GPT Image 1 OpenAI vs Wan 2.6 Alibaba

GPT Image 1

• + Perfect compliance with spatial positioning for all objects.
• + Clean, minimalist aesthetic with high resolution.
• + Excellent rendering of light through translucent curtains.

• − The sphere appears more matte than expected for a 'blue sphere' contrast to the glass.
• − The glass cube edges appear slightly impossible in their thickness/joining.

Wan 2.6

• + Highly realistic textures on the vintage book and weathered wooden table.
• + Beautiful, complex light interactions including caustics and reflections.
• + Very natural-looking plant integration.

• − The plant is more 'next to' the cube than 'behind' it as seen through the glass.
• − Small artifacts in the reflection on the table surface.

GPT Image 1

• + Excellent preservation of the car's model and specific design details
• + Great lighting integration between the man, the car, and the environment
• + Realistic camera angle and motion blur on the wheels

• − The man's facial features and dreadlock style are slightly altered from the source
• − The road texture in the foreground is a bit grainy

Wan 2.6

• + Successfully maintains the man's facial likeness and specific plaid coat from the source
• + Vibrant and scenic California coastline background
• + Accurately places the driver in a right-hand drive configuration consistent with the console edit

• − The car's proportions are distorted, making it look much longer and narrower than a Rolls-Royce
• − Loss of detail on the car's front grille and hood area compared to the source

GPT Image 1

• + Excellent skin and hair texture realism
• + Captures a narrow depth of field and bokeh realistically
• + Composed with a strong cinematic focus on the subject's face

• − Lacks the requested motion blur from passing cars; background traffic is static and out of focus
• − The bicycle chain and wheel spokes are simplified and lack structural detail

Wan 2.6

• + Stronger environmental storytelling with tools and a wet street aesthetic
• + Better representation of 'light rain' through visible droplets on skin and clothing
• + Captures the 'imperfect framing' requested in the prompt

• − Anatomical errors in the hands with cluttered, merged fingers
• − The rain droplets on the jacket look like static plastic beads rather than liquid
• − The bicycle geometry is slightly warped at the front wheel

GPT Image 1

• + Exquisite engraving detail on the plate armor
• + Highly realistic skin textures with subtle scaring and grime
• + Excellent atmospheric lighting that feels naturally integrated

• − The beads in the hair are very subtle and blend in too much
• − Less variation in the underlayer textures compared to model B

Wan 2.6

• + Strong adherence to the 'beads' prompt with clear, vibrant decorative elements
• + Excellent contrast between the plate, leather straps, and frayed cloth underlayers
• + Dynamic lighting and very expressive, lifelike eyes

• − The dirt on the face looks slightly digital or 'painted on' in patches
• − The torch in the background is a bit distracting and lessens the 'shallow depth of field' effect on the sparks

GPT Image 1

• + Excellent typography rendering with almost perfect spelling
• + High-quality, appetizing food photography
• + Clean layout that closely follows the minimalist prompt

• − The grid is slightly cut off at the bottom
• − Only shows Appetizers and Pizza sections clearly, missing a distinct 'Mains' heading

Wan 2.6

• + Includes all requested headings: Appetizers, Pizza, and Mains
• + Features a distinct grid of photos and vibrant colorful accents in the branding
• + Better overall balance of the entire menu page

• − Poor text quality with significant gibberish and typos
• − Some food images in the grid look repetitive or messy
• − Layout is a bit cluttered compared to the minimalist request

GPT Image 1

• + Excellent typography rendering with crisp, fiery outlines
• + Very clean and professional composition suitable for a high-end ad
• + Accurately represents all burger layers in a clear, exploded view

• − Missed the '6' in the price tag, displaying '.99' instead
• − The background is relatively static compared to the burger's motion

Wan 2.6

• + High emotional energy with smoke and realistic fire effects
• + Accurately includes the full price '€6.99' in the starburst
• + Dynamic tilted composition enhances the 'exploded' and 'motion' feel

• − Main title text is slightly muddy and less legible than Model A
• − The 'starburst' for the price looks like a 2D sticker slapped onto a 3D scene

GPT Image 1

• + Excellent text legibility and spelling accuracy.
• + Very clean, centered composition.
• + Renders a realistic chalk grit texture on the letters.

• − The handwriting feels a bit too uniform, bordering on a digital font look despite the texture.
• − Does not use 'elegant cursive' for the title as requested.
• − The framing is a tight crop that feels less like a 'cozy café' environment.

Wan 2.6

• + Successfully uses a beautiful cursive style for the title text.
• + Much more realistic 'cozy café' atmosphere with depth of field and environmental details like chalk dust.
• + Very natural variation in handwriting slant and pressure.

• − Slightly less crisp text rendering compared to the other model.
• − Minor artifacts in the chalk dust at the bottom of the frame.

GPT Image 1

• + Excellent cinematic lighting and dark, moody atmosphere
• + Strong anatomical rendering of the horse and textures on the spacesuit
• + Higher level of fine detail in the horse's coat and muscle definition

• − The pose is somewhat static for a 'cinematic' scene
• − Failed the negative constraint; the prompt requested 'horse on top' to be surreal, but it showed the standard rider.

Wan 2.6

• + Vibrant color palette with striking nebulae and light rays
• + Dynamic sense of motion with the horse's mane and dust effects
• + Cleaner reflection on the astronaut's visor

• − Failed the logical constraint; specifically ignored the 'horse on top, not vice versa' instruction
• − Slightly less realistic rendering of the horse's legs compared to Model A

GPT Image 1

• + Excellent replication of the specific plaid pattern and coat style from Image 2
• + Successfully added accessories like the watch that were present in the reference
• + High visual quality with clean lighting and fabric textures

• − Significantly altered the subject's face and changed the vitiligo pattern from Image 1
• − Modified the background landscape, removing various details on the sand

Wan 2.6

• + Perfectly preserved the original background from Image 1
• + Highly accurate preservation of the subject's face, hair, and vitiligo patterns
• + Correctly included the sunglasses from the reference image whereas Model A forgot them

• − The plaid pattern on the scarf is slightly simplified compared to the source
• − The jacket is cropped at the waist, failing to demonstrate the fit of the jeans and full outfit

GPT Image 1

• + Excellent composition with a focused, realistic close-up on the capybara's expression.
• + Higher photographic quality with realistic textures on the fur and the hat materials.
• + Very accurate adherence to the 'both paws on steering wheel' and professional attire instructions.

• − The passenger is heavily blurred, making it harder to discern her phone and expression.
• − The interior is very dark, which obscures some of the requested taxi details.

Wan 2.6

• + Great environmental storytelling with rainy streets and vibrant Manhattan neon lights visible through the glass.
• + Both characters are in clear focus, making the passenger's 'bored' expression visible.
• + Includes more external taxi details like the checkerboard trim and roof light.

• − The steering wheel placement is physically impossible, appearing to emerge from the center console/passenger side.
• − The capybara's paws look somewhat distorted and do not grasp the wheel realistically.
• − The composition feels a bit crowded and flatter compared to the cinematic depth of Model A.

GPT Image 1

• + Excellent typographic consistency and gothic style.
• + Cleaner, more atmospheric lighting that feels cinematic.
• + Perfectly captures the 'dark parchment' texture and moody aesthetic.

• − Failed to include '7pm' and mislabeled the location under the 'TIME' header.
• − The jack-o-lantern and trees are very subtle/dark compared to the prompt's focus.

Wan 2.6

• + Highly accurate adherence to all text details including date, time, and location.
• + Richly detailed border containing both the requested thorns and spiderwebs.
• + Stronger visual hierarchy with a more vibrant jack-o-lantern and twisted trees.

• − The font for the secondary scroll banner is a bit generic compared to the main title.
• − The 'parchment' effect is limited to the edges rather than the whole background.

GPT Image 1

• + Excellent preservation of the original facial anatomy and skin texture
• + Correctly applies a thick, dark hair texture that matches the beard style
• + Accurate preservation of the background and original lighting environment

• − The hairline is a bit unnatural and sits too high on the forehead
• − The hair volume is slightly exaggerated and looks somewhat like a wig

Wan 2.6

• + Natural and realistic hair flow and texture
• + Flawless integration of the hair with the ears and temple area
• + Highly convincing hairline and hair density

• − Slightly alters the shape of the person's face/forehead compared to the source image
• − Small artifact where the hair meets the top frame of the glasses

GPT Image 1

• + Excellent typography with clean, centered layout and correct stacking.
• + High-quality 3D clay-like textures match the 'cartoon scene' and 'soft' descriptors perfectly.
• + Superior composition that fills the frame more effectively while maintaining the requested isometric view.

• − The salmon texture is slightly simplified, appearing more like plastic than realistic PBR textures.

Wan 2.6

• + Stronger adherence to 'realistic PBR' for the sushi materials, especially on the shrimp and rice.
• + Good isometric perspective and clean background.

• − Typography layout is flawed, placing the flag icon in the middle of a line rather than below the text.
• − Shadow clipping occurs at the bottom of the diorama base.
• − The text occupies too much of the upper frame compared to the relatively small subject.

GPT Image 1

• + Excellent caricature style with exaggerated facial features that still maintain the subject's likeness
• + Creative integration of all elements including a hockey-playing dog on the TV screen
• + The denim shirt from the original photo is accurately preserved

• − The watercolor texture is a bit messy around the hands

Wan 2.6

• + Clean digital vector art style
• + Includes multiple dogs and professional studio equipment like headsets and lights
• + Clear rendering of the hockey stick and puck

• − The facial features are generic and do not resemble the woman in the source image
• − The 'caricature' aspect is more of a standard cartoon than an exaggerated caricature

GPT Image 1

• + Excellent character composition with all four animals clearly visible and interacting
• + Consistent, soft lighting that backlights the fur beautifully
• + Better anatomical accuracy on the animals' paws and expressions

• − The kitten's tail is missing/obscured
• − Butterflies are slightly less detailed compared to Model B

Wan 2.6

• + Beautiful atmosphere with strong god rays and sparkling dew effects
• + Each animal has a distinct personality and clear tail visible
• + High level of detail in the butterflies and foreground flora

• − The fox's front-right paw is anatomically messy
• − The kitten's front-left paw looks slightly distorted

GPT Image 1

• + Successfully captures a textured, hand-painted watercolor aesthetic reminiscent of classic anime backgrounds.
• + Accurately recreates the character's facial expressions in a simplified artistic style.
• + Maintains the warm, nostalgic, and soft pastel color palette requested.

• − Simplifies the plaid pattern of the shirt too much compared to the source.

Wan 2.6

• + Stays very close to the source image's composition and specific details like the plaid pattern on the shirt.
• + Features clean linework and clear character recognition.
• + Incorporates a pleasant watercolor wash texture.

• − The white speckle overlay feels like an artificial filter rather than part of the illustration style.
• − The character style leans more towards modern manga/manhwa than the specific 'Studio Ghibli' aesthetic requested.

GPT Image 1

• + Strong adherence to all parts of the motion prompt
• + Successfully added movement to the dog's fur/tail and the woman's hair
• + Includes a large quantity of flying leaves providing a sense of wind

• − Introduced some texture inconsistencies on the woman's face and original clothing details
• − Leaves are somewhat blurry and lack high-definition detail

Wan 2.6

• + Excellent preservation of the source image's identity, face, and clothing details
• + Clean, high-quality rendering of the flying leaves
• + Subtle but natural-looking hair movement that respects the original lighting

• − The wind effect is less energetic compared to the prompt's request
• − The dog remains static, failing to include it in the 'lively' feel of the scene

GPT Image 1

• + Excellent typography with perfect spelling and correct accent marks.
• + Strong minimalist vector style that feels modern yet vintage.
• + Balanced composition with the banner at the base.

• − Ignored the 'light background' request, providing a black background instead.
• − The steam element is a bit thick and less elegant than Model B.

Wan 2.6

• + Successfully used a light background with subtle vintage texture as requested.
• + The 'Est. 1720' banner is integrated into the icon in a creative way.
• + The steam graphic is delicate and well-rendered.

• − The text 'Caffè' uses a backtick instead of a proper accent grave.
• − The composition feels slightly off-balance with the banner only on the right side.

GPT Image 1

• + Excellent adherence to the six requested steps and iconography.
• + High-quality vector aesthetic with a consistent NASA-inspired color palette.
• + Legible text and logical layout that functions as an actual infographic.

• − Includes a typo in the footer ('EARLLUNAR' rather than individual steps).
• − The icon-to-label mapping is slightly misaligned in the middle row.

Wan 2.6

• + Features a clean, minimalist design.
• + Accurate spelling for the astronaut names.

• − Completely failed to include the requested infographic steps (Launch, Orbit, etc.).
• − Looks more like a towel or fabric texture than a vector poster.
• − Lacks the specific iconography and flat-vector style requested.