GPT Image 1 Mini OpenAI vs HiDream I1 Full HiDream AI

GPT Image 1 Mini

• + Excellent adherence to lighting instructions
• + Highly realistic glass and book textures
• + Precise placement of the blue sphere on the bottom of the cube

• − The plant is more to the side than directly behind the cube

HiDream I1 Full

• + Successfully placed the plant behind the glass cube
• + Interesting wood grain on the table

• − The blue sphere is floating unnaturally
• − Strong optical distortions and ghosting/reflections of the sphere that look like extra objects
• − Texture of the book looks less like an actual physical object

GPT Image 1 Mini

• + Excellent skin texture and realistic facial details
• + Accurate shallow depth of field and bokeh consistent with a 50mm lens
• + Effective placement of beads of water on the bicycle and jacket

• − The 'motion blur from passing cars' is present but very subtle in the background
• − The hand anatomy is slightly lumpy upon close inspection

HiDream I1 Full

• + Strong reflection on the wet pavement
• + Good sense of place with the blurred Japanese street background

• − Anatomy and scale issues: the man appears to be sitting on a miniature bicycle that is missing its rear wheel
• − The skin texture lacks the 'natural' and 'non-stylized' qualities requested, looking oily or plastic
• − The bicycle design is nonsensical with a frame that merges into the man's leg

GPT Image 1 Mini

• + Excellent textural detail on the engraved plate armor and leather strap
• + Coherent and natural lighting that realistically interacts with the skin and metal textures
• + Superior facial anatomy and a subtle, believable interpretation of battle-weariness

• − The braids are very subtle and lack the requested beads
• − Lower contrast colors compared to the other model

HiDream I1 Full

• + Strict adherence to hair details including the braids and beads along the crown
• + Vibrant torchlight effects and highly visible bokeh sparks
• + Dramatic eye color and sharp facial features

• − The scars look like digital paint strokes rather than skin texture
• − The armor engraving and leather straps look less realistic and more like plastic modeling
• − Proportions of the braids and facial symmetry feel slightly unnatural

GPT Image 1 Mini

• + Strict adherence to text labels with perfect spelling for all requested sections.
• + Very clean, minimalist grid layout that looks institutional and professional.
• + Excellent variety in food imagery, showing distinct pizzas, appetizers, and mains.

• − Lacks actual menu item text or descriptions, leaving large empty spaces.
• − The food photos are slightly repetitive in their top-down angle.

HiDream I1 Full

• + Includes menu item names and pricing, providing a more functional design layout.
• + High-quality, appetizing food photography with good lighting.
• + Uses a hierarchy of font weights to distinguish headers from descriptions.

• − Text rendering is gibberish/hallucinated for specific menu items.
• − Visual variety is poor as all four food photos are essentially the same pizza dish.
• − Combines Appetizers and Pizza into a single section header, unlike the prompt request.

GPT Image 1 Mini

• + Excellent text rendering with no spelling errors.
• + Realistic chalk texture and natural variations in the lettering.
• + Accurately followed the specific menu items and prices requested.

• − The title was requested in elegant cursive but is rendered in print capitals.
• − The composition is a bit tight at the bottom edge.

HiDream I1 Full

• + Followed the request for cursive handwriting for the title.
• + The chalk coffee illustration adds to the cafe aesthetic.

• − Failed significantly on text accuracy, resulting in illegible 'gibberish' words.
• − Ignored the specific list of menu items provided in the prompt.
• − The text style looks more like a digital font than natural chalk handwriting.

GPT Image 1 Mini

• + Excellent texture on the astronaut suit and horse hair.
• + Balanced cinematic composition with the moon providing a secondary focal point.
• + High level of anatomical detail on the horse's musculature.

• − Completely failed the negative constraint to put the horse on top.
• − Color palette is a bit monotonous/muted.

HiDream I1 Full

• + Dynamic lighting with vibrant highlights on the horse and suit.
• + Includes various celestial elements like planets and nebulae for a more 'space' feel.
• + Clean image clarity without much grain.

• − Failed the negative constraint to place the horse on top of the astronaut.
• − Noticeable artifacts such as the floating bridle straps and extra horse legs or hooves in the background blurred area.

GPT Image 1 Mini

• + Atmospheric cinematic lighting that truly feels like a taxi interior at night.
• + Highly detailed fur texture and realistic passenger expression.
• + The check-patterned hat provides a more authentic NYC taxi driver aesthetic.

• − One of the capybara's paws is poorly defined against the steering wheel.
• − The passenger's hands look slightly unnatural and blurry.

HiDream I1 Full

• + Excellent adherence to the request for both front paws on the steering wheel.
• + Very clean and sharp rendering of the capybara and the taxi cab framing.
• + Legible text on the driver's cap adds to the prompt adherence.

• − The lighting is too bright and even for a 'night' scene, lacking the moodiness of the prompt.
• − The passenger's hand holding the phone has significant anatomical errors (too many fingers/segments).
• − The composition feels a bit cramped with the woman appearing to sit next to the driver rather than in the back seat.

GPT Image 1 Mini

• + Excellent typography with perfect spelling and placement.
• + Authentic vintage parchment texture and moody cinematic lighting.
• + Fully adheres to all text and layout requirements of the prompt.

• − The thorns in the border are a bit subtle and hard to see against the dark background.

HiDream I1 Full

• + Strong contrast between the parchment and the moody night sky background.
• + Good visual iconography including the bats and the jack-o-lantern.

• − Failed significantly on text rendering, resulting in gibberish like 'Yootmber' and '30 cm'.
• − Missing the specific year and location requested in the bottom details.

GPT Image 1 Mini

• + Successfully added a thick, full head of hair that matches the original color and lighting.
• + Preserved the composition, foreground clothing, and overall background scenery effectively.
• + Maintained the facial features of the original person reasonably well during the hair addition.

HiDream I1 Full

• + Enhanced the skin texture and beard details to be very sharp.

• − Completely failed the main editing instruction by keeping the top of the head bald.
• − Failed to preserve the background, changing a desert landscape into a plowed field with windmills.
• − Introduced severe artifacts, including a bizarre black object merging with the face and beard.

GPT Image 1 Mini

• + Excellent text rendering and includes the requested flag icon.
• + Very clean 3D isometric execution with high clarity.
• + Perfectly follows the color and background specifications.

• − The 'diorama base' is a bit simple compared to the platform in model b.

HiDream I1 Full

• + Features a more detailed diorama base with legs.
• + Provides a wider variety of sushi types within the scene.

• − Failed to include the requested flag icon.
• − Text for 'SUSHI' is significantly smaller and less balanced than the prompt's layout.
• − Minor rendering artifacts on the left edge of the image.

GPT Image 1 Mini

• + Successfully incorporates all requested elements: tv news desk, dog, and hockey equipment.
• + Follows the 'caricature' style perfectly with exaggerated features and hand-drawn texture.
• + Preserves the denim jacket from the source image.

• − The hand holding the microphone has some anatomical sketching artifacts/extra fingers.

HiDream I1 Full

• + Maintains a strong facial resemblance to the subject in the source image.
• + High-quality digital illustration style with clean lines.

• − Completely misses the 'hockey' requirement of the prompt.
• − The 'tv show anchor' element is very weak, only vaguely implied by the background monitor.
• − The dog is rendered at an awkward, tiny scale on the subject's shoulder.

GPT Image 1 Mini

• + Excellent adherence to the prompt, including all four specific animal types (dog, cat, bunny, fox).
• + High visual quality with realistic fur textures and convincing dynamic movement.
• + Effective atmospheric lighting with visible god rays and dew sparkles.

• − The bunny's scale relative to the other animals is slightly small, though within the realm of realistic variation.

HiDream I1 Full

• + Warm and vibrant color palette with a clear sunrise effect.
• + Cute, expressive eyes as requested in the prompt.

• − Missing the baby bunny entirely, failing on prompt adherence.
• − The fox's neck and body proportions appear slightly distorted.
• − The animals are mostly static rather than 'tumbling and chasing' as requested.

GPT Image 1 Mini

• + Excellent source preservation, maintaining the composition and poses of the original meme perfectly.
• + The soft pastel texture and colored pencil aesthetic match the requested hand-painted feel.
• + Capture the specific expressions of the subjects with great accuracy in the new style.

HiDream I1 Full

• + Successfully captures a more modern anime/Ghibli-inspired clean art style.
• + Intelligently adds thematic details like Japanese signage and floral patterns.

• − Poor source preservation as it changes the man's pose and replaces his glance with a smile, losing the context of the meme.
• − Alters the subjects' appearances too drastically, including adding facial hair and hats not present in the source.

GPT Image 1 Mini

• + Excellent source preservation, maintaining the woman's face and the dog from the original.
• + Successfully added wind-blown hair and falling leaves while keeping the composition stable.
• + High visual quality with natural lighting integration for the new elements.

• − The wind direction in the hair is slightly inconsistent with the static nature of the clothing.

HiDream I1 Full

• + Strong sense of movement and 'energetic' feel as requested.
• + Beautifully rendered wind-blown hair and dynamic leaf placement.

• − Completely failed to preserve the source image; the dog is missing and the woman's face has changed.
• − The pose was changed from a walk to a skip/run, losing the original context.
• − Visual artifacts present on the hands and shoes.

GPT Image 1 Mini

• + Excellent typography with perfect spelling of the name and date.
• + Accurately represents the cloche with steam as requested.
• + Strong vintage aesthetic with wood-cut style texturing.

• − Failed to provide a light background as specifically requested.
• − The layout is a bit tall and heavy for a minimalist logo.

HiDream I1 Full

• + Followed the background color and tone instructions perfectly.
• + Clean, vector-style minimalist illustration.
• + Good use of the banner element for the date.

• − Completely omitted the primary commercial text 'Caffè Florian'.
• − The cloche dome is stylized in a way that looks more like a coffee cup inside a bell than a standard cloche.
• − Typography on the banner is slightly generic.

GPT Image 1 Mini

• + Successfully followed all 6 requested steps in the correct order.
• + Text and numbering are perfectly legible and accurate.
• + Excellent flat-vector style with consistent line weights and iconography.

• − The 'Translunar' trajectory icon is an abstract loop rather than a clear path between celestial bodies.
• − The crop at the bottom cuts off the crew and footer text.

HiDream I1 Full

• + Attractive color palette and bold typography for the title.
• + Good use of space and symmetry in the center graphic.

• − Failed to include all 6 prompt steps, stopping after step 3.
• − Contains nonsensical text, including prompt fragments like '+ UNG' and '+ ICON'.
• − Includes an incorrect planet depiction that looks more like Saturn than Earth or the Moon.