GPT Image 1 OpenAI vs Qwen Image 2512 Alibaba

GPT Image 1

• + Excellent photographic quality with realistic textures on the book and wooden table.
• + Perfect logical coherence with the sphere clearly inside the cube and the book resting naturally on top.
• + Very clean glass rendering with believable reflections and refractions.

• − The sphere appears to be floating slightly above the bottom surface of the cube.

Qwen Image 2512

• + Natural handling of reflections on the interior faces of the glass cube.
• + Follows all spatial instructions in the prompt correctly.
• + Good bokeh effect on the background plant.

• − The cube's glass appears more like a mirror on some faces, obscuring the interior slightly.
• − The texture of the red book is a bit less realistic compared to Image A.

GPT Image 1

• + Excellent shallow depth of field and bokeh quality
• + Highly realistic skin textures and lighting
• + Captures the 'repairing' action more convincingly

• − Slightly less motion blur on the cars than requested
• − Bicycle geometry is a bit simplified in the rear

Qwen Image 2512

• + Captures the 'imperfect framing' and wider street context well
• + Better sense of the wet pavement reflections
• + Includes more visible rain streaks

• − The subject is posing rather than repairing the bike
• − The bike proportions are warped, appearing too small for the man
• − Significant anatomical errors with the hands

GPT Image 1

• + Excellent depiction of ornate, swirled engraving on the plate armor
• + Subtle and realistic integration of grime and skin texture
• + Strong focus on a battle-worn facial expression

• − Lacks the requested leather straps and cloth underlayer detail
• − The beads in the hair are very sparse compared to the multiple braids shown in Model B

Qwen Image 2512

• + Perfectly captures all elements including leather straps and cloth textures
• + Accurate bead detailing in multiple complex braids
• + Dynamic lighting with visible torch and high-quality bokeh sparks

• − The scars look slightly 'painted on' compared to the skin texture
• − The armor engraving is slightly less intricate than Model A's

GPT Image 1

• + Excellent typography with clean, almost-perfect English text
• + High-quality, realistic food photography with vibrant colors
• + Strong minimalist aesthetic that looks like a real professional layout

• − Layout is slightly cut off at the bottom
• − Text descriptions beneath names are repetitive gibberish

Qwen Image 2512

• + Successfully creates a full document mockup view within a grid
• + Good use of color blocking for organizational sections

• − Text is largely illegible and contains numerous spelling errors
• − Food photos are small and lack the clarity of the competing model
• − Layout feels cramped and cluttered compared to the minimalist prompt

GPT Image 1

• + Excellent typography with a consistent glowing ember effect on all text elements.
• + High-quality photorealistic textures on the burger ingredients.
• + Superb adherence to the 'exploded' concept with clear separation of all layers.

• − Failed to include the '6' in the price, displaying '.99' instead of '€6.99'.
• − The starburst shape is a bit simplified and digital looking.

Qwen Image 2512

• + Perfect accuracy on all required text strings, including the price.
• + Highly dynamic sense of motion with debris and sauce droplets flying out.
• + Rich, multi-layered composition with flames in the foreground and background.

• − The 'LIMITED ONLY' text is missing the word 'TIME'.
• − The typography style is less cohesive than Model A, with various fonts and effects mixed together.

GPT Image 1

• + Excellent chalk texture throughout all characters
• + Perfect spelling for all requested menu items
• + Consistent letterforms that look authentically like the same person's handwriting

• − The 'cursive' requirement for the title is not fully met as it is more of a print style
• − Composition is a bit crowded vertically

Qwen Image 2512

• + Successfully followed the request for elegant cursive in the title and menu items
• + Great environmental storytelling with the blurred café background
• + Better use of space and layout for a menu board

• − One spelling error ('Risitto' instead of 'Risotto')
• − The chalk texture is a bit too smooth and looks slightly like a digital brush

GPT Image 1

• + Excellent cinematic lighting and texture on both the suit and horse fur.
• + Higher overall resolution and artistic detail in the background.
• + Stronger surreal atmosphere with the dark starfield and planetary curve.

• − Failed the negative constraint; the astronaut is riding the horse instead of vice versa.

Qwen Image 2512

• + Clear facial detail within the helmet visor.
• + Good anatomical representation of the horse and tack equipment.
• + Sharp contrast between the subject and the bright limb of the Earth.

• − Failed the crucial negative constraint; depicts a human riding a horse.
• − Lighting on the astronaut is inconsistent with the dark background space.

GPT Image 1

• + Excellent texture on the capybara's fur and the leather taxi hat.
• + Cinematic lighting that accurately reflects a New York night scene.
• + Subtle and professional expression on the capybara's face.

• − The passenger's hand holding the phone is slightly distorted and blurry.
• − The capybara's paws are somewhat indistinctly placed on the steering wheel.

Qwen Image 2512

• + Clearer depiction of both front paws on the steering wheel as requested.
• + The passenger's expression more accurately captures the 'bored/annoyed' look.
• + Very sharp focus and bright, vibrant colors throughout the scene.

• − The capybara's paws look more like primate hands with claws rather than capybara paws.
• − The taxi sign on the hat is a generic crest rather than the requested 'TAXI' text.

GPT Image 1

• + Excellent typography style that matches the gothic aesthetic
• + Very clean layout with clear hierarchical design
• + High artistic quality of the jack-o-lantern illustration

• − Confused the 'Time' and 'Location' labels in the footer
• − Missing the requested '7pm' time entirely

Qwen Image 2512

• + Successfully included all requested text labels and event details without error
• + Detailed border with prominent thorns and webs
• + Strong cinematic lighting with vibrant contrast

• − Includes a spelling error in the main header ('Hallowern')
• − The banner scroll cuts into the bottom text area slightly awkwardly

GPT Image 1

• + Excellent typography with clean, bold text and perfect spacing.
• + High-quality 3D clay-like textures that perfectly match the 'soft refined' request.
• + Superior lighting and shadow work on the diorama base.

• − The sushi variety is quite limited with only two types shown.
• − The flag icon is a bit large compared to the text.

Qwen Image 2512

• + Includes a wider variety of sushi (nigiri and maki).
• + Creative dioroma base with tiny grass and floral details.

• − Text rendering is lower quality with inconsistent outlines and spacing.
• − The 'JAPAN' text is slightly off-center to the left.
• − The overall image clarity is lower than Model A.

GPT Image 1

• + Excellent sense of motion and 'tumbling' as requested in the prompt
• + Superior lighting with realistic volumetric god rays and soft glows
• + Highly expressive, joyful facial expressions on all four animals

• − The fox kit has slightly unusual dark paws that look a bit like hooves or rounded stubs
• − The butterflies are less detailed compared to those in the second image

Qwen Image 2512

• + Exquisite detail in individual fur strands and butterfly wings
• + The animals are neatly grouped, making for a clear family-portrait style composition
• + Strong adherence to the '8K' and 'hyper-photorealistic' aesthetic

• − Static composition that fails to capture the 'playfully chasing' and 'tumbling' part of the prompt
• − Anatomical oddity where the kitten appears to lack a complete body behind the puppy's leg
• − The lighting feels a bit more artificial and post-processed compared to the natural look of image A

GPT Image 1

• + Clean vector emblem style
• + Accurate spelling including the accent on the 'e'
• + Good textured brown and cream tones

• − Generated on a black background instead of the requested light background
• − Minimalist steam is a bit too simplified and looks thin

Qwen Image 2512

• + Perfect adherence to the light background and subtle texture request
• + Excellent illustrative vintage style
• + Clear, legible typography and banner

• − The 'i' in Florian has an unnecessary dot above it in a script that doesn't usually feature it that way

GPT Image 1

• + Excellent adherence to the clean, flat-vector style requested.
• + Clear and legible typography with correct spelling for the most part.
• + Strict adherence to the specified NASA-inspired color palette.

• − Layout is a bit chaotic with text labels not aligning perfectly with their icons.
• − Has a spelling error in 'EARLLUNAR' and misses Lunar Orbit as a distinct step.

Qwen Image 2512

• + Strong composition that feels like a professional poster layout.
• + High level of detail on the spacecraft icons while maintaining a vector look.
• + Includes more of the requested sequential steps in a logical flow.

• − Numerous spelling errors in technical terms (e.g., 'TranslauraJ', 'Desceeint').
• − Text rendering is smaller and less crisp than its competitor.
• − Included the prompt text 'Steps stop at landing' literally into the image.