GPT Image 1 OpenAI vs Imagen 4.0 Generate 001 Google

GPT Image 1

• + Excellent adherence to the 'behind' and 'partially visible' instruction for the plant.
• + Realistic glass textures with convincing refractions at the edges.
• + Accurate lighting and shadows consistent with the specified window source.

• − The blue sphere appears slightly too large to be described as 'small'.
• − The sphere appears to be floating without a visible support mechanism.

Imagen 4.0 Generate 001

• + High-quality textures on the red book, including realistic spine details.
• + Good use of reflections within the glass cube surfaces.
• + The sphere is floating centrally, which creates a clean, symmetrical composition.

• − The plant is completely obscured by the cube rather than being 'partially visible through' it.
• − The left face of the cube looks more like a mirror than transparent glass.

GPT Image 1

• + Excellent skin texture and hyper-realistic facial details
• + Perfect representation of cinematic shallow depth of field
• + Natural, moody color grading that matches the 'light rain' setting

• − The bike anatomy is slightly warped near the rear derailleur
• − Background cars lack the requested 'motion blur' effect

Imagen 4.0 Generate 001

• + Strong environment details with bright reflections on wet pavement
• + Good sense of place with the yellow taxi in the background
• + Captures the activity of repairing more dynamically

• − Nonsensical text on the bike frame
• − Skin texture looks slightly processed and less natural than Model A
• − The tool held in the hand is blending into the bike's chain/frame

GPT Image 1

• + Exceptional photographic realism in skin texture and eyes
• + Sophisticated use of lighting and bokeh to create a cinematic atmosphere
• + Natural-looking hair braids with integrated beads

• − Very dark overall, obscuring some details in the armor and underlayers
• − Less focus on the leather straps mentioned in the prompt

Imagen 4.0 Generate 001

• + Clear visibility of ornate leather straps and buckles as requested
• + Stronger direct lighting from the torch source
• + Excellent engraving detail on the pauldrons

• − Skin texture appears somewhat plastic or CGI-like compared to Model A
• − The hair beads look like modern metallic caps rather than integrated beads
• − Visible artifact on the torch flame where it meets the edge

GPT Image 1

• + Excellent font legibility for the primary headers and prices
• + Food photography looks appetizing and consistent in lighting and style
• + Very clear categorization aligned with the prompt's requested sections

• − Placeholder text beneath titles is repetitive and misspelled ('Apperoiation descrigion')
• − The layout cuts off at the bottom, making it feel incomplete as a full menu design

Imagen 4.0 Generate 001

• + Stronger adherence to the 'grid' layout request with geometric design elements
• + Includes all three requested sections (Appetizers, Pizza, Mains) clearly within the frame
• + Better 'modern minimalist' aesthetic with creative use of color accents and whitespace

• − Text rendering is significantly worse with distorted and illegible characters
• − Prices are somewhat ambiguous in currency or scale compared to Model A

GPT Image 1

• + Perfect adherence to the glowing fiery text effect for all elements.
• + Excellent food photography aesthetic with high textural detail on the patty and bun.
• + Clean and impactful composition suitable for a professional advertisement.

• − Incorrect price rendering, showing '.99' instead of '6.99'.
• − The motion feels slightly more static compared to the diagonal explosion of the other model.

Imagen 4.0 Generate 001

• + Accurate text rendering for the price and all requested phrases.
• + Dynamic diagonal composition creates a stronger sense of mid-air motion.
• + Includes more varied ingredients like onions and pickles which add to the visual complexity.

• − The white text lacks the requested 'fiery, glowing effect' specified in the prompt.
• − The food textures appear slightly more like 3D renders than photorealistic photography.
• − The starburst effect looks a bit low-resolution compared to the rest of the image.

GPT Image 1

• + Excellent text rendering with no spelling errors.
• + Realistic chalk texture and authentic handwritten variation.
• + Perfect adherence to the requested menu items and prices.

• − The 'elegant cursive' request for the title was interpreted as simple script rather than flowing cursive.
• − The lighting is a bit dark and monochromatic.

Imagen 4.0 Generate 001

• + Correctly included all requested menu items and prices.
• + Good framing with a wooden chalkboard border.

• − Included meta-commentary like 'Tittle Menu' and 'Footer' into the actual image text.
• − Significant text hallucinations and gibberish between menu items.
• − The handwriting looks more like a digital marker font than authentic chalk.

GPT Image 1

• + Excellent high-detail rendering of the horse and astronaut textures
• + Cinematic lighting and composition

• − Completely failed the semantic inversion in the prompt, placing the astronaut on top
• − Lacks the requested surrealism beyond the basic concept

Imagen 4.0 Generate 001

• + Vibrant, surreal color palette with galaxy textures on the horse
• + High visual quality and clear details

• − Failed the negative constraint to have the horse on top of the astronaut
• − Generative artifacts present in the extra glowing lines and fuzzy hoof contact

GPT Image 1

• + Excellent photorealism and cinematic lighting
• + Natural-looking textures on the capybara's fur and the jacket
• + Perfectly captures the 'bored' expression of the businesswoman

• − The paws are merged with the steering wheel in a physically impossible way
• − The composition is a bit tight, cutting off some of the taxi environment

Imagen 4.0 Generate 001

• + Comprehensive layout showing more of the car interior and Manhattan street
• + Good interpretation of the 'taxi driver cap' with specific TAXI branding
• + Clearer separation of the front and back seat areas

• − The passenger's phone is unnaturally small and the hands are slightly warped
• − The capybara's paws have sharp, protruding talons that look more like a monster than a capybara
• − Lower level of photorealism compared to Model A, appearing more digital

GPT Image 1

• + Perfectly captures the requested 'vintage gothic' style with a muted, classic aesthetic.
• + Excellent composition with a very clear and legible font hierarchy.
• + The parchment texture and framing feel authentic to a real physical invitation.

• − The lighting on the pumpkin is a bit flat compared to the surrounding environment.
• − The border thorns are less pronounced than in the other model.

Imagen 4.0 Generate 001

• + Strong use of 'cinematic lighting' with a glowing blue atmosphere and bright jack-o-lantern.
• + Includes all requested textual elements with creative gothic typography.
• + The thorns and webs in the border are highly detailed and dynamic.

• − The overall style leans more toward a modern digital illustration/cartoon than the requested 'vintage' parchment look.
• − The large vertical scroll element on the left feels unbalanced and out of place.
• − The 'You are invited' banner text uses a generic sans-serif font that clashes with the gothic theme.

GPT Image 1

• + Perfectly follows all text layout instructions including specific words and flag icon position.
• + Highly consistent 3D cartoon art style with soft, clean textures as requested.
• + Excellent adherence to the color palette, specifically the solid light blue background.

• − The 3D clay-like textures, while 'soft', lean more toward stylization than realistic PBR materials.
• − The rice grains are somewhat oversized for a 'miniature' look.

Imagen 4.0 Generate 001

• + Impressive realistic textures and PBR shaders, especially on the salmon roe and tuna.
• + Captures a genuine miniature diorama aesthetic with realistic lighting and shadows.

• − Completely failed to include the requested text ('JAPAN', 'SUSHI') and flag icon.
• − Failed to provide the requested solid light blue background.
• − The 45-degree isometric projection is less precise than Model A's composition.

GPT Image 1

• + Excellent adherence to the 'god rays' and warm golden lighting request.
• + Achieves a high level of photorealistic detail in the fur texture.
• + Natural dynamic composition that feels like a captured action shot.

• − The fox kit has a slightly distorted front paw.
• − The kitten's pose and anatomy look a bit awkward where it meets the meadow.

Imagen 4.0 Generate 001

• + Successfully includes all requested animals clearly in a 'tumbling' interaction.
• + High level of detail in the dew drops on the flowers.
• + Vibrant and colorful interpretation of the wildflower meadow.

• − Has a more digital painting/illustration style rather than 'hyper-photorealistic'.
• − Lighting feels flat and lacks the atmospheric depth of god rays present in the other image.
• − The kitten's anatomy, specifically its midsection and legs, is confusingly rendered.

GPT Image 1

• + Excellent typography with correct accents and vintage serif styling
• + Strong texture that gives a tangible letterpress or stamped feel
• + Accurate interpretation of a vintage banner and cloche

• − Failed the prompt's request for a light background by using solid black
• − Monochromatic approach ignores the cream part of the tone request

Imagen 4.0 Generate 001

• + Perfect adherence to the light background and warm brown/cream color palette
• + Clean, minimalist vector aesthetic that is very professional
• + Accurate spelling and correct inclusion of the banner and steam

• − The 'E' in 'EST' is slightly taller than the other letters in the banner
• − Typography is less 'classic' and more modern/clean compared to Model A

GPT Image 1

• + Excellent adherence to the requested NASA-inspired color palette.
• + Includes all astronauts and specific lunar landing site details.
• + Clean, minimalist flat-vector aesthetic that feels like a cohesive poster.

• − Several spelling errors in the labels (e.g., 'EARLLUNAR').
• − The layout of the steps is somewhat disjointed and non-linear.

Imagen 4.0 Generate 001

• + Near-perfect text rendering with zero spelling errors.
• + Logical diagrammatic flow showing the trajectory from Earth to Moon.
• + Crisp vector graphics with professional iconography.

• − Fails to include all requested steps, missing 'Descent' and 'Landing'.
• − The color palette uses vibrant greens and oranges that deviate from the 'muted' NASA-inspired request.