DALL-E 3 OpenAI vs Vidu Q2 ShengShu Technology

DALL-E 3

• + Excellent photorealistic lighting and textures
• + Highly creative interpretation with a detailed miniature landscape inside the sphere

• − Failed the spatial instructions by placing the book inside the cube and the sphere on top of the book
• − The 'glass cube' is more of a wooden frame with glass panels

Vidu Q2

• + Perfect adherence to the spatial requirements of the prompt
• + Accurate representation of all requested objects in their correct positions
• + Realistic caustic reflections from the glass onto the table

• − The plant is slightly cut off at the top of the frame
• − The glass cube has a minor artifacts where the top edge meets the book

DALL-E 3

• + Excellent composition with a unique 'foreground frame' perspective.
• + Atmospheric lighting and reflections that perfectly capture a rainy evening mood.
• + Strong cinematic aesthetic with a well-balanced layout.

• − The subject appears somewhat emaciated and stylistically exaggerated rather than natural.
• − Anatomical issues with the feet and hands are visible.
• − Lacks the requested motion blur from passing cars; the vehicle is static.

Vidu Q2

• + Captured realistic skin textures and fine details on the hands and face.
• + Includes subtle motion blur on the passing car as requested.
• + The bicycle shows realistic wear and tear, and the puddle reflections are convincing.

• − The composition is cluttered and lacks a clear focal point.
• − The framing feels accidental rather than 'imperfect' in an artistic sense.
• − The bicycle structure is physically impossible with two front sections merging.

DALL-E 3

• + Excellent depiction of warm torchlight reflecting off the metal and skin.
• + Highly detailed texture on the fabric and leather elements in the foreground.
• + Effective use of shallow depth of field and bokeh.

• − Failed to include the requested braided hair with beads.
• − The skin texture looks slightly over-processed and 'plastic' in high-contrast areas.

Vidu Q2

• + Successfully followed all prompt instructions, including the braided hair with beads.
• + Balanced composition that shows a wider range of the ornate plate armor.
• + Clear and realistic skin texture with subtle scarring and dirt.

• − The lighting is a bit flat compared to the 'warm torchlight' requested.
• − Some armor details, such as the rivets, appear slightly warped or inconsistent.

DALL-E 3

• + Provides a variety of layout options in a collage format.
• + Excellent food photography that looks professional and appetizing.
• + Strong grid-based composition that adheres to the request.

• − Text is largely nonsensical and garbled.
• − Shows four layouts rather than one focused menu design.

Vidu Q2

• + Closer to a functional menu with clear price placeholders.
• + Clean, bold sans-serif fonts that are mostly legible.
• + Better categorization of sections like Pizza and Mains.

• − Food photos are repetitive and lack variety compared to the prompt.
• − Some spelling errors in section headers (e.g., 'Apecizen').
• − Layout feels slightly more generic than Model A.

DALL-E 3

• + Excellent photographic texture on the meat and fresh vegetables
• + Dynamic composition with embers and flying debris creating a sense of impact
• + Sophisticated lighting with a strong glow reflecting off the bottom of the bun

• − Multiple spelling errors in the text including 'AGIC BURGR' and 'Limiited'
• − Failed to follow the starburst requirement for the price tag

Vidu Q2

• + Perfect text rendering for all requested phrases including the price
• + Accurately followed the starburst requirement for the price tag
• + Strong adherence to the 'fiery, glowing effect' for the typography

• − The burger ingredients look slightly more plastic and less photorealistic than Image A
• − The euro symbol is substituted with a generic currency symbol

DALL-E 3

• + Excellent chalk texture and artistic flourishes
• + The composition feels like a professional menu layout
• + Strong lighting and atmospheric café context

• − Numerous spelling errors including 'GRILILLED', 'OCCTUS', and 'TRUFLE'
• − The numbers are wildly inaccurate relative to the prompt (e.g., $234 instead of $24)
• − Text becomes unreadable gibberish in the smaller sections

Vidu Q2

• + Successfully rendered the date 'APRIL 30, 2026' perfectly
• + Captured the 'handwritten cursive' and 'slant' requested in the prompt
• + Much better adherence to the specified pricing ($28 for octopus)

• − Several spelling errors like 'Musshoom', 'Lemepun', and 'Octopd'
• − The last item is a mess of repeated words ('Cookies Chiir cpokies')
• − Chalk texture is less realistic than Model A, looking a bit like a digital brush

DALL-E 3

• + Excellent cinematic atmosphere with a vast starry background
• + Smooth blending of the horse into the ethereal environment

• − Failed the negative constraint: the astronaut is riding the horse, not vice versa
• − Lower level of intricate detail on the astronaut's suit compared to the competitor

Vidu Q2

• + High level of sharp, intricate detail on the spacesuit and horse tack
• + Creative galaxy-textured coat on the horse
• + Dynamic composition with vibrant colors

• − Failed the negative constraint: the astronaut is riding the horse, not vice versa
• − The horse's frontal anatomy looks slightly distorted where the neck meets the chest

DALL-E 3

• + Excellent fur texture rendering and lighting on the capybara's face.
• + High visual clarity and a cohesive cinematic art style.
• + Intelligent detail in the background with a 'Capybara' sign on a building.

• − Fails to include the human businesswoman in the back seat entirely.
• − The taxi cap is black instead of the requested yellow.

Vidu Q2

• + Successfully includes both the capybara driver and the businesswoman in the back seat.
• + Accurately represents the businesswoman's bored expression as she looks at her phone.
• + Followed more prompt constraints including the yellow color of the hat.

• − The transition between the capybara head and the human-like body is slightly awkward.
• − Some anatomical issues with the capybara's hands/paws on the steering wheel.

DALL-E 3

• + Exquisite gothic aesthetic with deeply layered 3D-like textures.
• + High visual quality with moody, cinematic lighting and a professional layout.
• + Captures the vintage 'scary' atmosphere perfectly with web and thorn details.

• − Fails significantly on text rendering, with most body text being illegible gibberish.
• − Missed the specific event details like 'The Arches' and the correct year in the text.

Vidu Q2

• + Successfully included all requested event details including location and date.
• + Stronger adherence to the specific text requests, including the banner text.
• + Good compositional balance with the central jack-o-lantern.

• − Several spelling errors in the main title and banner (e.g., 'Intoviztion', 'invieed').
• − Visual style is more illustrative and lacks the 'cinematic' polish and depth of the competitor.
• − The date contains a typo (30.70.2025).

DALL-E 3

• + Excellent 3D isometric execution with clean textures.
• + Great lighting and shadow work that enhances the miniature effect.
• + High level of rendering detail on the sushi and diorama base.

• − Failed to place the text 'JAPAN' and 'SUSHI' at top-center as requested.
• − Text rendering is incomplete, missing the word 'SUSHI' entirely.

Vidu Q2

• + Followed the text layout instructions perfectly with 'JAPAN' and 'SUSHI' at top-center.
• + Strong variety in sushi types depicted while maintaining the cartoon aesthetic.
• + Successfully included all elements including the flag icon and solid background.

• − The diorama base is less 'isometrically' accurate than Model A.
• − Minor artifacting on the textures of the shrimp and gunkan sushi.

DALL-E 3

• + Excellent adherence to lighting requests with clear 'god rays' and dew sparkles.
• + Very consistent stylistic rendering of fur and expressive eyes.
• + Cohesive composition that feels intimate and magical.

• − Has a distinct digital illustration/CGI feel rather than 'hyper-photorealistic'.
• − Strange 'butterfly-animal' hybrids where butterfly bodies have been replaced with fuzzy heads.

Vidu Q2

• + Achieves a much higher level of photorealism as requested.
• + Successfully depicts the movement and 'playfully chasing' aspect of the prompt.
• + All four requested animal types are present and identifiable.

• − Composition includes an extra puppy that was not requested.
• − The kitten's tail has a slightly awkward anatomical connection to its body.
• − Butterflies feel a bit 'pasted on' with some lacking natural drop shadows.

DALL-E 3

• + Excellent visual quality with clear, professional vector-style graphics.
• + Captures the vintage aesthetic perfectly with stippling textures and balanced composition.
• + Includes the required cloche and date elements in a coherent design.

• − Failed to include the specific name 'Caffè Florian', substituting it with 'Coffee House'.

Vidu Q2

• + Successfully incorporated a banner as requested.
• + Closely followed the requested color palette of warm browns and creams.

• − Significant text hallucinations and spelling errors like 'Farmiin' and 'Esttt'.
• − The cloche handle is detached and floating awkwardly above the steam.
• − Lacks the professional polish and clean lines of a vector logo.

DALL-E 3

• + Exhibits a professional and complex layout consistent with high-end infographics.
• + Captures the requested navy, muted red, and light gray NASA palette perfectly.
• + Includes stylistically consistent, textured circles that evoke lunar surfaces.

• − Fails to follow the specific 6-step sequence requested in the prompt.
• − Contains significant text gibberish and visual repetition across the three layout options.
• − The inclusion of Space Shuttle-style imagery is historically inaccurate for the Apollo 11 mission.

Vidu Q2

• + Attempts to follow a sequential step-by-step layout as requested.
• + The vector icons (lunar module, rocket, earth) are clean and have consistent line weights.
• + Adheres reasonably well to the flat-vector style with subtle gradients.

• − Severe spelling errors throughout (e.g., 'Alfonch' for Apollo, 'Aeulth Orot').
• − Fails to meet the 'navy' color palette requirement, using mostly light gray and white.
• − Logic errors in icons, such as the lunar module being repeated for 'descent' and 'landing' without variation.