Nano Banana Pro vs GPT Image 1.5
Head-to-head across 16 challenges
Nano Banana Pro
44.4%
win rate
Ties
11.1%
GPT Image 1.5
44.4%
win rate
Challenge Results
Fantasy Warrior
Text-to-Image“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”
AI Judge Analysis
Nano Banana Pro
- + Excellent engraving details on the plate armor
- + Very clear and lifelike eyes with realistic reflections
- + Better representation of 'braided hair with small beads' throughout the hairstyle
- − The torch light source on the right looks a bit flat and cut off
- − Some leather strap textures are slightly softer than the metalwork
GPT Image 1.5
- + Beautiful cinematic lighting with a more natural bokeh effect
- + Extreme detail on the skin texture and scars
- + The gold cross and intricate filigree add great visual interest
- − The hair braids are less distinct and partially blended into the background/armor
- − The lighting on the face is slightly too localized, feeling a bit 'studio-lit' despite the setting
Verdict: Both models followed the prompt exceptionally well, producing high-quality cinematic portraits. Gemini 3 Pro Image Preview provides better adherence to the hair bead request and has more crisp armor engravings, while GPT Image 1.5 offers superior skin texture realism and a more dynamic, warm lighting composition.
Candid Street Photography
Text-to-Image“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”
AI Judge Analysis
Nano Banana Pro
- + Excellent photographic quality with a film-like aesthetic and realistic skin textures.
- + Strong adherence to the 'imperfect framing' and 'candid' aspects of the prompt.
- + Atmospheric wet pavement reflections and convincing light rain effects.
- − The bike's construction is slightly physically inconsistent near the pedals.
- − Movement blur on the passing car is fairly subtle.
GPT Image 1.5
- + Successfully captures a shallow depth of field with a very blurred background.
- + Hand and tool details are well-rendered for the task described.
- + Good color contrast and vivid reflections on the wet asphalt.
- − The skin texture appears slightly smoothed compared to the 'natural texture' request.
- − Lacks the specific 'motion blur from passing cars' requested as the car in the background is sharp.
Verdict: Gemini 3 Pro Image Preview provides a much more authentic 'candid street photo' look, effectively capturing the grit and imperfect framing requested. While GPT Image 1.5 produces a clean, high-quality image, it failed to incorporate the motion blur on the cars and feels more like a staged portrait than a candid street snap.
Modern Clean Menu
Text-to-Image“Modern minimalist restaurant menu design, white background with colorful food photos in grid, sections for appetizers/pizza/mains, bold sans-serif fonts, vibrant accents, clean professional layout for casual dining.”
AI Judge Analysis
Nano Banana Pro
- + Excellent visual composition with a modern, artistic layout featuring abstract geometric shapes.
- + The $2 imes 3$ grid of food photos is perfectly aligned and high-quality.
- + Uses a more sophisticated sans-serif font family that feels professional and trendy.
- − The text content is largely gibberish/placeholder text.
- − Repeats item names (e.g., 'Bruschetta' four times) instead of unique menu items.
GPT Image 1.5
- + Excellent text rendering with clear, legible, and relevant English menu items and descriptions.
- + Strong logical organization with clear visual separation between the three requested sections.
- + Accurate pricing and coherent descriptions that match the food categories perfectly.
- − The layout is more traditional and slightly less 'modern minimalist' than requested.
- − The 'grid' of photos is somewhat irregular compared to the requested uniform grid style.
Verdict: Gemini 3 Pro Image Preview produces a much more stylish and professional-looking design mockup with superior composition and aesthetic appeal, though the text is nonsensical. GPT Image 1.5 succeeds in providing functional, readable text and logical menu items, but the overall design feels more like a standard flyer than a modern minimalist brand identity. Gemini is the winner for visual design, while GPT is better for content accuracy.
Man and Car in California
Editing“Make a photo of the man driving the car down the California coastline”
AI Judge Analysis
Nano Banana Pro
- + Excellent preservation of the man's identity, including his specific hairstyle, scarf, and plaid coat.
- + Faithfully maintains the interior and exterior details of the white convertible from the source image.
- + Dynamic composition with a realistic feeling of motion and a beautiful California coastal background.
- − The steering wheel placement is slightly awkward relative to the man's hands.
GPT Image 1.5
- + Strong background accuracy for a California coastal road.
- + Preserves the man's facial features and Hairstyle well.
- − The man's scale and positioning inside the car are anatomically incorrect, making him look too small or deep in the seat.
- − The hand on the steering wheel has significant structural artifacts.
- − Loss of detail on the plaid coat compared to Model A.
Verdict: Gemini 3 Pro Image Preview is the clear winner as it successfully merges the two source images with high fidelity. It perfectly preserves the man's specific clothing (plaid coat and infinity scarf) and the car's interior, whereas GPT Image 1.5 struggles with the man's scale and hand anatomy.
Geometric Composition
Text-to-Image“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”
AI Judge Analysis
Nano Banana Pro
- + Excellent photorealistic texture on the wooden table and antique book
- + Naturalistic lighting with soft window light and realistic refractions
- + Accurate plant visibility through the glass pane
- − The sphere is very small, bordering on being a marble rather than a central subject
GPT Image 1.5
- + Clean, vibrant colors and sharp focus
- + The blue sphere is clearly defined and matches the prompt's focus
- + Good composition with all elements clearly visible
- − Reflections in the glass and sphere are slightly simplified
- − The plant behind the glass lacks the refractive distortion expected in thick glass
Verdict: Both Gemini 3 Pro and GPT Image 1.5 adhered perfectly to the prompt requirements. Gemini 3 Pro is the winner because it achieves a significantly higher level of photorealism, particularly in the weathered texture of the wood and the subtle dust/imperfections on the glass, whereas GPT Image 1.5 has a slightly more rendered, CGI appearance.
Bald man challenge
Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
Nano Banana Pro
- + Perfect preservation of the base image including facial features, glasses, and background.
- + Highly realistic hair texture and natural integration with the existing beard and sideburns.
- + Accurate lighting that matches the outdoor setting of the original photo.
GPT Image 1.5
- + Successfully adds a full head of hair that appears thick and natural.
- + Maintains the overall character and environment of the source image.
- − Substantially alters the subject's facial structure, especially the forehead and brow area.
- − The glasses and eye area look different compared to the original, losing the 'preservation' aspect.
Verdict: Gemini 3 Pro Image Preview performed a flawless edit, seamlessly integrating realistic hair onto the original subject without changing any other pixels in his face or the surroundings. GPT Image 1.5, while providing a good hairstyle, essentially re-generated the man's face, resulting in a different person who merely resembles the original subject and losing the specific details of the glasses and facial expressions.
Night Sky Transformation
Editing“Change the scene to night: a deep, dark sky with subtle, glistening stars visible behind the mountain.”
AI Judge Analysis
Nano Banana Pro
- + Perfect preservation of the town's geometry and lighting.
- + Excellent transition of the sky while maintaining the overall scene composition.
- + Realistic lighting on the mountain peak following the removal of the sunset.
- − The sky feels slightly purple/brown near the horizon compared to a deep black night.
GPT Image 1.5
- + Created a very deep, dark night sky as requested.
- + Realistic star density and Milky Way details.
- − Lost a significant amount of detail in the village and foreground due to extreme darkening.
- − The town lights feel less vibrant and some structures become indistinguishable.
- − The mountain silhouette is slightly less defined against the sky.
Verdict: Gemini 3 Pro Image Preview is the winner because it successfully transitions the scene to night while perfectly preserving all the intricate details of the village and the mountain from the source image. GPT Image 1.5 achieves a darker, more realistic night sky, but it does so by crushing the shadows to the point where much of the foreground detail is lost.
Adorable Baby Animals in Sunny Meadow
Text-to-Image“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”
AI Judge Analysis
Nano Banana Pro
- + Perfectly captures all four requested animals with distinct, clear character designs.
- + Excellent rendering of motion and joy with the running poses.
- + Rich, vibrant colors and clear 'god rays' lighting from the canopy.
- − The eyes look slightly more 'cartoonish' or illustrative than hyper-photorealistic.
- − The butterfly appears somewhat pasted on rather than integrated into the lighting.
GPT Image 1.5
- + Higher level of photorealism in the fur textures and lighting integration.
- + Better capture of the 'tumbling together' aspect of the prompt.
- + Atmospheric lighting with realistic dew sparkles and sun glare.
- − The fox's anatomy is slightly awkward where it blends with the other animals.
- − The kitten's paws have an inconsistent number of toes/pads.
Verdict: Both models followed the prompt exceptionally well, but GPT Image 1.5 achieves a higher level of photorealism and captures the 'tumbling' interaction more naturally. Gemini 3 Pro is brighter and more balanced in its composition, but the animals have an slightly more stylized, 'Disney-like' appearance compared to the requested hyper-photorealism.
Over-the-top cartoon caricature
Editing“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
AI Judge Analysis
Nano Banana Pro
- + Perfectly captures the 'caricature' art style with exaggerated features and line work.
- + Creatively integrates all themes into one cohesive scene, including 'Dog & Hockey News' text.
- + Includes numerous humorous details like dogs in hockey gear and a hockey-themed shirt.
- − The facial resemblance to the source image is slightly lost in the heavy stylization.
- − Some elements, like the many dogs, make the composition feel a bit cluttered.
GPT Image 1.5
- + Maintains a much stronger facial resemblance to the woman in the source image.
- + High visual quality with realistic lighting and texture rendering.
- + Clearly depicts all requested elements including the TV studio, hockey player, and dogs.
- − The style feels more like a '3D render' or 'digital painting' than a traditional caricature.
- − The hockey stick held by the dog looks awkwardly merged with the table.
Verdict: Gemini 3 Pro Image Preview best captures the spirit of the 'caricature' request with an exaggerated, humorous, and stylized illustration that merges all themes perfectly. GPT Image 1.5 produces a more polished, high-fidelity image that preserves the original person's face better, but it feels less like a caricature and more like a standard AI portrait.
Victorian Greenhouse Oasis
Text-to-Image“Hyper-photorealistic interior of a lush Victorian glass greenhouse filled with exotic tropical plants, vibrant blooming orchids, tall ferns, colorful butterflies in flight, sunlight filtering through ornate glass roof creating realistic caustics and dew on leaves, intricate iron framework visible, misty atmosphere, 8K masterpiece.”
AI Judge Analysis
Nano Banana Pro
- + Excellent depiction of a wide variety of tropical plants including detailed ferns and bromeliads.
- + The scale of the Victorian architecture feels grand and accurately reflects the period style.
- + Butterflies are integrated with varying depths, enhancing the sense of space.
- − The lighting feels a bit flat and several butterflies appear to be 'stickers' placed on top of the glass without proper depth cues.
- − The 'misty' effect is localized more to the background glass than the actual air of the room.
GPT Image 1.5
- + Stunning lighting with visible light rays and realistic caustics that create a magical atmosphere.
- + The dew drops on the leaves are rendered with high detail, perfectly matching the prompt.
- + The central composition and the light filtering through the roof create a much stronger focal point.
- − The butterflies are somewhat large and few in number compared to the vast space.
- − The architectural details are slightly more repetitive and less 'grand' in scale than Model A.
Verdict: GPT Image 1.5 is the clear winner due to its superior handling of light and atmosphere; the volumetric rays and dew drops on the foliage make the scene feel much more 'hyper-photorealistic' as requested. While Gemini 3 Pro Image Preview provides a better sense of large-scale architecture and a wider variety of plants, it lacks the depth and sophisticated lighting effects seen in the competing image.
Heroic Super Hero Portrait
Text-to-Image“Hyper-photorealistic full-body portrait of a female superhero standing triumphantly on a New York skyscraper rooftop at golden sunset, wearing a classic modest superhero costume with flowing cape, chest emblem, gloves, and boots in red and blue colors, practical design, short hair, strong determined heroic expression looking into the distance, powerful confident stance with hands on hips and cape billowing dramatically in the wind, detailed urban cityscape background, warm natural sunlight with sharp shadows and fabric highlights, ultra-sharp textures on suit, hair, and concrete, 8K masterpiece, empowering family-friendly style.”
AI Judge Analysis
Nano Banana Pro
- + Excellent composition with a strong cinematic feel and atmospheric depth.
- + Very high level of photorealism in the textures of the rooftop, skin, and fabric.
- + Natural lighting and shadows that integrate the subject perfectly into the environment.
- − Failed to follow the 'hands on hips' instruction, posing with hands at her sides instead.
- − The character feels slightly small within the frame for a 'full-body portrait'.
GPT Image 1.5
- + Perfectly followed the 'hands on hips' pose instruction.
- + Very clear and detailed cityscape featuring recognizable New York architecture.
- + The costume design is vibrant and matches the 'red and blue' requirement well.
- − Large, anatomical errors in the hands, with too many fingers on one glove.
- − The lighting on the subject feels artificial and 'pasted on' compared to the background.
- − The short skirt design leans away from the 'modest' and 'practical' request in the prompt.
Verdict: Gemini 3 Pro Image Preview produces a much more realistic and visually cohesive image with superior lighting and atmospheric quality, though it misses the specific 'hands on hips' pose. GPT Image 1.5 captures the requested pose and iconic Superman-style aesthetic better, but suffers from significant anatomical defects in the hands and a less believable composite between the subject and the background.
Neutral Expression to Genuine Smile
Editing{
"action": "image_edit",
"reference": "uploaded neutral portrait",
"change": "Warm genuine Duchenne smile: lips curved up, slight natural teeth, soft eye crinkles, subtle cheek raise",
"details": "Realistic smiling skin (dimples if present, soft cheek shadows), slightly brighter eyes; keep exact eye shape/color/iris",
"preserve_exact": "Face identity/structure, eyes/nose/lips/eyebrows, hair, skin texture/pores/freckles, makeup, clothing, head pose, background, lighting, shadows, framing",
"no_changes": "No face shape change, no new features, no gaze shift, no hair/clothing/lighting/background edits",
"style": "Ultra-photorealistic 8K portrait, sharp face focus, natural soft lighting, realistic skin glow"
}
AI Judge Analysis
Nano Banana Pro
- + Perfectly preserves original texture and fine details like freckles
- + Subtle, realistic smile that creates natural eye crinkles
- + Excellent adherence to 'preserve exact' constraints for identity and lighting
- − The smile is a bit more closed than requested, showing very little teeth
GPT Image 1.5
- + Accurately captures the 'natural teeth' and 'cheek raise' requested in the prompt
- + Maintains high quality skin texture and facial identity
- + Strong interpretation of a 'warm genuine Duchenne smile'
- − Slightly more change to the overall face shape due to the wider smile
- − Fine freckle detail is slightly softened compared to the source
Verdict: Both models performed exceptionally well, which is rare for such precise editing tasks. Gemini 3 Pro Image Preview offers nearly perfect preservation of the original image's texture and detail, but its smile is very subtle; GPT Image 1.5 followed the specific smile secondary instructions (teeth, cheek raise) much better while still keeping the person's identity remarkably intact.
Studio Ghibli Anime Style
Editing“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
AI Judge Analysis
Nano Banana Pro
- + Excellent preservation of the source image's layout and character poses.
- + Captures the Studio Ghibli line art and flat cel-shading style perfectly.
- + High clarity and crisp details while maintaining a hand-painted texture in the background.
- − The background elements (tram, street lamps) are entirely new additions not present in the original.
- − The facial expressions are slightly softened compared to the intensity of the source image.
GPT Image 1.5
- + Strong application of warm, nostalgic lighting and dreamy pastel tones.
- + Good character similarity while adapting to an anime aesthetic.
- + Background preserves the blurry, generic urban feel of the original more closely than Model A.
- − The image is excessively hazy and blurry, losing too much detail in the foreground characters.
- − The man's hand/arm anatomy is slightly distorted compared to the source.
- − Character faces feel a bit generic and less 'Ghibli' in their specific line work compared to Model A.
Verdict: Gemini 3 Pro Image Preview is the winner for its superior execution of the Ghibli art style, featuring clean lines and hand-painted textures that look like a still from a film. While GPT Image 1.5 captures the warm lighting and mood well, the final result is far too blurry and lacks the distinct character design associated with the requested style. Gemini 3 Pro also does a better job of maintaining the composition and identifiable features of the original people.
Golden Hour Stroll
Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
Nano Banana Pro
- + Excellent preservation of the original person face and dog features.
- + The hair-blowing effect is very natural and follows a logical wind direction.
- + The leaves feel more integrated with the background depth.
- − The orange autumn leaves contrast sharply with the very green summer trees in the background.
GPT Image 1.5
- + Effective application of wind effect to the hair.
- + Good quantity of flying leaves creates an energetic feel.
- + Maintains the overall lighting and color palette well.
- − Noticeable change to the subject's facial features compared to the source image.
- − The dog's face and fur details have been subtly altered/smoothed.
- − The leaves appear somewhat flat and lack motion blur.
Verdict: Gemini 3 Pro Image Preview does a superior job of preserving the identity of the person and the dog from the source image while applying the requested edits. GPT Image 1.5 adds more leaves, but unfortunately re-generates the faces of the subjects, losing the original likeness.
Vintage Cafe Logo
Text-to-Image“Vintage minimalist restaurant logo for "Caffè Florian", retro cloche dome with steam and "Est. 1720" banner, classic typography, warm brown and cream tones, subtle texture on light background, vector emblem style.”
AI Judge Analysis
Nano Banana Pro
- + Perfect text rendering for both the name and the date banner.
- + Accurately followed the 'light background' prompt requirement.
- + Elegant illustration style that feels authentically vintage and vector-like.
- − The steam curls are a bit more ornate than 'minimalist' might suggest.
GPT Image 1.5
- + Strong composition with a nicely integrated cloche and banner.
- + Good use of warm brown and cream tones in the illustration.
- − Failed the 'light background' instruction by using a solid black background.
- − Typography in the main name is slightly irregular and less professional.
- − Texture on the cloche is a bit grainy rather than 'subtle' vector texture.
Verdict: Gemini 3 Pro Image Preview is the clear winner as it followed all instructions, including the light background and precise text rendering. GPT Image 1.5 failed to provide the light background and had less polished typography, making it less suitable for a professional logo concept.
Apollo 11: Journey to Tranquility
Text-to-Image“Create a clean, modern vector infographic poster about the Apollo 11 mission. NASA-inspired palette (navy, white, muted red, light gray). Flat-vector style, crisp lines, consistent iconography, subtle gradients only. Steps (stop at landing): 1. Launch (Saturn Vicon) 2. Earth Orbit (Earth + orbit ring icon) 3. Translunar (trajectory arc icon) 4. Lunar Orbit (Moon + orbit ring icon) 5. Descent (lunar module descending icon) 6. Landing (lunar module on the surface icon) Small supporting elements (minimal text): • Crew strip: three silhouette icons with only last names: Armstrong, Aldrin, Collins. • Landing site marker: Moon pin labeled "Tranquility" only. Layout constraints: generous margins, large readable labels, clean background with subtle stars. Vector-only, print-poster look, high resolution.”
AI Judge Analysis
Nano Banana Pro
- + Excellent narrative flow with a continuous trajectory line connecting all steps.
- + Very clean typography and professional layout that feels like a real educational poster.
- + Higher level of detail in the lunar module and astronaut iconography.
- − The 'Descent' icon is slightly cluttered compared to the other clean vectors.
- − Text on the trajectory line is a bit jumbled.
GPT Image 1.5
- + Strong adherence to the NASA-inspired color palette.
- + Very clean, minimalist flat-vector icons for the Earth and Moon.
- + Good vertical layout for step-by-step reading.
- − The image is cropped at the top and bottom, cutting off text.
- − Lacks a cohesive narrative line connecting the steps unlike Model A.
- − Some repetitive elements like the globe icon for 'Launch' and 'Earth Orbit'.
Verdict: Gemini 3 Pro Image Preview provides a much more cohesive infographic with a clear visual journey (the S-curve trajectory) that actually explains the mission steps. GPT Image 1.5 has nice individual icons but fails on the composition, resulting in a disconnected list that is also awkwardly cropped.
Nano Banana Pro
Gemini 3 Pro with image generation capabilities. Combines advanced reasoning with the ability to generate and edit images.
GPT Image 1.5
OpenAI's state-of-the-art image generation model with better instruction following and adherence to prompts