GPT Image 1 OpenAI vs GPT Image 1.5 OpenAI

GPT Image 1

• + Excellent material rendering with realistic glass thickness and refraction
• + Clean and high-resolution textures on the sphere and book
• + Precisely following the spatial prompt regarding the plant being behind the cube

• − The sphere appears to be floating slightly above the bottom surface of the cube

GPT Image 1.5

• + Beautiful lighting and reflections on the sphere's surface
• + Realistic integration of the sphere sitting on the base including a reflection
• + Good adherence to the requested layout

• − The glass cube is missing its top plane; the book appears to be resting directly on the side walls
• − Lower resolution/clarity in the plant leaves compared to Image A

GPT Image 1

• + Excellent preservation of the car's design including the grill and badge.
• + Successful integration of the man's likeness and hairstyle into the driver seat.
• + Great sense of motion and environment with realistic lighting and motion blur on the wheels.

• − The man appears to be looking down rather than at the road.

GPT Image 1.5

• + Natural facial expression and pose for a driving scenario.
• + The background and scenery are very convincing for a coastal drive.
• + Good preservation of the man's facial features and distinctive hair.

• − Crop is a bit tight, losing the front of the car which was a focus of the source image.
• − The hand on the steering wheel has some anatomical distortion.

GPT Image 1

• + Excellent skin texture and hyper-realistic facial details
• + Strong '50mm lens' look with soft bokeh
• + Great atmospheric handling of rain droplets on the bike and clothing

• − Anatomy and perspective issues where the man's hand meets the bicycle chain/gears
• − The bicycle design is overly simplified and structurally illogical in the rear frame area

GPT Image 1.5

• + More complex and realistic bicycle mechanics including a chain, gears, and a tool tray
• + Better storytelling with the inclusion of the tool tray and secondary umbrellas in the background
• + Dynamic composition with visible rain streaks

• − The facial texture is slightly smoother and less detailed compared to Model A
• − The car in the background looks a bit static despite the prompt for motion blur

GPT Image 1

• + Exceptional engraved metal texture with realistic patina
• + Mature and weary facial expression fits the 'battle-worn' prompt perfectly
• + Natural skin texture and subtle color grading

• − Lighting is somewhat flat compared to Model B
• − Fails to show leather straps requested in the prompt

GPT Image 1.5

• + Includes all elements like beads, leather straps, and cross iconography
• + Dynamic lighting with strong torchlight reflections and bokeh sparks
• + High contrast and sharp details on the hair and fabric

• − Skin texture appears slightly airbrushed or overly smooth under the dirt
• − Scars look a bit tropically placed rather than natural wear

GPT Image 1

• + Excellent photographic quality and lighting of food items
• + Bold and clean typography that adheres well to a modern minimalist style

• − Nonsensical placeholder text with frequent typos like 'Apperoiation descrigion'
• − Missing the 'Mains' category header despite showing mains in the grid

GPT Image 1.5

• + Includes all requested text categories: appetizers, pizza, and mains
• + Legible, realistic menu item names and descriptions
• + Professional grid-based layout that feels like a real commercial menu

• − Image resolution/sharpness is slightly lower than Model A
• − Text alignment on the bottom section is slightly cramped

GPT Image 1

• + Excellent typography rendering with clean, glowing neon effects.
• + Very clean composition with minimal clutter, making the food items stand out.
• + Good photorealistic texture on the bun and patty.

• − Failed to include the price correctly, displaying '.99' instead of '6.99'.
• − Less 'explosive' energy compared to the other image, feeling more static.

GPT Image 1.5

• + Successfully rendered all text prompts including the correct €6.99 price.
• + Highly dynamic composition with a strong sense of motion and chaotic energy.
• + Rich detail with added ingredients like onions and cascading sauce drips.

• − The background is a bit busy, which slightly distracts from the central product.
• − The text 'MAGIC BURGER' is partially cropped at the very top of the frame.

GPT Image 1

• + Excellent text legibility throughout the entire menu.
• + Perfect spelling and numerical accuracy for all requested items.
• + Uniform chalk texture that feels consistent across the entire board.

• − The handwriting style is a bit too uniform, bordering on a digital font look.
• − Fails the request for 'elegant cursive' for the title, using a simple print instead.

GPT Image 1.5

• + Successfully uses elegant cursive for the title as requested.
• + Handwriting has a more authentic, variable feel with realistic chalk dust and smudging on the board.
• + Captures the 'cozy café' aesthetic more effectively through board texture.

• − The cursive handwriting is much harder to read compared to Model A.
• − The 'Grilled Octopus' line has a slight visual smudge that makes the text look a bit messy.

GPT Image 1

• + Successfully replicates the complex leg-crossing pose from Image 1.
• + Accurately represents the scarf and glasses from Image 2.
• + Matches the yellow background and lighting of the source image.

• − The face has significant distortions and does not closely resemble the character in Image 2.
• − Notable anatomical issues with the feet and hands.

GPT Image 1.5

• + Excellent character resemblance, keeping the face and hair almost identical to Image 2.
• + Accurately replicates the pose while maintaining better anatomical proportions.
• + Includes clothing details like the text on the sweatshirt from the character reference.

• − The scarf is slightly merged with the sweater fabric in some areas.
• − One foot has an extra toe/misshapen structure.

GPT Image 1

• + Successfully follows the surreal instruction of a horse riding an astronaut.
• + Clean and cinematic lighting with high contrast.
• + Good anatomical rendering of the horse despite the unusual composition.

• − The astronaut's legs and lower body are awkwardly integrated with the horse's back.
• − Slightly muddy details in the dark space background.

GPT Image 1.5

• + Very high detail in the lunar surface and nebula background.
• + Dynamic action with the dust and lunar lander adding to the scene.

• − Completely failed the negative constraint/spatial instruction: shows an astronaut riding a horse instead of a horse on top.
• − Overly busy composition with many clashing celestial elements.

GPT Image 1

• + Excellent preservation of the person's face, skin markings, and facial features.
• + Accurately replicates the coat texture and the specific plaid pattern of the scarf.
• + Maintains the background lighting and grain of the source image well.

• − The scarf is missing the fringe tassels visible in the target outfit.
• − The watch chosen (leather strap) does not match the gold metal link watch from the target image.

GPT Image 1.5

• + Includes the fringe tassels on the scarf and the gold link watch from the target outfit.
• + Captures the full-body pose including the jeans and leg positioning reasonably well.

• − Crops out the top half of the person's face, failing the instruction to keep the person's exact face unchanged.
• − The skin marking on the chin does not match the original person's vitiligo pattern.
• − Missing the sand on the skin that was present in the source image.

GPT Image 1

• + Excellent fur texture and photographic lighting integration
• + Captures the bored expression of the passenger perfectly
• + The taxi cap has a high-quality leather/fabric texture

• − The capybara's right paw is gripping the wheel in a slightly distorted, hand-like way
• − The interior feels a bit dark and lacks specific taxi equipment

GPT Image 1.5

• + The paws are more anatomically correct for a capybara while still gripping the wheel
• + Includes realistic taxi details like the meter on the dashboard
• + The composition provides a wider, more authentic view of the car's interior

• − The passenger's face is slightly less detailed and more blurred than Model A
• − The capybara's hat has a slightly more 'clip-art' feel to the logo compared to the rest of the image

GPT Image 1

• + Excellent typography rendering with zero spelling errors.
• + Clean and legible layout suitable for an actual invitation.
• + The lighting on the jack-o-lantern and ground is subtle and cinematic.

• − Failed to include the specific '7pm' time detail.
• − The location is incorrectly labeled as 'TIME: The Arches, NYC'.

GPT Image 1.5

• + Accurately included all requested text details including the date, time, and location.
• + High-quality 'vintage' aesthetic with a detailed thorny border and parchment texture.
• + Dynamic background featuring a graveyard and castle silhouette that adds to the gothic theme.

• − The 'frights' text on the scroll has some minor tailing/artifacting at the end.
• − The layout feels more cluttered than Model A due to the high amount of texture.

GPT Image 1

• + Excellent preservation of the jacket and background
• + Adds a very thick head of hair as requested

• − The hairline is unnaturally low, obstructing part of the forehead and distorting the eye area
• − Noticeable texture mismatch between the added hair and the original beard

GPT Image 1.5

• + Natural-looking curly hair texture that blends perfectly with the existing beard
• + Realistic hairline and forehead preservation
• + Maintains original facial structure and eye shape

• − Slightly changes the bridge of the nose and the frames of the glasses compared to the original

GPT Image 1

• + Excellent adherence to the 'cartoon scene' style with soft, rounded shapes.
• + Extremely clean text rendering and layout.
• + Strong execution of the 45-degree isometric perspective and diorama base.

• − Very limited variety of sushi compared to the prompt's potential.
• − Textures are perhaps a bit too simplified/plasticky for 'realistic PBR'.

GPT Image 1.5

• + Superb material quality with realistic PBR textures on the wood, ceramic, and fish.
• + Includes a wider variety of sushi items and accessories like the teapot and soy sauce.
• + Accurate text rendering and placement.

• − A bit cluttered compared to the 'minimal garnish' and 'small raised diorama' request.
• − The 'cartoon' aspect is less pronounced than in Image A.

GPT Image 1

• + Excellent realization of the caricature art style with hand-drawn watercolor aesthetics.
• + Great source preservation by keeping the subject's denim jacket from the original photo.
• + Creative integration of the theme by having a dog as the hockey player on the TV screen.

• − The facial exaggeration is quite extreme, bordering on creepy.
• − The background is very simple and less 'news-like' compared to the other model.

GPT Image 1.5

• + Successfully incorporates all elements into a busy, humorous scene including a dog in a hockey helmet.
• + High level of detail in the studio setting with cameras, lighting, and news tickers.
• + The caricature face maintains a better likeness to the source person despite the exaggeration.

• − Changed the subject's clothing from the original denim to a red dress.
• − The hand holding the microphone has anatomical issues with the thumb and finger placement.

GPT Image 1

• + Excellent dynamic composition with a clear sense of movement and 'tumbling'.
• + Consistent lighting and beautiful god rays that create a cinematic atmosphere.
• + Distinct separation between the four animals, making Each easy to identify.

• − The fox kit has unnatural dark brown/black paws that look like gloves.
• − The kitten's eye anatomy is slightly distorted compared to real cats.

GPT Image 1.5

• + Highly detailed fur texture and adorable 'toe beans' visible on the kitten.
• + Excellent 'dew sparkles' effect and warm golden sunrise lighting.
• + Captures a very charming, cuddly interaction between the animals.

• − More static composition compared to Model A; less of a 'chasing' feel.
• − The bunny looks somewhat squashed and less realistic between the other animals.
• − The kitten has an extra toe or claw-like protrusion on its raised paw.

GPT Image 1

• + Expertly captures the hand-painted, textured aesthetic of traditional Ghibli backgrounds.
• + Maintains the exact compositional balance and poses of the source image.
• + The facial expressions perfectly translate the 'distracted boyfriend' meme into an anime style.

• − The colors are a bit muted compared to some of Ghibli's more vibrant outdoor scenes.

GPT Image 1.5

• + Excellent 'dreamy' lighting effect with a beautiful sunlight glow.
• + High visual quality with clean lines and soft gradients.
• + Preserves the plaid pattern on the shirt accurately.

• − The style leans closer to modern digital anime/shoujo than the specific hand-painted Ghibli aesthetic.
• − The woman in the red dress has a somewhat generic face compared to Model A.

GPT Image 1

• + Excellent preservation of the woman's face and original features
• + Hair movement looks very natural and follows the direction of the wind well
• + Dog's tail is modified to show realistic movement

• − Flying leaves are small and clumped, looking almost like debris
• − Overall brightness and saturation decreased noticeably from the source

GPT Image 1.5

• + Successfully added a large number of colorful flying leaves
• + Hair movement is dynamic and symmetrical
• + Maintains clarity and color saturation closer to the source image

• − The woman's face has changed slightly, losing some of the likeness to the original
• − Hair strands look a bit more artificial and 'pasted on' compared to the source

GPT Image 1

• + Clean minimalist design that fits modern vector standards
• + Precise typography and accurate accent on 'Caffè'
• + Superior texture consistency across the graphic

• − Single color choice is slightly less dynamic than requested 'brown and cream' palette
• − Failed the background requirement by using a black background instead of a light one

GPT Image 1.5

• + Excellent use of warm brown and cream tones as requested
• + Detailed vector emblem style with sophisticated shading on the cloche
• + Captured the 'retro' feel more effectively with the banner design

• − Failed the background requirement by using a black background instead of a light one
• − Typography for 'Caffè' is slightly less legible than Model A

GPT Image 1

• + Excellent adherence to the 'navy, white, muted red' color palette
• + Clean flat-vector style with professional-looking silhouettes
• + Accurate spelling of the crew names

• − Confusing layout where labels do not clearly align with their respective icons
• − Nonsense word 'EARLLUNAR' replaces 'Lunar Orbit' in the footer area
• − Lacks the specific steps/numbers requested in a logical sequence

GPT Image 1.5

• + Highly organized grid layout that clearly follows the requested 6-step sequence
• + Strong visual narrative with distinct panels for each phase of the mission
• + Perfect spelling for all main phase labels and crew names

• − Introduced green on Earth which was not in the 'NASA-inspired' specific palette
• − Minor spelling deviation in 'Tranquillity' (double 'L' is British variant, but 'Tranquility' was in prompt)