GPT Image 1 OpenAI vs GPT Image 2 OpenAI

GPT Image 1

• + Excellent refraction and lighting through the glass cube.
• + The blue sphere has a pleasing matte texture.
• + Clear adherence to the soft left-side lighting requested.

• − The glass cube has internal glass structural lines that make it look more like a frame than a solid cube.
• − The perspective of the cube base feels slightly tilted relative to the tabletop.

GPT Image 2

• + Natural glass transparency and realistic reflections on the cube surfaces.
• + High-quality texture on the red book cover.
• + Solid composition with realistic plant visibility through the glass.

• − The blue sphere's shadow on the bottom glass panel is slightly inconsistent with the light source.
• − The glass walls look a bit like double-paned glass rather than a single solid sheet.

GPT Image 1

• + Excellent skin texture and moody photographic lighting.
• + Strong bokeh and rain effects that create a cinematic atmosphere.
• + Composition flows well with the subject’s posture and focal point.

• − The red bicycle frame has some physical inconsistencies and nonsensical parts.
• − The motion blur of the cars in the background looks more like static bokeh than motion.

GPT Image 2

• + Successfully captures the requested 'imperfect framing' and candid street style.
• + Includes realistic Japanese text and urban background elements for authenticity.
• + Better representation of 'motion blur' on the passing vehicles.

• − The lighting is a bit flat compared to the 'cinematic' request.
• − Skin texture on the subject is slightly smoother and less detailed than Model A.

GPT Image 1

• + Excellent depiction of ornate engraved plate armor
• + Authentic battle-worn expression and grim skin textures
• + Dramatic use of warm torchlight and high-contrast shadows

• − Lighting is very dark, obscuring details on the right side
• − The braided hair beads are less distinct

GPT Image 2

• + Exceptional detail on leather straps and cloth underlayer as requested
• + Lifelike eyes and very clear skin textures with faint scars
• + Perfect execution of the braided hair with small beads

• − Lighting feels more like soft sunlight than localized torchlight
• − Overall composition is slightly less 'gritty' than the prompt implies

GPT Image 1

• + Excellent food photography quality with high-resolution details
• + Clean, minimalist layout that feels modern and spacious

• − Nonsense filler text with frequent spelling errors like 'descridion'
• − Weak composition that cuts off the bottom half of the menu

GPT Image 2

• + Outstanding professional layout with logical sections for Appetizers, Pizza, and Mains
• + Extremely readable text with realistic item names and descriptions
• + Complete branding including a logo, social media icons, and dietary tags

• − Some minor food proportions look a bit small compared to the text blocks

GPT Image 1

• + Excellent centering and clean layout for a professional ad feel.
• + Accurate rendering of the requested text phrases.
• + High photorealism on the vegetable textures and burger bun.

• − Missed the first digit of the price, displaying '.99' instead of '6.99'.
• − The motion is quite static compared to the 'dynamic' request.

GPT Image 2

• + Perfect adherence to all text requirements including the specific '€6.99' price.
• + Highly dynamic composition with sauce splashes and tilted angles that convey motion.
• + Superior 'fiery' textures on the text and background elements.

• − The composition is slightly more cluttered than Model A.
• − Some sauce drips defy physics in a way that slightly lowers photorealism.

GPT Image 1

• + Excellent text legibility across all sections
• + Consistent chalk texture throughout the board
• + Accurate rendering of specific menu prices and items

• − The handwriting style is too uniform and lacks natural cursive variation in the title
• − The layout is very tight to the edges, feeling less like a natural cafe scene

GPT Image 2

• + Successfully followed the requirement for elegant cursive for the title
• + Great atmospheric lighting and composition that feels like a real café setting
• + Captures natural variations in letter size and slant more effectively

• − The text is slightly less crisp and clear compared to Image A
• − Includes a decorative underline flourish not specifically requested

GPT Image 1

• + Matches the yellow background and red ottoman from the source image
• + Correctly applies the black clothing and scarf from the second image

• − The facial likeness is poor and looks somewhat distorted
• − The hand anatomy on the left is severely deformed and simplified
• − Fails to capture the tilt of the head from the pose reference correctly

GPT Image 2

• + Excellent facial likeness and hair matching from the character reference
• + Perfectly replicates the complex cross-legged pose and arm angles from Image 1
• + Successfully integrates the scarf and sunglasses while maintaining natural-looking skin tones

• − The scarf's length and hanging position are slightly unrealistic given the extreme tilt of the body

GPT Image 1

• + Excellent cinematic lighting and atmosphere
• + Realistic textures on the spacesuit and horse's coat
• + High visual quality with a balanced composition

• − Failed the primary prompt instruction of having the horse on top

GPT Image 2

• + Adhered perfectly to the unusual prompt instruction of having the horse on top
• + Correctly interpreted the surreal nature of the request
• + Impressive rendering of the NASA logo and suit details

• − The transition between the horse's legs and the astronaut is anatomically messy
• − The harness/bridle logic is a bit confused in the center

GPT Image 1

• + Excellent fabric texture and realistic lighting on the pea coat.
• + Successfully captures the watch and plaid pattern from the reference image.
• + Maintains a high level of facial detail and similarity to the original person.

• − The position of the hands in the pockets resulted in some slight anatomical awkwardness in the right arm.
• − Cropped the original image significantly, losing some of the background context.

GPT Image 2

• + Near-perfect preservation of the original background and framing.
• + Very accurate replication of the specific plaid scarf and gold watch from Image 2.
• + Maintains the exact vitiligo patterns on the face and the specific hair styling of the original subject.

• − The transition between the neck and the collar of the coat is slightly rough with some minor artifacts.
• − The left hand in the pocket has a slightly blurred, less defined appearance compared to the rest of the image.

GPT Image 1

• + Excellent fur texture and photorealism on the capybara.
• + The capybara's paws are positioned very naturally on the steering wheel.
• + Very clean rendering of the 'TAXI' text on the cap.

• − The composition feels slightly tighter, cutting off more of the car door and environment.

GPT Image 2

• + Good use of local NYC branding cues like the 'NYC' circle logo on the hat.
• + Dynamic angle that shows more of the car's exterior and the city environment.
• + Captures the 'bored' expression of the passenger very effectively.

• − The left paw (from our perspective) has a slight anatomical clipping issue with the steering wheel.
• − The texture of the capybara's snout is slightly more 'painterly' compared to the high realism of Model A.

GPT Image 1

• + Excellent typography for the main title and banner scroll
• + Clean, cinematic lighting that draws focus to the jack-o-lantern

• − Confused the event details, merging the location into the 'Time' field and omitting the 7pm text

GPT Image 2

• + Includes all specific event details correctly (Date, Time, Location)
• + Rich, intricate border design that adheres perfectly to the thorns and webs prompt
• + Creative inclusion of 'The Arches' and NYC skyline in the background

• − The parchment texture is a bit busy, making some fine text slightly harder to read against the background

GPT Image 1

• + Excellent adherence to the 'minimal' garnish and plate request
• + Perfect execution of soft, 3D cartoon textures
• + Clean and readable typography and flag icon

• − Simple composition might feel sparse to some users

GPT Image 2

• + Detailed diorama base with stone textures and miniature pagoda
• + Professional 3D graphic design for the text header
• + Vibrant colors and high-quality material rendering on the seafood

• − Ignored the 'minimal garnish' instruction, filling the scene with many elements
• − The chopsticks are incorrectly rendered being partially embedded in the base

GPT Image 1

• + Excellent character interaction with the cat's paw reaching out toward the bunny.
• + Warm, cohesive lighting with very prominent god rays.
• + Clean, minimalist background that keeps focus on the animals.

• − The fox's front right leg has an anatomically awkward, dark, stump-like appearance.
• − The background butterfly in the top left is slightly less integrated into the lighting.

GPT Image 2

• + Dynamic composition with a more varied and colorful wildflower meadow.
• + Superior detail on the individual floral elements and grass blades.
• + The animals display more distinct action poses, particularly the pouncing fox.

• − The fox exhibits anatomical issues, with a strange tail/leg hybrid visible behind it.
• − The kitten's raised paw features an irregular number of toes/claws.

GPT Image 1

• + Clean vector-style icon suitable for modern branding
• + Accurate spelling of the name including the accent
• + Good balance between the cloche and the typography

• − Ignores the light background request, using a black background instead
• − Minimalist style is a bit too simple, lacking the requested vintage texture

GPT Image 2

• + Excellent adherence to the vintage aesthetic and light textured background
• + High-quality typography with period-accurate flourishes
• + Superior visual appeal with detailed hatching and framing

• − The 'E' in 'Caffè' is stylized in a way that slightly degrades legibility compared to Model A

GPT Image 1

• + Strong minimalist flat-vector aesthetic
• + Follows the muted red and navy color palette effectively
• + Clean typography for crew names

• − Poor layout with disorganized text labels that don't align with icons
• − Spelling errors like 'EARLLUNAR'
• − Missing the Lunar Orbit step entirely

GPT Image 2

• + Perfect adherence to all 6 requested steps in a logical sequence
• + Excellent graphic design composition with clear sections and headers
• + Near-perfect text rendering and iconography

• − The detail level is slightly higher than 'flat-vector', leaning into more complex illustrations
• − Minor artifacts in the NASA meatball logo