Gemini 3.1 Flash with image generation capabilities. High-efficiency image generation model with support for text rendering, reference images, search grounding, and thinking mode. The efficient counterpart to Gemini 3 Pro Image.
Settled by community votes across 15 shared challenges, with an AI judge weighing in on each.
Nano Banana 2
28.5
arena score
#1 of 48 in Text-to-Image
Best Text-to-Image right now
Top 2 in Image Editing
Skill signature
· Text-to-Image
GPT Image 2
28.4
arena score
#2 of 48 in Text-to-Image
Top 2 in Text-to-Image
Top 3 in Image Editing
Vote tally
Where the votes landed
Nano Banana 2
0%
win rate
Ties
0%
GPT Image 2
0%
win rate
Shared challenges
15
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent text rendering on the book spine
- + Highly realistic reflections and refractions through the glass
- + Very high photographic detail in the wood grain and foliage
- − The glass structure is slightly inconsistent at the top edges where it meets the book
GPT Image 2
- + Perfect geometric cube shape and alignment
- + Clean, minimalist composition
- + Accurate implementation of soft directional lighting
- − The plant is positioned mostly beside the cube rather than behind it as requested
- − Lacks the fine texture detail found in the competitor
Verdict: Nano Banana 2 produces a superior photographic result with impressive text rendering and complex refractions, though the cube's construction is slightly imperfect. GPT Image 2 creates a cleaner geometric cube but fails to place the plant directly behind the glass, which was a specific requirement for the 'partially visible through glass' prompt instruction.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent cinematic atmosphere with vibrant reflections on wet asphalt.
- + Highly realistic skin texture and apparel detailing on the subject.
- + Very convincing '50mm' aesthetic with beautiful bokeh and background depth.
- − Minor anatomical/mechanical issues with how the wrench interacts with the chain.
- − The wrench tool itself has a slightly warped, AI-generated appearance.
GPT Image 2
- + Good adherence to the 'motion blur from passing cars' instruction.
- + Features authentic looking Japanese signage and a natural street environment.
- + Effective use of shallow depth of field to isolate the subject.
- − The man's hands are poorly rendered, appearing fused and lacking clear digit definition.
- − The lighting is somewhat flat and lacks the 'cinematic' rain effect requested compared to the other model.
- − Significant anatomical errors in the man's forearm and hands.
Verdict: Nano Banana 2 is the superior image, capturing a stunning cinematic quality that perfectly aligns with the 'light rain' and '50mm lens' prompts. GPT Image 2 follows the motion blur instruction well, but fails significantly in rendering the subject's hands and lacks the atmospheric textures found in the competing image.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent adherence to the 'battle-worn' description with realistic dirt and gore textures.
- + Highly detailed engraving on the plate armor including runic script and heraldry.
- + Perfect implementation of bokeh sparks and torchlight atmosphere.
- − The hand gripping the sword hilt has anatomical irregularities in the digit placement.
- − The hair beads are slightly less distinct than the ones in the competing image.
GPT Image 2
- + Incredible facial realism with lifelike eyes and subtle skin textures like freckles.
- + Very clean rendering of fine braids and individual beads.
- + Beautiful lighting and shallow depth of field effects.
- − The character looks less 'battle-worn' and more groomed compared to the prompt requirements.
- − The armor engraving is slightly softer and less defined than in the other image.
Verdict: Both models performed exceptionally well on this complex prompt. Nano Banana 2 captured the 'battle-worn' grit and ornate armor details more effectively, while GPT Image 2 produced a more photorealistic face with cleaner hair rendering. Nano Banana 2 is the winner because it better embodies the specific aesthetic of a hardened paladin requested in the prompt.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent text legibility and alignment
- + Includes all requested sections (appetizers, pizza, mains) clearly labeled
- + High photo quality with a very logical grid layout
- − Minor spelling errors in fine print descriptions
- − The logo design is a bit generic compared to the other
GPT Image 2
- + Fantastic modern aesthetic with high design value
- + Excellent integration of graphic icons and social media handles
- + Near-perfect text rendering and spelling
- − Does not strictly follow a 'grid' layout for photos, opting for a horizontal row approach
- − The 'NOVA' branding takes up a significant amount of vertical space
Verdict: Both models performed exceptionally well, but GPT Image 2 creates a more professional and aesthetically pleasing design that looks like a real-world modern menu. Nano Banana 2 followed the 'grid' instruction more literally, but GPT Image 2 has superior typography and artistic balance.
Magic Burger Explosion: Fiery Photorealism Challenge
Text-to-Image“Ad for 'Magic Burger'. Dynamic, exploded burger with all components (bun, patty, cheese, lettuce, tomato, sauce) suspended in mid-air. Emphasize photorealistic detail and a sense of motion. Dark, fiery background with glowing embers. Integrate text: 'MAGIC BURGER' as a prominent title, 'LIMITED TIME ONLY' as a secondary message, and '€6.99' in a starburst, all rendered with a fiery, glowing effect.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent typography with a professional, clean font layout.
- + Highly realistic texture on the grilled patty and fresh vegetables.
- + Centered, balanced composition that feels like a completed advertisement.
- − The 'starburst' for the price is a bit simple compared to the other fiery effects.
- − The sauce splash is slightly less dynamic than the competition.
GPT Image 2
- + Outstanding fiery effects on all text elements and the starburst border.
- + Very dynamic 'exploded' view with a more extreme angle on the top bun.
- + The price tag integration is creative and matches the overall theme perfectly.
- − The 'LIMITED TIME ONLY' text is slightly less legible due to the heavy fire effects.
- − The lettuce texture looks a bit more plasticky compared to the patty.
Verdict: Both models followed the prompt exceptionally well, producing high-impact, professional-grade ads. Nano Banana 2 has slightly better food realism and a cleaner layout, while GPT Image 2 excels in the 'fiery' branding and dynamic motion of the burger components. Nano Banana 2 is the winner for its superior balance and photorealistic vegetable textures.
Chalkboard Menu
Text-to-Image“Handwritten-style chalkboard menu in a cozy café, all text rendered in the exact same realistic chalk handwriting style with natural variations in letter size, slight slant, and chalk texture — no printed or digital fonts anywhere on the board. Title at the top in elegant cursive chalk handwriting: ‘TODAY’S SPECIALS – APRIL 30, 2026’. Below it, three menu items also in the same handwritten chalk style: ‘Truffle Mushroom Risotto – $24’, ‘Grilled Octopus with Lemon & Herbs – $28’, ‘Brown Butter Chocolate Chip Cookies – $9’. At the very bottom, smaller text in the identical handwritten chalk style (slightly smaller but still clearly legible with the same handwriting characteristics): ‘All items made fresh daily • Ask about our gluten-free options’. Warm ambient café lighting, visible chalk dust and smudges, realistic handwriting imperfections, no clean printed text anywhere.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent text legibility and formatting.
- + Strong chalk texture and smudge effects for realism.
- + Accurately completed the partial text 'Brown But...' with logical menu items.
- − The background bokeh shows some generic AI artifacts in the human distant shapes.
GPT Image 2
- + Very realistic, slightly shaky handwriting style that feels authentic to a real person.
- + Excellent integration of the chalkboard into the physical environment.
- + Natural chalk texture that varies in thickness and pressure.
- − The text is thinner and slightly harder to read than Model A.
- − Limited depth in the background compared to Model A.
Verdict: Both models performed exceptionally well on a difficult text rendering task. Nano Banana 2 produced very clean, bold text that stands out beautifully, while GPT Image 2 achieved a slightly more authentic 'imperfect' handwriting style that feels like it was written by a café employee. Nano Banana 2 is the winner for superior legibility and composition.
Pose & Character Mashup
EditingEdit instruction
“Use Image 1 as the exact pose reference and Image 2 as the character reference. Recreate the person/character from Image 2 in the exact dynamic pose and body position from Image 1. Keep the exact face, hair, clothing style/details, and expression from Image 2. Match the lighting and environment of Image 1. The final image must show the character from Image 2 performing the precise action/pose from Image 1 with perfect anatomy and natural integration.”
Source
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Strong facial likeness to the character in Image 2
- + Accurately replicates the character's clothing and scarf details
- − The pose is upright and fails to replicate the extreme horizontal torso tilt from Image 1
- − The lighting on the character's back doesn't match the studio lighting of the background
GPT Image 2
- + Excellent adherence to the extreme dynamic pose and torso angle of Image 1
- + Perfectly matches the lighting and environment of the source image
- + Successfully integrates all character elements (sunglasses, scarf, hair) into the complex pose
- − Slight distortion in the perspective of the lower hand
- − The face is slightly less sharp than in Model A
Verdict: GPT Image 2 is the clear winner as it successfully recreated the difficult, dynamic pose and body position from Image 1 while maintaining character consistency. Nano Banana 2 failed to match the core instruction of the pose, instead presenting the character standing relatively upright which does not reflect the source image's energy.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent cinematic lighting and color palette
- + High level of detail in the nebula and Earth background
- + Dynamic composition with a sense of motion
- − The horse is floating above rather than 'riding' the astronaut
- − The reins are not being held by anything, and the horse's anatomy is slightly elongated
GPT Image 2
- + Strong adherence to the literal 'riding' instruction with the horse positioned on top
- + Interesting surreal interpretation of the astronaut acting as a mount
- + Consistent texture between the moon surface and the astronaut suit
- − Anatomical issues where the horse's legs merge into the astronaut's shoulders
- − Static and less 'cinematic' composition compared to the other model
- − The background stars look like a flat texture
Verdict: GPT Image 2 followed the specific 'horse on top' instruction much more literally by depicting the horse actually saddled onto the astronaut's back, whereas Nano Banana 2 simply showed them floating near each other. However, Nano Banana 2 has significantly better visual quality, lighting, and a more professional cinematic aesthetic, while GPT Image 2 suffers from awkward anatomical merging where the horse meets the astronaut.
Outfit Transfer Challenge
EditingEdit instruction
“Use Image 1 as the base person. Dress them in the exact elaborate outfit from Image 2 (including all layers, accessories, jewelry, and shoes). Carefully adapt the clothing to the body shape and pose in Image 1 while maintaining realistic fabric behavior, correct proportions, and perfect lighting/shadow matching. Keep the person’s exact face, hair, and background completely unchanged.”
Source
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Successfully preserved the subject's vitiligo patterns on the hands and fingers.
- + Accurately included the sunglasses from the second image.
- + Maintained the exact skin textures and sand details on the face from Image 1.
- − The head and neck proportions seem slightly disconnected from the body.
- − The sunglasses are a bit larger and more opaque than in the source image.
GPT Image 2
- + Better anatomical integration of the neck and shoulders with the heavy coat.
- + High fidelity in replicating the scarf and coat details from Image 2.
- + Maintains the lighting and sharp focus of the background and wooden structure.
- − Failed to preserve the vitiligo pattern on the hands, turning them into a solid skin tone.
- − Missed the sunglasses entirely despite the instruction to include all accessories.
- − The subject's pose was slightly altered, losing the lean against the post.
Verdict: Nano Banana 2 was the more successful model because it adhered to the specific detail of preserving the subject's vitiligo on their hands and including all accessories like the sunglasses. GPT Image 2 produced a cleaner anatomical composite but failed to maintain the unique physical characteristics of the person and missed required items from the outfit.
The Capybara Taxi Driver
Text-to-Image“Photorealistic scene inside a yellow New York taxi at night. A capybara is driving, wearing a yellow taxi driver cap and a dark jacket. It has a calm, professional expression and both front paws on the steering wheel. In the back seat sits a human businesswoman in a coat, looking at her phone with a completely normal, bored expression (as if this is just another normal ride). Through the windows you can see the streets of Manhattan at night with blurred lights. Realistic taxi interior, photorealistic, detailed fur and fabric, 35mm lens, night lighting with reflections, shallow depth of field.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent attention to detail with the inclusion of a 'Captain Cappy' hack license.
- + Incredible sharpness and composition showing the full interior of the taxi.
- + Superior text rendering on the fare meter and interior signage.
- − The driver cap is black/navy rather than the requested yellow.
GPT Image 2
- + Correctly followed the color instruction for the yellow taxi driver cap.
- + The capybara's expression and posture perfectly match the 'professional' deskripter.
- + High cinematic quality with natural lighting from the city street.
- − Composition is tighter, losing the interesting foreground taxi details seen in the other model.
- − The passenger is slightly more out of focus.
Verdict: Nano Banana 2 provided a much more creative and detailed scene by including a personalized hack license and a clear view of the taxi's dashboard, though it missed the specific color for the cap. GPT Image 2 followed the color prompts more accurately but felt like a more standard close-up. Nano Banana 2 is the winner for its environmental storytelling and high-quality detail.
The Halloween Invitation
Text-to-Image“Vintage gothic Halloween party invitation. Dark parchment poster, spooky border with webs and thorns, central glowing jack-o-lantern, bats, twisted trees, moody night sky. Add elegant gothic title text saying "Halloween Party Invitation", a small scroll banner saying "You are invited to a night of frights", and event details at the bottom: Date: 30.10.2026 Time: 7pm Location: The Arches, NYC Spooky but polished, cinematic lighting, square format.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent text legibility and accuracy for all requested fields
- + Clean and professional-looking gothic border with integrated spiderwebs
- + Strong cinematic lighting with a vibrant glowing jack-o'-lantern
- − Composition is a bit generic compared to Model B
- − The night sky background lacks complex atmospheric depth
GPT Image 2
- + Highly detailed 'vintage parchment' texture with rich, gritty details
- + Very creative interpretation of 'The Arches' and the NYC skyline in the background
- + Sublime gothic typography for the main title
- − Small text at the bottom is less legible due to the dark, busy background
- − The scroll banner is slightly warped
Verdict: Both models followed the prompt exceptionally well, but GPT Image 2 edges ahead due to its superior artistic depth and the creative inclusion of the 'The Arches' bridge and NYC skyline in the background. Nano Banana 2 is a very strong contender with perfect text rendering, but GPT Image 2 captures the 'vintage' and 'moody' atmosphere with more complexity.
Isometric Miniature Diorama Scenes
Text-to-Image“Create a clear, 45° top-down isometric miniature 3D cartoon scene of Japan's signature dish: sushi, with soft refined textures, realistic PBR materials, gentle lighting, on a small raised diorama base with minimal garnish and plate. Solid light blue background. At top-center: 'JAPAN' in large bold text, 'SUSHI' below it, small flag icon. Perfectly centered, ultra-clean, high-clarity, square format.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent high-clarity text rendering
- + Very realistic PBR material textures on the wood and fish
- + Clean and sophisticated layout and lighting
- − The diorama base is more of a plate than a miniature landscape
- − The text layout is slightly off-center with the flag position
GPT Image 2
- + Stronger 'miniature diorama' feel with the inclusion of the stone lantern and leaves
- + Better 3D/cartoon aesthetic for the text to match the prompt
- + Superior source preservation of the 45 degree isometric perspective
- − The text is a bit crowded vertically
- − Micro-texturing on the sushi rice looks slightly more plastic/less detailed than Model A
Verdict: Nano Banana 2 produces a very clean and professional image with superior material realism, but GPT Image 2 better captures the 'miniature 3D cartoon scene' and 'diorama' aspects of the prompt by incorporating thematic elements like the stone lantern and a more defined base. While Nano Banana 2 has cleaner text, GPT Image 2 feels more like the requested isometric miniature art style.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent depiction of god rays and sunrise lighting
- + Sharp focus across all animals and butterflies
- + Vibrant and diverse wildflower meadow with clear dew sparkles
- − The fox kit's proportions are slightly off, making it look a bit more like a mature fox than a baby
- − Butterflies look a bit flat and pasted on
GPT Image 2
- + The animals are grouped more tightly, better following the 'tumbling together' prompt
- + The fox kit has a very realistic, baby-like appearance and expression
- + Stronger sense of motion and playfulness in the poses
- − The kitten has five paws or an anatomical error in how it is reaching out
- − Some blurring on the butterflies and background elements is a bit heavy-handed
Verdict: Both models captured the essence of the prompt well, but Nano Banana 2 provided a much cleaner and technically sound image with superior lighting and atmospheric effects. While GPT Image 2 had a more dynamic and 'wholesome' tumbling composition, the anatomical issues with the kitten and the slightly muddy background make Nano Banana 2 the better choice.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent curved text integration
- + Clean circular badge composition
- + Accurate rendering of the requested 'Est. 1720' banner
- − The steam lines are somewhat simplified
- − The 'F' in Florian is slightly oversized compared to other letters
GPT Image 2
- + Elegant and detailed engraving-style cross-hatching
- + Sophisticated frame design with high visual appeal
- + Clearer and more artistic steam rendering
- − The word 'FLORIAN' has inconsistent kerning and slight letter distortion
- − The banner texture is a bit muddy compared to Model A
Verdict: Both models followed the prompt exceptionally well, producing high-quality vintage logos with accurate text. Nano Banana 2 (Image A) is the winner as a functional logo due to its superior vector-like clarity and more balanced typography, whereas GPT Image 2 (Image B) exhibits slight letterform warping despite its beautiful artistic textures.
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.”
Nano Banana 2
GPT Image 2
AI Judge Analysis
Nano Banana 2
- + Excellent adherence to the 'flat-vector' style and consistent iconography.
- + Highly legible text with professional, modern typography.
- + Clean and balanced composition that feels like a real infographic.
- − One of the crew silhouettes is white/unfilled, which slightly breaks consistency.
- − The 'Translunar' icon is a bit abstract compared to others.
GPT Image 2
- + Successfully incorporates all six requested steps in a clear chronological progression.
- + High level of detail in the lunar module and crew patch graphics.
- + Includes several additional thematic elements like the NASA meatball and mission logo.
- − The style leans more toward '3D render' or 'illustration' than the requested 'flat-vector'.
- − The bottom-right section with the pin has some slightly cluttered line work.
- − Text rendering on the 'Translunar Trajectory' is a bit smaller and harder to read.
Verdict: Nano Banana 2 followed the stylistic constraints much better, delivering a truly clean, flat-vector infographic with a professional color palette. GPT Image 2 provided a more comprehensive layout with more accurate NASA branding, but it missed the 'flat' aesthetic, opting for more detailed, shadowed illustrations. Nano Banana 2 is the winner for superior design consistency and adherence to the requested art style.
Next steps
Explore each model
OpenAI's state-of-the-art image generation model with arbitrary resolution up to 4K and strong instruction following