GPT Image 1.5 OpenAI vs HiDream I1 Full HiDream AI

GPT Image 1.5

• + Excellent realism with sharp textures on the table and book.
• + High-quality glass reflections and refractive effects.
• + Perfect composition with the plant naturally behind the cube.

• − The sphere is quite large, pushing the limits of 'small blue sphere'.

HiDream I1 Full

• + Successfully creates the desired soft window light with shadows.
• + The sphere sizing is more accurate to the 'small' descriptor.

• − The sphere appears to be levitating inside the cube rather than sitting on a surface.
• − Distorted reflections at the edges of the glass cube look unnatural.
• − Overall image clarity is softer and less detailed than its counterpart.

GPT Image 1.5

• + Excellent adherence to the 'imperfect framing' and 'candid' aspects of the prompt.
• + Highly realistic skin textures and wet surfaces with detailed rain droplets.
• + Effective use of motion blur on the background car as requested.

• − The bike's mechanical details are slightly messy and non-functional upon close inspection.

HiDream I1 Full

• + Very clean reflections on the wet pavement.
• + Good central composition and color balance.

• − The man is straddling a tiny, disproportionate bicycle that looks like a toy.
• − Fails to include the requested motion blur on passing cars.
• − The hands and bicycle frame have significant anatomical and structural AI artifacts.

GPT Image 1.5

• + Exceptional photographic realism in skin texture and eyes
• + Highly detailed engraving on the armor and texture on the leather straps
• + Naturally integrated torchlight reflections and bokeh

• − Hair beads are present but less distinct than requested

HiDream I1 Full

• + Strong adherence to the 'hair braided with small beads' request
• + Clear representation of an ornate plate armor design
• + Good use of bokeh and shallow depth of field

• − The facial features and hair look slightly over-sharpened and artificial
• − Skin scars look like markings rather than realistic texture
• − The lighting on the face is a bit flat compared to the metal

GPT Image 1.5

• + Perfect text rendering with legible, meaningful content for all menu items.
• + Follows all prompt instructions including the specified categories (Appetizers/Pizza/Mains) and grid layout.
• + Excellent professional composition that looks like a real, usable design.

• − The food images, while high quality, are slightly generic stock-style photos.

HiDream I1 Full

• + Adheres to the color scheme and minimalist white background request.
• + Includes food photos in a grid-like arrangement.

• − Text is garbled and contains nonsensical gibberish instead of actual menu items.
• − The images only show pizza regardless of the section title, failing on variety.
• − Layout is less professional with awkward spacing and low-quality font rendering.

GPT Image 1.5

• + Excellent adherence to the specific text requested in the prompt.
• + Very realistic chalk texture with dusty smudges and varying pressure.
• + The handwriting looks authentic and consistently styled across the board.

• − The composition is a bit plain, focusing only on the board surface.

HiDream I1 Full

• + Includes a nice frame and coffee cup illustration for cafe ambiance.
• + Text is clear and legible for a digital generation.

• − Failed almost completely on the text content requested, showing nonsensical words.
• − The lettering looks like a digital font rather than authentic chalk handwriting.
• − The prices and menu items do not match the prompt description.

GPT Image 1.5

• + Excellent surface textures and cinematic lighting.
• + High level of detail in the astronaut's suit and moon lander.
• + Balanced composition with interesting background elements like Saturn and Earth.

• − Failed the negative constraint; the astronaut is on top of the horse, not the horse on top of the astronaut.
• − A bit cluttered with asteroids and dust that distract from the main subject.

HiDream I1 Full

• + Clean aesthetic with a clear focus on the subject.
• + Strong sense of speed and motion in the horse's gallop.
• + Correctly placed the subject in an orbit-like environment.

• − Failed the negative constraint; the astronaut is riding the horse.
• − Lower level of fine detail compared to the competitor.
• − The lighting on the horse feels a bit flat compared to the suit.

GPT Image 1.5

• + Excellent cinematic lighting and photorealism that feels like a real movie still.
• + Precise texture on the capybara's fur and the leather steering wheel.
• + Captures the 'bored' expression of the passenger perfectly with natural depth of field.

• − The capybara's paws look slightly more like hands/claws than natural capybara feet.

HiDream I1 Full

• + Good adherence to the requested colors and taxi driver cap style.
• + The composition clearly shows both the driver and the passenger in a wider shot.

• − The lighting is flat and looks more like a digital render than a photorealistic scene.
• − The passenger's hand holding the phone has anatomical irregularities.
• − The capybara's head is looking sideways while his paws are steering straight, creating a disconnected look.

GPT Image 1.5

• + Perfect text rendering of the complex prompt details including the date, time, and location
• + Excellent vintage atmosphere with a cohesive dark parchment color palette
• + Beautiful thorns and web border that feels integrated into the theme

• − The 'Invitation' text is slightly overlapped by the background flourish but remains legible

HiDream I1 Full

• + Clean, readable layout with distinct layers
• + Good adherence to the elements requested like the scroll and jack-o-lantern

• − Significant text hallucinations and errors in the date, time, and location
• − The 'parchment' looks like a clean paper cutout on top of a background rather than an aged poster
• − The overall aesthetic is more like a modern clip-art collage than a vintage gothic poster

GPT Image 1.5

• + Excellent source preservation, keeping the original man's face and background intact.
• + Perfectly followed the edit instruction with a natural-looking head of curly hair.
• + Matches the original lighting and texture of the source image seamlessly.

• − None notable; the edit is near-flawless for the requested task.

HiDream I1 Full

• + Successfully added hair to the back of the head.

• − Failed to preserve the source image, completely changing the man's facial features and the background.
• − Added nonsensical artifacts like the black object floating in front of the beard.
• − The top of the head remains bald, which contradicts the 'full head of hair' instruction.

GPT Image 1.5

• + Excellent PBR material textures, showing realistic wood grain and ceramic glaze.
• + Perfectly followed text and icon instructions, including the flag.
• + Rich, balanced composition with high-quality model details like the teapot and soy sauce bottle.

• − The scene is slightly less 'minimal' than requested due to the extra props.
• − Chopstick placement is a bit awkward across the front edge.

HiDream I1 Full

• + Strong adherence to the '3D cartoon' and 'minimal' aesthetic.
• + Very clean isometric perspective and simple geometry.
• + Accurate text rendering for both 'JAPAN' and 'SUSHI'.

• − Missed the flag icon instruction entirely.
• − Materials look like basic plastics rather than the requested 'realistic PBR materials'.
• − Background has a slight texture/artifacting instead of being a solid blue.

GPT Image 1.5

• + Excellent adherence to all prompts including the TV anchor setting, dogs, and hockey.
• + Successfully captures a true 'caricature' style with exaggerated features while maintaining likeness.
• + High visual quality with clean text rendering and vibrant colors.

• − The hand holding the microphone is slightly distorted with inconsistent finger lengths.

HiDream I1 Full

• + Successfully maintains the caricature style and incorporates a dog.
• + Keeps the denim jacket from the original image while stylizing it.

• − Fails to include any hockey-related elements.
• − The TV anchor profession is only vaguely hinted at by the small screen in the background rather than a main theme.
• − The dog's placement and the character's hands are compositionally awkward.

GPT Image 1.5

• + Excellent adherence to the prompt by including all four distinct animals: puppy, kitten, bunny, and fox.
• + Highly detailed fur textures and realistic anatomy for all subjects.
• + Beautiful lighting with visible god rays and dew sparkles that enhance the atmosphere.

• − The fox kit has a slightly awkward paw placement on the right.

HiDream I1 Full

• + Cutesy, idealized art style with very large, expressive eyes.
• + Clean composition with a clear central focus on the characters.

• − Failed to include the baby bunny, instead providing two kittens.
• − The fox has a disjointed anatomy where the tail appears to be growing from its side or another invisible creature.
• − The '8K masterpiece' level of detail is lacking compared to the competition, feeling more like a standard digital illustration.

GPT Image 1.5

• + Excellent preservation of the original source image's composition and poses.
• + Captures the 'nostalgic' and 'warm' lighting requested in the prompt perfectly.
• + Accurately replicates the character expressions from the original meme in an illustrative style.

• − The style leans more toward generic soft anime/shoujo than distinct Ghibli hand-painted textures.

HiDream I1 Full

• + Stronger adherence to the specific 'Studio Ghibli' art style in terms of character design and line work.
• + Creatively interprets the background with Ghibli-esque town details and signage.

• − Completely fails to preserve the source image's iconic poses and emotional narrative.
• − Lacks the sense of conflict or 'distraction' that defines the original meme.
• − Added unnecessary floral patterns to the clothing that weren't in the source.

GPT Image 1.5

• + Excellent preservation of the subject's face, body, and clothing from the source image.
• + Accurately maintains the background elements including the dog, the bridge, and the pond.
• + Adds motion through swirling hair and flying leaves as requested.

• − The leaf placement looks slightly like a static layer overlay rather than integrated motion.
• − A minor artifact appears in the leash loop which has been reshaped.

HiDream I1 Full

• + Successfully captures an 'energetic and lively' feel by changing the woman's pose to running.
• + Effective lighting and integration of flying leaves.

• − Completely fails to preserve the source image, removing the dog entirely.
• − The subject's facial features and clothing have changed significantly, making it a new image rather than an edit.
• − Anatomy issues with the subject's left hand and right leg.

GPT Image 1.5

• + Excellent typography rendering for both the main title and banner
• + Strong vector emblem aesthetic with sophisticated stippling texture
• + Accurate inclusion of all requested elements including the cloche and banner

• − Failed the 'light background' instruction by using a solid black background

HiDream I1 Full

• + Correctly followed the light background and warm cream tones instruction
• + Clean, minimalist flat vector style
• + Accurate banner text rendering

• − Completely missed the primary text 'Caffè Florian' from the logo
• − The cloche dome is stylized in a way that looks more like a window or bell

GPT Image 1.5

• + Excellent adherence to the 6-step chronological structure requested.
• + Clear, legible text for all mission phases.
• + Clean and consistent flat-vector style with a great NASA-inspired palette.

• − The Saturn V rocket in the 'Launch' section is slightly distorted and lacks verticality.
• − Minor overlap of objects in the Lunar Orbit panel.

HiDream I1 Full

• + Strong heading typography for the title.
• + Good use of negative space for a modern poster feel.

• − Failed to include all 6 requested steps, stopping after 3-4 phases.
• − Included literal prompt text like '+ ICON' and 'UNLY' in the final image.
• − Incorrect iconography, such as a Saturn-like ringed planet for Earth Orbit.