GPT Image 1 OpenAI vs Z-Image Turbo Alibaba

GPT Image 1

• + Excellent adherence to the 'soft window light from the left' instruction.
• + High visual clarity and realistic textures on the sphere and book.
• + Strong composition with a clearly defined plant visible through the glass.

• − The sphere appears slightly large for the description of 'small'.
• − The bottom of the glass cube has a heavy metallic base not explicitly requested.

Z-Image Turbo

• + Accurately depicts a 'small' blue sphere as requested.
• + The glass cube looks more like a standard hollow glass vessel.
• + Good surface reflections on the wooden table.

• − The plant is extremely blurry and barely recognizable in the background.
• − The lighting is somewhat flat and lacks the clear directional quality requested.
• − There are some minor artifacts on the edges of the book.

GPT Image 1

• + Excellent shallow depth of field and bokeh effects
• + Highly realistic skin textures and fine details on the bicycle and clothing
• + Strong cinematic lighting and atmospheric mood with wet pavement reflections

• − The 'imperfect framing' requested is somewhat subtle as the composition is quite balanced
• − Lacks significant motion blur for the passing cars

Z-Image Turbo

• + Successfully captures an older Japanese man with a red bicycle
• + Good rendering of the wet pavement and rain streaks

• − The man is pushing or standing by the bike rather than repairing it
• − Composition lacks the shallow depth of field and 'cinematic' feel requested
• − Anatomical issues with the right hand merging into the handlebars

GPT Image 1

• + Excellent typography rendering with almost perfect spelling.
• + High-quality, appetizing food photography that looks professional.
• + Clean and balanced layout that adheres strictly to the minimalist prompt.

• − The grid structure is a bit basic, feeling more like a cropped page than a full menu sheet.
• − Missing the 'Mains' category header despite showing a main course image.

Z-Image Turbo

• + Strong grid layout with a high volume of food images.
• + Bold use of vibrant orange accents as requested.
• + Effective use of negative space around the menu borders.

• − Significant spelling errors and garbled text (e.g., 'PIZZA MANS', 'SE IIION').
• − Food photography is slightly less realistic with some AI artifacts in the pizza toppings.
• − Composition is a bit cluttered compared to the minimalist request.

GPT Image 1

• + Excellent adherence to the 'exploded' concept with space between all ingredients.
• + High-quality photorealistic food textures, especially the lettuce and bun.
• + Text is perfectly integrated with the requested ember/fiery glow effect.

• − The price in the starburst is missing the '6' digit, reading only '€ .99'.
• − The composition is a bit tight at the top and bottom edges.

Z-Image Turbo

• + Clean, readable text that includes all requested characters correctly.
• + Dynamic environment with physical fire and depth-of-field effects.
• + Vibrant colors and a polished commercial aesthetic.

• − Failed the 'exploded' instruction as the ingredients are mostly stacked normally.
• − The burger looks more 'tilted' than 'suspended' or 'exploding'.
• − The starburst effect is a yellow glow rather than the requested 'fiery, glowing effect' seen in the title.

GPT Image 1

• + Excellent chalk texture throughout the lettering
• + Followed the prompt to complete the 'Brown Butter' text intelligently
• + Perfectly legible and aesthetically pleasing layout

• − Missed the request for 'elegant cursive' for the title
• − The handwriting is a bit too uniform, bordering on looking like a font

Z-Image Turbo

• + Realistic smudging and dust on the chalkboard surface
• + Varied letter sizes and slanting feel very human
• + Bold, clear presentation

• − Spelling error: 'Mustroom' instead of 'Mushroom'
• − Missed the request for elegant cursive in the title
• − The chalk strokes look a bit more like a digital brush than actual chalk on board

GPT Image 1

• + Excellent cinematic lighting and atmospheric depth in the space background.
• + Highly detailed texture on the astronaut suit and horse's coat.
• + Clearer visual of Earth and distant celestial bodies enhancing the setting.

• − The reins seem to disappear or merge oddly with the astronaut's hand.
• − The horse's back hooves have a slightly strange shape/orientation.

Z-Image Turbo

• + Bright, clear focal point with good color contrast between the white suit and brown horse.
• + The horse's mane and tail have a dynamic, flowing appearance.

• − Failed the specific prompt instruction 'horse on top, not vice versa' by placing the astronaut on top of the horse.
• − The background is relatively flat and less cinematic than the competitor.
• − The lighting on the astronaut doesn't naturally match the deep space environment.

GPT Image 1

• + Excellent photorealistic lighting and shallow depth of field.
• + Accurate text rendering on the hat.
• + High-quality fur texture and realistic passenger expression.

• − The capybara's paws look slightly human-like in their grip.
• − Visible exterior roof light is partially cut off at the top.

Z-Image Turbo

• + Good inclusion of safety elements like the seatbelt.
• + Clear side profile of the capybara showing the professional expression.
• + Accurate placement of both front paws on the steering wheel.

• − The passenger appears to be in the front passenger seat rather than the requested back seat.
• − The hands/paws on the steering wheel have a distorted, primate-like anatomy.

GPT Image 1

• + Exceptional typographic layout and legibility
• + Atmospheric cinematic lighting on the jack-o-lantern
• + Clean and professional graphic design style

• − Confused the Time and Location fields in the bottom text
• − The parchment texture is very subtle compared to the request

Z-Image Turbo

• + Stronger adherence to the 'dark parchment' and 'thorns' prompt elements
• + Included all required text fields correctly
• + Dynamic and layered composition

• − Typo in the location text ('Archves' instead of 'Arches')
• − The 'You are invited...' banner is floating awkwardly without a scroll behind it
• − Overall image is a bit cluttered

GPT Image 1

• + Successfully added a thick, full head of hair as requested
• + Preserved the glasses and majority of the original facial features
• + Maintained the original background and lighting well

Z-Image Turbo

• + Improved the texture and color of the beard
• + High overall image resolution and clarity

• − Failed the primary edit instruction by leaving the person largely bald
• − Altered the face shape and removed the person's glasses
• − Failed to preserve the original person's identity and facial features

GPT Image 1

• + Perfectly followed the text instructions, including the correct Japanese flag.
• + Higher complexity in the 3D scene with multiple sushi types and condiments.
• + Superior rendering of textures, particularly the subsurface scattering effect on the salmon.

• − The text 'JAPAN' is slightly off-center compared to the 'SUSHI' text.

Z-Image Turbo

• + Clean, minimalist aesthetic that fits the 'miniature' prompt well.
• + Good use of material depth on the diorama base.
• + Excellent centering of the text elements.

• − Used the Chinese flag instead of the requested Japanese flag for a Japan-themed image.
• − The sushi composition is very basic compared to the other model.
• − Visible artifacting/blurriness on the top of the salmon piece.

GPT Image 1

• + Perfectly follows the caricature instruction with exaggerated features and a watercolor art style.
• + Successfully incorporates all requested elements: news anchor desk, hockey equipment, and a companion dog.
• + Maintains recognizable features from the source image, such as the hair color and denim shirt.

• − The transition to a hand-drawn style means total loss of the original photographic background.
• − The hockey stick on the desk is slightly warped in perspective.

Z-Image Turbo

• + Excellent preservation of the original image's lighting, skin texture, and background composition.
• + Subtly adds a small dog in the background, maintaining a realistic photographic style.

• − Completely failed the main instruction to create a 'caricature' and make it 'exaggerated'.
• − Missing the 'tv show anchor' and 'hockey' elements entirely.
• − The added dog is blurry and lacks detail.

GPT Image 1

• + Excellent dynamic motion with animals appearing to run and tumble.
• + Superior lighting effects with clear god rays and a warm golden hour glow.
• + Better integration of the animals into the environment with accurate shadows and fur interactions.

• − The fox's front paws look a bit structurally odd or blurry.
• − The butterflies are somewhat simplified in detail compared to the foreground subjects.

Z-Image Turbo

• + High clarity on the textures of the animals' fur.
• + Successfully includes all requested subjects in a clear, centered group.

• − The composition feels more static and staged than 'playfully chasing'.
• − The lighting is flatter with less pronounced god rays compared to Image A.
• − The kitten's facial structure appears a bit distorted and unnatural.

GPT Image 1

• + Excellent adherence to the 'Studio Ghibli' style with hand-painted watercolor textures.
• + Perfectly captures the character expressions and layout of the original meme in an illustrative format.
• + Subtle use of soft pastel colors and dreamy lighting as requested.

• − The plaid pattern on the shirt is simplified to horizontal lines, losing part of the original detail.

Z-Image Turbo

• + High preservation of the original subjects' realistic faces and the specific plaid pattern of the shirt.

• − Completely failed the stylistic edit; the image remains a photograph rather than a Ghibli-inspired illustration.
• − The colors are slightly muted but do not meet the 'soft pastel' or 'hand-painted' criteria.
• − The woman on the right has her expression changed from angry to neutral, losing the context of the meme.

GPT Image 1

• + Excellent adherence to the hair motion request with strands visibly wind-swept.
• + Abundant flying leaves create a strong sense of dynamic motion throughout the frame.
• + Strong preservation of the original subjects' faces and general appearance.

• − The dog's leash has been shortened and the handle has become a messy loop disconnected from the hand.
• − The overall lighting is slightly harsher and more saturated than the source image.

Z-Image Turbo

• + Preserves the original structure of the image, including the dog's leash and handle, much better than Model A.
• + Maintains a natural, soft lighting that matches the source image closely.

• − The hair motion is very subtle, failing to fully capture the 'blowing in the wind' instruction.
• − Fewer flying leaves result in a less 'energetic and lively' feel compared to the other model.
• − Significant alterations to the face of the woman compared to the original source.

GPT Image 1

• + Excellent typography with correct inclusion of the accent mark
• + Strong vector logo composition with a centered banner
• + Good texture within the brown elements

• − Prompt requested a light background; it provided a black one
• − The banner shape is slightly irregular at the ends

Z-Image Turbo

• + Adhered perfectly to the request for a light background with subtle texture
• + Clean, professional typography and iconography
• + High icon clarity and symmetry

• − Missing the 'Est. 1720' banner, opting for horizontal lines instead
• − The 'è' accent mark is stylized as a leaf/drop, which may hinder legibility

GPT Image 1

• + Excellent adherence to the requested NASA-inspired color palette and flat-vector style.
• + Contains all the requested textual elements for mission steps, even with minor spelling errors.
• + Includes creative and relevant supporting details like the crew names and sillhouettes.

• − Has several typos including 'EARLLUNAR' and inconsistent placement of labels.
• − The layout is a bit cluttered and the flow of the steps is confusing to follow.

Z-Image Turbo

• + Clean, modern layout with plenty of white space that fits the minimalist infographic aesthetic.
• + High-quality vector icons, particularly for the Earth and Moon.
• + Very crisp lines and clear typography.

• − Significant spelling errors in every headline including 'APOLIO E 11', 'Translurian', and 'Descenty'.
• − Fails to include all requested steps and icons, missing the specific 'Lunar Orbit' and 'Landing' surface icons.