GPT Image 2 OpenAI vs Wan 2.6 Alibaba

GPT Image 2

• + Excellent photographic quality with very clean glass reflections
• + High-detail texture on the red book cover
• + Strong adherence to the 'soft window light' requirement

• − The plant appears more next to/behind the cube than partially visible 'through' it as requested

Wan 2.6

• + Successfully shows the plant distorted and visible through the glass cube
• + Realistic weathered texture on the red book
• + Strong lighting interaction with reflections and shadows on the table

• − The glass cube has some structural inconsistencies in its thickness and geometry
• − The blue sphere has a slightly unnatural glow/internal reflection

GPT Image 2

• + Excellent photographic realism with very natural skin textures
• + Perfectly captures the 'imperfect framing' and 'candid' street photography style requested
• + Very high technical clarity on foreground objects like the toolbox and bike components

• − The rain is barely visible, appearing more like a damp day than light rain
• − Motion blur on the passing vehicle is present but feels a bit static compared to Model B

Wan 2.6

• + Atmospheric depiction of rain with visible droplets on clothing and pavement
• + Stronger cinematic feel with pronounced bokeh and neon reflections
• + Better representation of motion blur from the passing car

• − Significant anatomy issues with multiple instances of extra fingers and mangled hand structure
• − The rain droplets on the jacket look like static glass beads rather than water
• − Slightly more stylized and lacks the raw 'candid' feel of Model A

GPT Image 2

• + Excellent high-frequency texture on the skin and etched metal surfaces
• + Natural blending of the braided hair with the head and shoulders
• + Subtle and realistic application of dirt and faint scars

• − Lighting is a bit flat compared to the requested warm torchlight effect
• − Beads in hair are very small and less distinct than requested

Wan 2.6

• + Stronger adherence to the lighting prompt with warm highlights and bokeh sparks
• + Excellent rendering of large, distinct beads and leather straps
• + Highly weathered and battle-worn feel with visible mud and scars

• − Some messy artifacts where the braid connects to the armor
• − Lower facial realism compared to the skin texture in Model A

GPT Image 2

• + Excellent text rendering with clear, legible, and relevant menu items.
• + Perfect execution of the requested grid layout with distinct sections for appetizers, pizza, and mains.
• + Professional and realistic food photography that matches the descriptions.

• − The layout is quite dense, which slightly pushes the limits of 'minimalist'.

Wan 2.6

• + Good use of vibrant accent colors in the border design as requested.
• + Clean white background consistency throughout the layout.

• − Text is largely unintelligible gibberish with severe artifacts.
• − The food photos are repetitive and do not match the section headers (e.g., pizzas under the 'Appetizers' heading).
• − Poor section organization and inconsistent alignment of text elements.

GPT Image 2

• + Excellent text rendering with impressive fiery glow effects.
• + Superior photorealistic detail in food textures like the patty and sesame seeds.
• + Perfect adherence to the starburst and layout requirements.

• − The composition is a bit crowded on the left side due to text size.

Wan 2.6

• + Atmospheric use of smoke and embers in the background.
• + Good sense of vertical motion and depth in the burger explosion.

• − Failed to render the price text within a fiery glowing effect, opting for a flat graphic look.
• − The 'LIMITED TIME ONLY' text is placed at the very bottom and lacks the specified fiery impact.
• − The sauce stream connecting the bun and burger looks slightly unnatural.

GPT Image 2

• + Excellent text legibility and accuracy
• + Realistic chalk texture with fine details
• + Natural, consistent handwriting style across all items

• − Composition is a bit static and centered
• − The chalk holder on the ledge looks slightly blended into the wood

Wan 2.6

• + Dynamic composition with a more authentic cafe atmosphere
• + Great lighting and smudge effects on the chalkboard
• + Higher visual interest with the 'messy' chalk dust

• − The letter 'P' in 'Specials' is poorly formed
• − Some inconsistencies in font weight between the title and the items

GPT Image 2

• + Excellent adherence to the logic-defying prompt request of a horse on top of an astronaut.
• + Highly detailed texture on the spacesuit and lunar surface.
• + Creative interpretation of the harness and saddle setup for the horse.

• − The astronaut's hands/gloves have an anatomically incorrect number of fingers (6 or more).
• − The horse's front legs/hooves are awkwardly integrated into the harness handles.

Wan 2.6

• + Beautiful cinematic lighting and vibrant nebula colors.
• + Solid composition with dynamic movement and high visual appeal.
• + Technically well-executed rendered details on the horse's mane and suit.

• − Completely failed the negative constraint/specific instruction for the horse to be on top.
• − Cliches interpretation of the prompt that ignores the primary surreal requirement.

GPT Image 2

• + Excellent preservation of the subject's face, hair, and vitiligo patterns
• + High fidelity to the specific outfit from Image 2, including the plaid scarf pattern and watch
• + Successfully expanded the frame downwards to accommodate the full outfit and pose

• − The lighting on the coat is slightly flatter than the original outdoor scene

Wan 2.6

• + Successfully applied sunglasses found in several reference outfits
• + Integrates the coat with realistic lighting and shadows

• − Significant loss of detail in the subject's face and eyes
• − Changed the scarf color/pattern significantly from the reference image
• − Cropped the bottom of the image, losing the jeans and hand-in-pocket detail requested

GPT Image 2

• + Excellent photorealism in textures, especially the capybara's fur and the jacket fabric.
• + Cinematic lighting and composition that feels like a high-end film still.
• + The scale of the capybara relative to the steering wheel and seat is very convincing.

• − The passenger's face is slightly blurry and lacks fine detail.
• − The perspective makes the capybara appear extremely large, almost filling the front of the car.

Wan 2.6

• + Provides a clear view of both the driver and the passenger as requested.
• + Captures the 'bored' expression of the businesswoman very effectively.
• + Includes realistic details like rain on the windshield and a taxi roof light.

• − The passenger is sitting in the front passenger seat instead of the back seat.
• − The capybara's hands/paws on the wheel look a bit distorted and unnatural.
• − The lighting is flatter and less cinematic compared to the other model.

GPT Image 2

• + Excellent typography with perfect spelling in all requested text fields
• + Superior adherence to the vintage parchment aesthetic
• + Intricate border detail including skulls, thorns, and webs that feel naturally integrated

• − The color palette is very monochromatic compared to Model B

Wan 2.6

• + Strong cinematic lighting with vibrant orange glow from the jack-o-lantern
• + Good use of depth and contrast with the blue night sky
• + Clear representation of twisted trees mentioned in the prompt

• − The typography is less 'elegant gothic' and more standard digital font
• − The border feels less like a cohesive vintage invitation and more like an overlay
• − Slightly less 'polished' look in the text rendering compared to Model A

GPT Image 2

• + Excellent 3D text rendering with shadows and outlines.
• + High-detail models for the sushi, base, and decorative elements.
• + Rich adherence to all prompt elements including the flag icon and complex lighting.

• − Includes more garnish and background elements than the requested 'minimal' amount.

Wan 2.6

• + Captures the 'minimal' aesthetic requested in the prompt perfectly.
• + Accurate isometric layout and solid blue background.
• + Clean, soft textures that match the 'miniature 3D cartoon' style.

• − Text rendering is flat and lacks the high-clarity 3D finish seen in the other model.
• − The sushi models look slightly more generic and less like realistic PBR materials.

GPT Image 2

• + Excellent character coherence where all animals feel like they occupy the same space
• + The fox and kitten exhibit highly detailed fur textures and realistic anatomy
• + Subtle but effective use of god rays and backlighting that creates a warm atmosphere

• − The bunny is partially hidden behind the grass, losing some of the 'playful' impact
• − Some butterflies in the background are low-resolution or generic

Wan 2.6

• + Fantastic dynamic composition with all animals clearly visible and interacting
• + Beautiful 'dew sparkles' effect through bokeh and light refraction in the foreground
• + Stronger god rays that perfectly match the requested sunrise light

• − The kitten's facial structure and eyes look slightly unnatural/uncanny
• − Floating dandelion seeds or artifacts appear a bit cluttered in the upper portion

GPT Image 2

• + Excellent typography with clean, professionally rendered text.
• + High-quality etching detail on the cloche and banner.
• + Superb layout with a cohesive frame and balanced composition.

• − The detail level is very high, pushing slightly away from the 'minimalist' request.

Wan 2.6

• + Successfully adopts a simpler, more minimalist vector style.
• + Accurate adherence to the warm brown and cream color scheme.

• − The banner is very small and awkwardly placed coming out of the cloche.
• − The steam icon is overly simplified and lacks the requested vintage feel.
• − The background texture looks like digital grunge brushes rather than subtle paper texture.

GPT Image 2

• + Excellent adherence to the multi-step infographic structure and specific iconography requests.
• + Highly accurate text rendering for the title, steps, crew names, and landing site.
• + Strong visual composition with a professional, clean vector aesthetic that fits the NASA theme.

• − The lunar module in step 5 contains slightly dense detail for a 'flat vector' style.

Wan 2.6

• + Successfully applied the requested navy, white, and red color palette.
• + Includes the requested crew names in a legible font.

• − Completely failed to generate the six-step infographic workflow requested in the prompt.
• − The background has a textured, fabric-like appearance instead of a clean digital vector finish.
• − Lacks the specific icons (Saturn V, Earth, Moon, etc.) requested.