Nano Banana 2 Google vs Vidu Q2 ShengShu Technology

Nano Banana 2

• + Excellent text rendering on the book's spine
• + Highly realistic textures on the wooden table and book cover
• + Very accurate representation of soft window light from the side

• − The glass cube missing its top panel, making the book appear to float on just the edges

Vidu Q2

• + More physically coherent glass cube structure with a visible top face
• + Dynamic composition with attractive shadows
• + Accurate adherence to all prompt elements

• − The plant is slightly less 'behind' the cube compared to Model A
• − Text on the book is garbled and unreadable

Nano Banana 2

• + Excellent preservation of the car's original design and details.
• + Seamless integration of the subject including his specific clothing (plaid coat, scarf).
• + High-quality, realistic lighting and dynamic motion blur on the wheels.

• − The driver is positioned too far to the right, appearing to sit in the passenger seat for a left-hand drive car.

Vidu Q2

• + Great scenic composition of the California coastline.
• + Correctly positions the driver on the left side of the vehicle.

• − Significant loss of detail on the man's clothing and facial features.
• − The car's proportions feel slightly warped, and the front grille has some alignment issues.
• − The shadow of the car on the road is inconsistent with the lighting of the environment.

Nano Banana 2

• + Excellent full-body composition that tells a story
• + Realistic street photography aesthetic with visible rain and environment
• + Hand and tool interaction is coherent

• − The bike's kickstand and chain geometry are slightly warped
• − The wrench scale looks a bit large for the task

Vidu Q2

• + Strong skin texture and realistic anatomical detail in hands
• + Excellent 'imperfect framing' that feels like a candid macro shot

• − Anatomical issues where the man's arm and hand are fused with the bike frame
• − Missing the requested motion blur for passing cars

Nano Banana 2

• + Excellent adherence to the 'battle-worn' request with realistic dirt, grime, and skin texture.
• + Highly detailed material rendering on the engraved armor, leather straps, and textured undergarments.
• + Perfect atmospheric lighting with visible torchlight reflections and bokeh sparks.

• − The character's hand gripping the sword has slightly muddy anatomical detail on the knuckles.

Vidu Q2

• + Clean and sharp rendering of ornate armor engravings and metal highlights.
• + Good implementation of hair braids with beads as requested in the prompt.

• − The character looks too pristine and 'clean' for a battle-worn description.
• − The lighting feels more like studio lighting with an orange rim than natural torchlight.
• − The texture on the skin and fabric is slightly smoother and less realistic compared to Model A.

Nano Banana 2

• + Excellent text legibility with coherent and meaningful names and pricing.
• + Superior layout adherence to the prompt, featuring clearly defined Appetizers, Pizza, and Mains sections.
• + High-quality, realistic food photography that correlates directly with the text items.

• − Minor typos in smaller descriptive text such as 'somon' instead of salmon.
• − Simple aesthetic might be perceived as a template rather than a custom creative design.

Vidu Q2

• + Vibrant and artistic use of color and rounded design elements for a playful vibe.
• + Dynamic layout with stylized photo frames and colorful accents.

• − Text is largely unintelligible gibberish, failing the requirement for a functional menu.
• − Layout is chaotic and fails to clearly distinguish the specific requested sections.
• − Images for 'Pizza' and 'Mains' sections do not accurately represent the food categories.

Nano Banana 2

• + Excellent typography with a glowing fire texture that matches the prompt perfectly
• + Superior textures on the patty and bun, providing a higher sense of photorealism
• + Integrated all required text elements, including the starburst for the price

• − The starburst design is slightly generic compared to the high-quality text above

Vidu Q2

• + Dynamic fire background with good color vibrancy
• + The exploded burger components are well-distributed

• − The price text includes a nonsensical character instead of the Euro symbol
• − The lighting on the burger feels more illustrative and less photorealistic than Image A
• − The text lacked the specific 'fiery glowing' effect requested for the price starburst

Nano Banana 2

• + Excellent text rendering with perfect spelling for almost all items.
• + Highly realistic chalkboard texture with chalk smudging and authentic handwriting variations.
• + Great overall composition and background café atmosphere.

• − The final word of the prompt was cut off in the input, but the model made a logical completion that fits perfectly.

Vidu Q2

• + Natural chalk texture on the lettering with a nice layered effect.
• + Good focus on the board within the frame.

• − Numerous spelling errors including 'Truffe Musshoom', 'Risoto', and 'Lemepun'.
• − Illegible text at the bottom and messy overlapping in the cookie description.
• − Incorrect price for the first item ($34 instead of $24).

Nano Banana 2

• + Excellent character preservation including face, sunglasses, and clothing details
• + Accurately recreates the yellow background and red ottoman from Image 1
• + Maintains the scarf texture and pattern from the character reference

• − The pose is significantly altered; the torso is much more upright compared to the deep forward lean in Image 1
• − Anatomical issues with the feet and how they intersect with the legs

Vidu Q2

• + Successfully replicates the specific bent-over torso angle and forward head lean of Image 1
• + Very high fidelity in character likeness, including hair and skin tone
• + Includes clothing details like the scarf and watch while adapting to the pose

• − The character is wearing shorts instead of the long black pants from Image 2
• − The scarf rendering becomes slightly detached or floating away from the neck in the middle section

Nano Banana 2

• + Successfully followed the difficult counter-intuitive instruction of the horse riding on top of the astronaut.
• + High cinematic quality with realistic space lighting and detailed textures on the spacesuit.
• + Creative surrealist interpretation of the prompt.

Vidu Q2

• + Aesthetically pleasing colors and cosmic effects.
• + Good lighting and clean character rendering.

• − Failed the primary prompt constraint by placing the astronaut on top of the horse.
• − The horse's legs and hooves have anatomical issues.

Nano Banana 2

• + Maintains the subject's unique vitiligo patterns very faithfully.
• + Successfully incorporates the watch and ring accessories from Image 2.
• + Captures the open-coat styling that shows the subject's torso.

• − Introduces a hallucinated second person in the background.
• − Adds sunglasses that were not present on the subject in Image 1 and look poorly composited.
• − The hand rendering is distorted, especially the right hand.

Vidu Q2

• + High resolution and clean rendering of the clothing textures.
• + Better integration of the scarf and coat layers relative to the subject's pose.
• + Follows the aesthetic of the outfit while maintaining the beach environment.

• − Significantly alters the subject's face, adding a mustache and changing facial structure.
• − Fails to preserve the exact vitiligo pattern from Image 1.
• − Completely changes the background scenery, ignoring the 'keep background completely unchanged' instruction.

Nano Banana 2

• + Excellent attention to detail with the NYC hack license and fare meter.
• + High photorealism in the textures of the capybara's fur and the leather seats.
• + Perfectly captures the 'bored' expression of the passenger as requested.

Vidu Q2

• + Followed the specific color instruction for a yellow taxi hat.
• + Good composition using a side profile to show both characters clearly.
• + Clear, vibrant city lights in the background.

• − The passenger looks a bit sad or concerned rather than bored.
• − The capybara's hands look slightly more primate-like than capybara-like.
• − Lighting on the capybara's face is a bit flat compared to the background.

Nano Banana 2

• + Perfect text rendering for all requested details, including date and location.
• + High-quality, cohesive artistic style with rich moody lighting.
• + Complex composition with detailed borders, skulls, and a gothic background.

• − The layout is a bit crowded, making the middle section slightly compressed.

Vidu Q2

• + Follows the basic structure of the prompt.
• + Clear central jack-o-lantern focal point.

• − Significant spelling errors in every piece of text ('Intovztion', 'invieed ta might', etc.).
• − Incorrect date (30.70.2025 instead of 30.10.2026).
• − Lower visual fidelity and more simplistic interpretation of the background elements.

Nano Banana 2

• + Excellent source preservation with no changes to the background or clothing.
• + Very natural-looking hairline and believable hair texture that matches the existing beard.

• − The hair feels a bit thin or receding compared to the request for a 'full, thick' head of hair.

Vidu Q2

• + Successfully provided a very thick, full head of hair as requested.
• + Maintains the likeness of the subject and preserves the lighting well.

• − Slightly less realistic hairline compared to Model A, especially around the temples.
• − The volume of the hair looks a bit stylized/wind-blown rather than natural.

Nano Banana 2

• + Excellent PBR materials with realistic textures on the fish and rice grains.
• + Perfect text rendering and layout for 'JAPAN' and 'SUSHI'.
• + High visual complexity and clarity in the isometric scene.

• − The diorama base has a bit more green moss/garnish than 'minimal' requested.
• − The 'cartoon' style leans more towards photorealism, though it retains a miniature feel.

Vidu Q2

• + Successfully captures the 'cartoon' aesthetic with soft, rounded 3D shapes.
• + Clean diorama base and simple composition.
• + Follows the request for a small flag icon attached to the text.

• − The text has slight artifacting and less professional typography than Model A.
• − Render quality is lower with some muddy textures in the sushi toppings.
• − The 'isometric' perspective is a bit flatter and less dynamic.

Nano Banana 2

• + Excellent caricature style with colored pencil texture.
• + Highly creative and humorous text integration like 'W-K9 NEWS'.
• + Strong character likeness preserved in the exaggerated caricature.

• − The 'holding the photo' hand effect is a bit generic for a direct digital edit.
• − Minor anatomical awkwardness in how she holds the hockey stick.

Vidu Q2

• + Very clean, vector-style digital illustration.
• + Great integration of the hockey rink background and anchor desk.
• + Excellent preservation of the source subject's features and outfit.

• − The subject's right hand holding the microphone has an extra finger/distorted anatomy.
• − The caricature style is less 'exaggerated' and more like a standard portrait compared to typical caricatures.

Nano Banana 2

• + Perfect adherence to the count of animals requested (one of each species).
• + Excellent anatomical realism for all four animals.
• + Beautiful lighting with clear god rays and atmospheric depth.

• − The transition between the puppy's back paw and the grass is a bit blurry.

Vidu Q2

• + Vibrant colors and a very high number of butterflies enhancing the scene.
• + Cute, expressive eyes on the kitten and bunny.

• − Failed the animal count by including two golden retriever puppies instead of one.
• − The fox's anatomy looks slightly distorted with very long, dark front legs.
• − Overall composition feels slightly more cluttered and less realistic than the competitor.

Nano Banana 2

• + Excellent adherence to the Ghibli art style with watercolor textures and soft line work.
• + Creatively enhances the background with flowers and European-style architecture consistent with Ghibli's aesthetic.
• + Maintains character poses while perfectly translating their expressions into an anime format.

• − The woman in the foreground is slightly more blurred than in the original source image.

Vidu Q2

• + Successfully translates the image into an anime style while preserving the original background structure.
• + Preserves the color saturation of the original red dress more accurately than Model A.
• + Closer preservation of the original spatial depth and blurring in the foreground.

• − The background feels more like a standard filter rather than the 'hand-painted' Ghibli look requested.
• − The character in the red dress has a slightly awkward facial structure compared to the source.

Nano Banana 2

• + Excellent wind-blown hair effect that looks natural and dynamic
• + Modified the subject's expression to match the energetic feel
• + Highly effective integration of blowing leaves

• − The facial expression change, while thematic, deviates significantly from the source identity
• − Noticeable artifact under the woman's right arm

Vidu Q2

• + Strong source preservation of the woman's face and expression
• + Effective hair movement that feels consistent with the breeze
• + Good distribution of flying leaves throughout the frame

• − The leaves appear somewhat flat and 'pasted on' with less motion blur than Model A
• − A small artifact is present above the dog's head

Nano Banana 2

• + Excellent typography with perfect spelling of 'Caffè Florian' and 'Est. 1720'.
• + Clean, professional vector emblem style with a cohesive vintage design.
• + Accurate representation of all prompt elements including the cloche, steam, and banner.

• − The steam is a bit simplified/abstracted compared to the rest of the detailed linework.

Vidu Q2

• + Successfully captured the requested warm brown and cream tones.
• + Good use of subtle texture on the background.

• − Severe spelling errors including 'FARMIIN', 'Esttt', and 'FopLi20'.
• − The cloche handle is poorly rendered and floating.
• − The steam appears inside the cloche, which is physically illogical for a closed dome.

Nano Banana 2

• + Excellent text rendering with almost no spelling errors.
• + Clean, consistent flat-vector illustration style that perfectly matches the prompt.
• + Follows all six requested steps in a logical, readable layout.

• − Includes a fourth unidentified astronaut silhouette alongside the three named mission members.

Vidu Q2

• + Generally adheres to the color palette and vector aesthetic.
• + Good icon consistency for the lunar modules.

• − Severe spelling and typography errors throughout all labels.
• − Fails to follow the chronological order of the requested steps, resulting in confusing numbering.
• − The Saturn V icon looks like a generic toy rocket rather than the specific NASA vehicle.