integrating the hockey theme into the background news monitors. Seedream 5.0 Lite produces a very clean and professional-looking illustration that preserves the original clothing, but the face is too stylized to be a personalized caricature of the specific person provided.
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
GPT Image 1.5
- + Excellent depiction of realistic fur texture and lighting on the animals.
- + Strong adherence to the 'tumbling together' part of the prompt with complex interaction.
- + Impressive lighting with god rays and dew sparkles that feel integrated into the scene.
- − The kitten has an anatomical error with a third paw appearing near its chest.
- − The composition is a bit crowded, making it slightly hard to distinguish the individual bodies of the animals.
Seedream 5.0 Lite
- + Clean, cute, and well-spaced composition allowing each animal to be clearly seen.
- + Accurate representation of all four requested animals with expressive eyes.
- + Bright, vibrant colors that enhance the 'joyful wholesome vibe'.
- − The style leans more toward 3D animation/digital art than the requested 'hyper-photorealistic' look.
- − The animals appear somewhat static and placed on the grass rather than 'tumbling' together.
- − The dew sparkles look like floating orbs rather than water droplets on grass.
Verdict: GPT Image 1.5 followed the stylistic prompt for 'hyper-photorealism' much better, achieving incredible fur detail and believable lighting, though it suffered from a minor anatomical glitch in the kitten's paws. Seedream 5.0 Lite produced a very charming and clean image, but it looks more like a high-quality Pixar film than a photograph. GPT Image 1.5 is the winner for its superior texture work and more dynamic interaction between the animals.
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
GPT Image 1.5
- + Excellent application of the 'soft pastel colors' and 'gentle lighting' requested in the prompt.
- + Captures a dreamy, nostalgic mood with hand-painted textures.
- + Preserves the composition and poses while successfully translating them into a specific art style.
- − The faces lean more towards generic Shoujo anime than the specific Ghibli aesthetic.
- − The heavy blooming and soft focus make some details of the original scene less clear.
Seedream 5.0 Lite
- + Highly accurate Ghibli-style character designs, especially the eyes and line work.
- + Strong preservation of the source image's background details and sharpness.
- + Perfectly captures the expressions of the original meme in an animated style.
- − Doesn't lean as heavily into the 'soft pastel' or 'dreamy background' request as the other model.
- − The lighting feels a bit flat compared to the requested 'warm, nostalgic' atmosphere.
Verdict: Seedream 5.0 Lite does an incredible job of capturing the character design and line art associated with Studio Ghibli, making the iconic 'distracted boyfriend' meme look like a direct frame from a movie. However, GPT Image 1.5 followed the stylistic modifiers of the prompt much more closely, providing the requested soft textures and nostalgic lighting, even if its character designs are more generic. Seedream 5.0 Lite is the likely winner for its superior ability to mimic the Ghibli identity while keeping the source image recognizable.
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
GPT Image 1.5
- + Excellent hair motion effect that looks natural and dynamic.
- + High density of leaves creates a strong sense of wind and energy.
- + Near-perfect preservation of the subject's face, clothing, and the dog.
- − Some leaves in the foreground appear slightly blurry or lower resolution than the background.
Seedream 5.0 Lite
- + Good hair motion effect that matches the prompt.
- + Leaves are clear and have a consistent green color palette.
- − The leaf placement feels a bit sparse and static compared to the 'dynamic' request.
- − Minor loss of detail on the woman's face compared to the original.
Verdict: GPT Image 1.5 is the winner as it successfully creates a much more 'energetic and lively' atmosphere with a high volume of swirling leaves and excellently rendered wind-blown hair. While Seedream 5.0 Lite followed the instructions, its interpretation was more conservative and the leaves look somewhat pasted on rather than part of a dynamic gust of wind.
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
GPT Image 1.5
- + Excellent typography with professional kerning and stylistic flourishes
- + Sophisticated use of texture and shading on the cloche
- + Accurate rendering of the accent in 'Caffè'
- − Ignored the request for a light background
- − Cloche stem/steam is slightly off-center
Seedream 5.0 Lite
- + Followed the background color instruction perfectly
- + Clean vector style that feels more aligned with a modern logo
- + Good text clarity
- − The accent on 'Caffè' is reversed (grave vs acute)
- − Banner geometry is slightly awkward and lacks the vintage refinement of Model A
- − Composition feels a bit fragmented with large gaps between elements
Verdict: GPT Image 1.5 produced a much more professional and aesthetically pleasing design with superior typography and texture, though it failed to use a light background. Seedream 5.0 Lite followed the background color prompt but fell short on typographic accuracy and overall design sophistication. GPT Image 1.5 is the preferred choice for a high-quality logo concept.
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
GPT Image 1.5
- + Strict adherence to all six requested steps with accurate iconography.
- + Excellent text rendering for names and steps.
- + Superior visual depth and professional vector styling while maintaining the requested palette.
- − The layout is slightly cramped with overlapping elements in the 'Launch' and 'Earth Orbit' icons.
- − The Saturn V rocket appears a bit chunky compared to the real-life proportions.
Seedream 5.0 Lite
- + Clean, minimalist layout that is very easy to read at a glance.
- + Included a clear 'Apollo 11' title which adds to the poster feel.
- + Properly aligned text throughout the infographic.
- − The Translunar trajectory (Step 3) is just a line and lacks the requested icon/craft detail.
- − The colors are slightly less 'NASA-inspired' and feel more like generic primary tones.
- − Missed the 'Saturn V' specific look, using a generic rocket shape.
Verdict: GPT Image 1.5 is the winner because it followed the step-by-step instructions with much higher fidelity, including specific icons for each phase and accurate supporting details like the crew names. While Seedream 5.0 Lite produced a very clean and readable layout, it failed to provide a meaningful icon for the 'Translunar' step and used more generic graphics overall.
GPT Image 1.5
OpenAI's state-of-the-art image generation model with better instruction following and adherence to prompts
Seedream 5.0 Lite
ByteDance's image generation model with built-in reasoning, example-based editing, and deep domain knowledge, supporting up to 3K resolution