Best Nano Banana Prompts: 30 Examples for Stunning AI Images (2026)

BLUF (Bottom Line Up Front)

Nano Banana prompts follow a specific structure that most beginners get wrong. The difference between a mediocre AI image and a stunning one comes down to six elements: subject, composition, action, environment, style, and exclusions. This guide gives you 30 field-tested prompts you can copy directly into Google AI Studio, Flowith, or any platform running Nano Banana models — plus the underlying formula so you can write your own. Nano Banana 2 generates images in 4-6 seconds at near-Pro quality, while Nano Banana Pro takes 10-20 seconds but delivers the absolute highest fidelity for hero images and print-ready assets.

Key Takeaways

• In one sentence: Nano Banana is Google’s AI image generation family — spanning Flash, Standard, and Pro tiers — and these 30 prompts cover every use case from blog heroes to cinematic scenes, each tested and ready to copy-paste.
• Key number: Over 200 million image edits have been processed across the Nano Banana family, with Nano Banana 2 generating images in 4-6 seconds compared to Pro’s 10-20 seconds — a 3-5x speed advantage at roughly 95% of Pro’s quality.
• Why it matters: The prompt structure matters more than the model you choose. A well-structured Flash prompt consistently outperforms a vague Pro prompt, which means mastering the 7-part formula saves both time and money.
• What to do next: Pick one prompt from each category below, paste it into Google AI Studio or Flowith, and generate your first image in under 10 seconds. Then use the prompt formula in Section 3 to customize it for your specific project.
• Related reading: Nano Banana AI Guide, AI Image Generation Guide, How to Write AI Prompts

The Nano Banana Prompt Structure Formula

Before diving into the 30 prompts, you need to understand the structure that makes them work. According to Google Cloud’s official prompting guide, Nano Banana 2 uses a reasoning layer from Gemini 3.1 to parse spatial relationships, lighting logic, and compositional intent before generating a single pixel. Short, vague prompts bypass that reasoning engine almost entirely — which is why “a beautiful sunset” produces generic results while a structured prompt produces magazine-quality output.

The 7-part formula that drives every prompt in this guide:

1. Subject — Who or what is the main focus? (“A 35-year-old female architect,” “A weathered leather journal,” “A robotic hummingbird”)
2. Action/Pose — What is the subject doing? (“reviewing blueprints at a drafting table,” “lying open on a rain-soaked park bench,” “hovering mid-flight with wings blurred”)
3. Environment — Where does this take place? (“inside a glass-walled modern office overlooking Tokyo at dusk,” “Central Park in autumn with golden leaves scattered,” “a bioluminescent underwater cave”)
4. Camera/Composition — How is the shot framed? (“shot on 85mm f/1.4 lens, shallow depth of field,” “wide-angle 24mm establishing shot,” “overhead flat lay composition”)
5. Lighting — What creates the mood? (“warm golden hour side lighting,” “dramatic Rembrandt lighting with a single key light,” “cool blue neon reflected on wet pavement”)
6. Style — What’s the visual aesthetic? (“editorial photography for Architectural Digest,” “Wes Anderson color palette with symmetrical framing,” “dark academia oil painting”)
7. Exclusions — What should the model avoid? (“no text overlays, no watermarks, no extra fingers,” “avoid cartoonish rendering, no oversaturated colors”)

This formula works because Nano Banana’s reasoning layer processes each element sequentially. Google’s own documentation confirms the model can track up to 5 characters and 14 objects within a single workflow — but only when you give it clear spatial and compositional instructions. Skip any element, and you’re leaving quality on the table.

Which Nano Banana Model Should You Use?

Before you start prompting, here’s the decision framework for choosing between models — because the right model depends on what you’re creating, not just what you can afford:

Nano Banana 2 (Flash) — Best for Speed and Volume

Built on Gemini 3.1 Flash Image, released February 2026. Generates images in 4-6 seconds. Achieves approximately 95% of Pro’s image quality in most scenarios. Supports resolutions up to 4K and all standard aspect ratios (1:1, 3:2, 2:3, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9) plus extended ratios like 1:4, 4:1, 1:8, and 8:1. Best for social media content, blog images, rapid iteration, and bulk generation. According to Grokipedia, the Flash architecture optimizes the inference path to boost generation speed by 3-5x while maintaining quality close to the Pro version.

Nano Banana Pro — Best for Quality-Critical Work

Built on Gemini 3 Pro Image, released November 2025. Takes 10-20 seconds per image. Delivers the absolute highest fidelity — richer textures, more natural lighting and shadows, superior spatial composition. Best for hero images, print campaigns, brand assets, and any deliverable where quality ceiling matters more than speed. Pro handles complex multi-element compositions with strict spatial relationships better than any other model in the family.

Nano Banana Standard (Original) — The Middle Ground

Based on Gemini 2.5 Flash Image, launched August 2025. The original model that started the Nano Banana phenomenon. Still capable for general-purpose image generation, though Nano Banana 2 has largely superseded it in speed and quality. Useful if you’re working with platforms that haven’t yet integrated the newer models.

Throughout the 30 prompts below, each one includes a model recommendation based on extensive testing. The pattern is straightforward: use Flash for anything you need fast, use Pro for anything that needs to be perfect. For a deeper look at the Nano Banana ecosystem and how it fits into the broader AI image generation landscape, see our complete guide.

Category 1: Editorial and Blog Hero Images (5 Prompts)

Blog hero images need to stop the scroll, communicate a topic instantly, and look professional at 1200x630px or wider. These prompts are designed for content creators, bloggers, and marketing teams who need publication-ready visuals without a photographer.

Prompt 1: AI Technology Editorial Hero

The prompt:

A minimalist editorial photograph of a human hand reaching toward a translucent holographic AI neural network floating in mid-air, set against a clean white studio background with subtle warm gradient lighting from the left. Shot on a 50mm lens at f/2.8 with shallow depth of field — the hand is tack-sharp while the holographic elements have a soft ethereal glow. Style: Apple product launch aesthetic meets MIT Technology Review cover. Color palette restricted to white, soft gold, and electric blue accents. 16:9 aspect ratio. No text, no watermarks, no extra fingers, no cluttered background elements.

What it creates: A clean, professional hero image perfect for articles about AI technology, machine learning concepts, or human-AI interaction. The minimalist aesthetic works across light and dark website themes.

Customization tips: Swap “AI neural network” for any tech concept — “blockchain nodes,” “quantum computing qubits,” “data visualization dashboard.” Change the color accents to match your brand. Replace “Apple product launch” with “Bloomberg Businessweek” for a more corporate feel.

Best model: Nano Banana Pro. The holographic translucency and precise lighting control benefit from Pro’s superior rendering of light interaction and fine detail.

Prompt 2: Productivity and Workflow Hero

The prompt:

An overhead flat-lay photograph of a modern workspace — a MacBook Pro with code on the screen, a ceramic pour-over coffee in a handmade mug, a Moleskine notebook with hand-drawn wireframes, wireless earbuds in their case, and a small succulent plant. All items arranged on a light oak desk with deliberate negative space between them. Soft diffused natural light from a window on the upper right. Shot from directly above with a 35mm lens. Style: Kinfolk magazine editorial photography. Warm neutral tones — cream, walnut, sage green, matte black. 16:9 aspect ratio. No hands visible, no faces, no brand logos, no text overlays.

What it creates: A versatile workspace flat-lay that works as a hero for productivity articles, tool reviews, remote work guides, or “best apps” roundups. The deliberate negative space leaves room for text overlay in design tools.

Customization tips: Add or remove items to match your article topic. For a finance article, swap the notebook for a calculator and financial reports. For a prompt engineering guide, replace items with printed prompt templates and colored sticky notes. Changing “Kinfolk” to “Monocle” shifts the vibe from cozy to polished.

Best model: Nano Banana 2 (Flash). Flat-lay compositions with defined objects are Flash’s sweet spot — the model handles the spatial arrangement cleanly, and the 4-6 second generation time lets you iterate quickly on item placement.

Prompt 3: Finance and Business Hero

The prompt:

A confident female CEO in her late 40s standing at a floor-to-ceiling glass window of a corner office on the 40th floor, overlooking a city skyline at golden hour. She is wearing a tailored navy blazer and holding a tablet displaying an upward-trending chart. Her reflection is visible in the glass. Shot from a low angle at 35mm, emphasizing power and stature. Warm amber light from the sunset floods the room, creating long shadows on the marble floor. Style: Forbes magazine portrait photography. Rich, authoritative color palette — navy, gold, warm amber, charcoal. 16:9 aspect ratio. No other people, no distracting background elements, no text on the tablet screen.

What it creates: A compelling editorial-style hero for business, leadership, finance, or entrepreneurship content. The golden hour lighting and low angle convey authority and success without feeling stock-photo generic.

Customization tips: Change the setting from a corporate office to a co-working space for startup content, or to a factory floor for manufacturing articles. Adjust the subject’s age, gender, and attire to match your audience. Swap “Forbes” for “Fast Company” to shift from traditional corporate to innovation-focused.

Best model: Nano Banana Pro. The reflection in the glass, the precise lighting interaction with the marble floor, and the detailed facial features all demand Pro’s superior rendering capabilities.

Prompt 4: Creative and Design Hero

The prompt:

A surreal double-exposure photograph blending a human silhouette profile with a vibrant watercolor galaxy — swirling nebulas in magenta, teal, and gold fill the figure's outline while the background remains pure matte black. Small geometric golden particles float around the edges of the silhouette. Shot as a fine art portrait at 85mm. Style: James Jean meets National Geographic astrophotography. High contrast, jewel-tone colors against absolute black. 16:9 aspect ratio. No face details visible inside the silhouette, no text, no harsh edges between the silhouette and the galaxy fill.

What it creates: A striking, artistic hero image for creativity, design thinking, innovation, or “big picture” concept articles. The double-exposure technique gives it an editorial art feel that stands out in social feeds.

Customization tips: Replace “watercolor galaxy” with “a dense forest canopy” for nature topics, “circuit board patterns” for tech, or “ocean waves” for mindfulness content. The silhouette-fill technique works with almost any subject matter. Change the background from black to deep navy for a softer feel.

Best model: Nano Banana Pro. Double-exposure effects require precise blending of two distinct visual layers, and Pro’s superior compositional reasoning handles the silhouette-fill boundary much more cleanly than Flash.

Prompt 5: Education and Learning Hero

The prompt:

A warm, inviting photograph of a diverse group of three young professionals (ages 25-35) collaborating around a large digital whiteboard in a bright, modern learning space. One person is sketching a mind map with a stylus, another is pointing at a specific node, and the third is taking notes on a laptop. The whiteboard displays colorful interconnected concept bubbles. Soft overhead diffused lighting with accent light from large windows on the left. Shot at 35mm f/4, wide enough to capture the full scene with everyone in focus. Style: WeWork brand photography meets TED conference aesthetic. Warm, optimistic palette — soft white, warm wood, pops of coral and teal. 16:9 aspect ratio. No stock-photo-style fake smiles, natural candid expressions, no visible brand names.

What it creates: A genuine-feeling collaborative learning scene for education, online course, tutorial, or team training content. The candid instruction (“no fake smiles”) pushes Nano Banana toward more natural expressions.

Customization tips: Reduce from three people to one for a solo-learner angle. Change the whiteboard content description to match your specific topic. Swap the modern learning space for a library, coffee shop, or home office. Adding “natural candid expressions, mid-conversation” further reduces the AI-generated stiffness.

Best model: Nano Banana Pro. Multiple people with specific poses, natural expressions, and interactive elements (pointing, sketching) is the exact scenario where Pro’s character consistency tracking outperforms Flash.

Category 2: Photorealistic Portraits (5 Prompts)

Portrait prompts demand the highest precision because humans instantly spot anything unnatural about faces, hands, and skin. These prompts use specific photography terminology — lens focal length, aperture, and lighting setups — because Nano Banana’s training data includes millions of captioned photographs where those terms correlate with specific visual qualities.

Prompt 6: Professional Headshot

The prompt:

A professional headshot of a 30-year-old South Asian male software engineer with short black hair, a neatly trimmed beard, and warm brown eyes. He is wearing a charcoal crew-neck sweater over a white collared shirt. Expression: confident half-smile, direct eye contact with the camera. Shot on an 85mm f/1.8 lens with creamy bokeh background in a muted sage green. Butterfly lighting setup with a silver reflector below to fill chin shadows. Natural skin texture with visible pores — no airbrushing, no plastic skin effect. Style: LinkedIn professional headshot by Peter Hurley. 4:5 aspect ratio. No jewelry, no visible ears, no busy background patterns.

What it creates: A believable, high-quality professional headshot suitable for LinkedIn profiles, team pages, or author bios. The Peter Hurley reference activates Nano Banana’s understanding of professional portrait lighting.

Customization tips: Change demographics, clothing, and expression to match your needs. The key detail is “natural skin texture with visible pores” — this single instruction dramatically reduces the plastic AI look. Swap the background color to match brand guidelines. Replace “butterfly lighting” with “Rembrandt lighting” for a more dramatic, editorial feel.

Best model: Nano Banana Pro. Portrait quality is where Pro’s advantage is most visible — skin texture, eye reflections, and hair detail are noticeably better.

Prompt 7: Environmental Portrait

The prompt:

An environmental portrait of a female ceramics artist in her mid-50s, standing in her sunlit pottery studio surrounded by shelves of unglazed terracotta pieces. She is wearing a clay-stained linen apron over a black turtleneck, holding a freshly thrown bowl. Her silver-streaked hair is pulled back in a loose bun. Expression: serene focus, slight crow's feet from years of smiling. The studio has large skylights casting diffused natural overhead light. Shot on a 50mm f/2.0 lens, subject sharp, background slightly soft. Style: Annie Leibovitz environmental portraiture for Vanity Fair. Earthy palette — terracotta, cream, warm gray, matte black. 3:2 aspect ratio. No modern electronics visible, no perfect skin, natural aging details preserved.

What it creates: A character-rich environmental portrait that tells a story about the subject’s craft. The specific aging details (“crow’s feet,” “silver-streaked hair”) and environment details create authenticity that generic portrait prompts miss.

Customization tips: Change the craft to match your article — a woodworker in their workshop, a chef in their kitchen, a programmer at their standing desk. The environmental portrait format works for any profession article. The key is specifying real environmental details (shelves of unglazed pieces, clay-stained apron) rather than generic workspace descriptions.

Best model: Nano Banana Pro. The combination of character detail, environmental complexity, and naturalistic lighting makes this a Pro-tier prompt.

Prompt 8: Dramatic Low-Key Portrait

The prompt:

A dramatic low-key portrait of an elderly Japanese man with deep wrinkles and kind eyes, lit by a single warm tungsten key light positioned 45 degrees to the right and slightly above. The left side of his face falls into deep shadow. He is wearing a traditional indigo-dyed cotton jacket (sashiko stitching visible). Background is pure black with no visible edges or surfaces. Shot on a 135mm f/2.0 telephoto lens for compressed, intimate framing. Style: Yousuf Karsh portrait photography with the tonal depth of a Caravaggio painting. Monochromatic warm tones — amber, deep brown, charcoal black. 4:5 aspect ratio. No smile, contemplative expression, no distracting elements, no color cast on the shadows.

What it creates: A museum-quality low-key portrait with dramatic chiaroscuro lighting. The Yousuf Karsh and Caravaggio references activate two of the strongest portrait lighting aesthetics in Nano Banana’s training data.

Customization tips: This lighting setup works for any subject — change demographics and clothing while keeping the single-key-light instruction. For a more modern feel, replace the tungsten light with “a single strip softbox” and change the style reference to “Peter Lindbergh.” Adding “catchlight visible in both eyes” ensures the eyes stay alive in the darkness.

Best model: Nano Banana Pro. Low-key lighting demands precise shadow rendering and subtle tonal gradations that Flash can approximate but Pro nails consistently.

Prompt 9: Casual Lifestyle Portrait

The prompt:

A candid lifestyle portrait of a 28-year-old Black woman laughing mid-conversation at a sidewalk cafe in Paris. She's wearing an oversized cream linen blazer, gold hoop earrings, and her natural curls are caught in a light breeze. One hand is wrapped around a small espresso cup, the other is mid-gesture. Shallow depth of field at 85mm f/1.4 — she's sharp, the cafe background is a soft blur of warm light and passing pedestrians. Late afternoon golden hour light from camera left. Style: street photography by Brandon Stanton (Humans of New York) meets Parisian editorial fashion. Warm, romantic palette — golden light, cream, soft brown, blush. 4:5 aspect ratio. Genuine laugh with eyes crinkled, not a posed smile, no visible teeth alignment issues, hands naturally positioned.

What it creates: A warm, authentic-feeling lifestyle portrait that avoids the stiff posed quality of most AI portraits. The “mid-conversation” and “genuine laugh” instructions push the model toward more natural human expressions.

Customization tips: Change the location and wardrobe for different vibes — a Brooklyn coffee shop, a Tokyo ramen bar, a London bookshop. The key instruction is “mid-gesture” — it tells the model to generate hands in a natural, dynamic position rather than resting, which reduces the common AI problem of awkward static hands.

Best model: Nano Banana 2 (Flash). Lifestyle portraits with bokeh backgrounds are well within Flash’s capabilities, and the faster generation lets you iterate on expression and hand positioning until you get the right candid feel.

Prompt 10: Conceptual Portrait

The prompt:

A conceptual portrait of a young woman with her eyes closed, face tilted slightly upward, as hundreds of tiny glowing golden butterflies emerge from the top of her head and disperse into a dark teal background. The butterflies transition from solid near her head to increasingly translucent and scattered at the edges of the frame. Her skin has a soft luminous quality as if lit from within. Shot at 85mm with the face in sharp focus and butterflies progressively softer. Style: a blend of Erik Johansson's surrealism and Gregory Crewdson's cinematic staging. Rich jewel tones — gold, deep teal, warm ivory skin. 4:5 aspect ratio. No visible hair — the butterflies replace it entirely, no harsh light sources, no distracting background elements beyond the gradient.

What it creates: A surreal, emotionally evocative portrait that blends realism with fantasy. The transitioning opacity of the butterflies (solid to translucent) creates a beautiful depth effect that elevates this beyond typical AI art.

Customization tips: Replace butterflies with any symbolic element — “glowing paper origami cranes,” “floating musical notes made of light,” “fragments of old photographs.” The framework is: realistic subject + surreal element emerging from a specific body part + opacity gradient. This template works for virtually any conceptual theme.

Best model: Nano Banana Pro. The progressive opacity gradient, luminous skin quality, and precise boundary between realistic face and surreal elements all require Pro’s advanced compositional reasoning.

Category 3: Product Photography (3 Prompts)

Product photography prompts are among the most commercially valuable AI image use cases. E-commerce businesses generating 500+ product images would spend $50,000-$100,000 on traditional photography — Nano Banana 2’s Flash-tier pricing makes bulk generation practical while you reserve Pro for hero shots. These prompts focus on the three most common product photography styles.

Prompt 11: Hero Product Shot (Cosmetics)

The prompt:

A luxury cosmetics product photograph of a frosted glass serum bottle with a gold dropper cap, positioned at a slight angle on a slab of raw white marble. A single drop of golden serum is suspended mid-fall from the dropper. Behind the bottle, two fresh eucalyptus sprigs and three small river stones create an asymmetric arrangement. Clean white infinity curve background. Three-point studio lighting: a large softbox above for even illumination, a strip light from the right for edge definition on the glass, and a silver reflector on the left for fill. Shot at 100mm macro f/5.6 for maximum product sharpness. Style: Glossier product photography meets Aesop brand aesthetic. Muted natural palette — frosted glass, gold, white marble, sage green, river stone gray. 1:1 aspect ratio. No visible reflections of the studio, no fingerprints on glass, no text on the bottle label.

What it creates: A premium product hero shot suitable for e-commerce listings, social media ads, or brand lookbooks. The suspended serum drop adds a dynamic element that static product shots lack.

Customization tips: Replace the serum bottle with any product — a watch, a perfume bottle, a tech gadget. The three-point lighting setup description is transferable. Change the styling props (marble, eucalyptus, stones) to match brand personality — dark wood and leather for masculine products, pink quartz and dried flowers for feminine.

Best model: Nano Banana Pro. The suspended droplet, glass transparency, and precise lighting reflections on the marble are all details where Pro’s quality advantage is clearly visible.

Prompt 12: Lifestyle Product Context Shot

The prompt:

A lifestyle context photograph of a pair of white leather wireless headphones resting on the arm of a mid-century modern walnut lounge chair. On the side table next to the chair: an open hardcover book face-down, a hand-thrown ceramic cup of matcha latte, and a small potted monstera plant. The room has white walls, a jute rug visible on the floor, and late morning sunlight streaming through sheer linen curtains creating soft shadow patterns. Shot at 35mm f/2.8 from a slight overhead angle. Style: Sonos brand photography meets Architectural Digest interior feature. Warm Scandinavian palette — white, walnut, sage, cream, natural jute. 16:9 aspect ratio. No people visible, no electronics other than the headphones, no visible cords or cables, no brand markings.

What it creates: A lifestyle product shot that tells a story about the customer’s world rather than just showing the product. This style converts better for premium products because it sells the aspiration, not just the item.

Customization tips: Swap the headphones for any product and adjust the room styling accordingly. A fitness tracker might sit on a yoga mat next to a water bottle and plants. A smart speaker might sit on a kitchen counter with fresh herbs and a cookbook. The framework: product + curated environment + lifestyle props + specific light source.

Best model: Nano Banana 2 (Flash). Interior lifestyle shots with defined objects and natural lighting are a strength of Flash, and the speed advantage matters when you’re generating multiple angle variations.

Prompt 13: Food Product Photography

The prompt:

An artisan food photograph of a sourdough loaf sliced to reveal its open, irregular crumb structure, placed on a rustic wooden cutting board with a serrated bread knife beside it. Scattered around: a small dish of hand-churned butter with visible salt crystals, a linen napkin bunched naturally, and a few wheat stalks laid diagonally. Warm side lighting from the left simulating morning kitchen light through a window, with a slight backlight creating a glow on the bread's crust edges. Shot at 50mm f/2.8 from a 45-degree angle. Style: Tartine Bakery cookbook photography by Gentl & Hyers. Rustic warm palette — golden crust brown, cream crumb, weathered wood, natural linen, soft wheat gold. 4:3 aspect ratio. No perfect uniform slices — the bread should look hand-torn and artisan, no plastic packaging, no modern kitchen appliances visible.

What it creates: A mouth-watering artisan food shot that feels authentic and hand-crafted. The specific instruction about “open, irregular crumb structure” and “hand-torn” prevents the over-perfected AI food aesthetic.

Customization tips: Swap the bread for any food product. The lighting formula (warm side light + slight backlight for edge glow) works universally for food photography. For restaurant content, add steam or condensation. For a cocktail, add ice with visible clarity and bubbles. The Gentl & Hyers reference can be replaced with “David Chang’s cookbooks by Alex Lau” for more vibrant, modern food styling.

Best model: Nano Banana 2 (Flash). Food photography without people or complex reflections is well-handled by Flash, and you’ll want the speed to iterate on composition and angle.

Category 4: Whimsical and Fantasy Art (5 Prompts)

Fantasy and whimsical prompts give Nano Banana the freedom to push beyond photorealism into imaginative territory. These are ideal for children’s content, creative branding, social media graphics, and editorial illustrations that need to stand out.

Prompt 14: Miniature World Scene

The prompt:

A whimsical miniature world inside an open antique pocket watch — tiny glowing mushroom houses line a winding cobblestone path, miniature lanterns hang from delicate wire arches, and a small waterfall cascades over the watch's internal gears into a tiny crystal-clear pool. Moss covers the watch mechanism like forest floor. A single tiny fox with a glowing tail sits at the edge of the pool. The watch rests on a bed of dried autumn leaves on a wooden table. Macro lens perspective at f/4, watch interior sharp, background soft. Style: a blend of Wes Anderson's miniature sets and Studio Ghibli environmental detail. Warm fairytale palette — amber, emerald, copper, candlelight gold. 1:1 aspect ratio. No human figures, no modern elements inside the watch world, the transition between real watch and miniature world should feel seamless.

What it creates: A deeply detailed miniature world scene that invites viewers to zoom in and discover details. This type of image performs extremely well on Instagram and Pinterest where users engage with intricate visual storytelling.

Customization tips: Change the container from a pocket watch to a teacup, a vintage camera, a snow globe, or an old book. The miniature-world-inside-an-object framework is infinitely adaptable. Swap the fairytale theme for sci-fi (tiny space station inside a lightbulb) or ocean (coral reef inside a seashell).

Best model: Nano Banana Pro. The seamless transition between real-world container and fantasy interior, plus the micro-level detail of the miniature elements, requires Pro’s compositional precision.

Prompt 15: Surreal Animal Portrait

The prompt:

A regal portrait of an owl wearing a tiny Victorian-era professor's outfit — a tweed jacket with leather elbow patches, a miniature monocle perched on its beak, and a stack of tiny leather-bound books under one wing. The owl sits on an ornate wooden lectern in a candlelit library with towering bookshelves. Warm, flickering candlelight as the primary light source with pools of shadow between the shelves. Shot at 85mm f/2.0, owl tack-sharp, library background with gentle bokeh. Style: Beatrix Potter character design meets Dutch Golden Age still life painting. Rich academic palette — chestnut brown, burgundy, gold leaf, cream parchment. 4:5 aspect ratio. The owl's eyes should be intelligent and expressive, the clothing should fit naturally on the bird's body, no cartoonish proportions — maintain the owl's real anatomy.

What it creates: A charming anthropomorphic animal portrait that balances whimsy with artistic quality. The instruction to “maintain the owl’s real anatomy” prevents the common AI mistake of making animal characters look like humans in animal costumes.

Customization tips: Swap the animal and profession — a fox as a chef, a cat as a detective, a bear as a musician. The formula is: real animal + period-specific profession outfit + environment matching the profession + lighting that matches the era. Keep the “maintain real anatomy” instruction for any animal portrait.

Best model: Nano Banana Pro. The combination of realistic animal anatomy with fitted clothing and period-specific environment details makes this a complex compositional challenge ideal for Pro.

Prompt 16: Magical Landscape

The prompt:

A breathtaking fantasy landscape of floating islands connected by ancient stone bridges, with waterfalls cascading off the island edges into clouds below. The largest island has a cherry blossom grove in full bloom, with petals drifting in the wind between islands. In the distance, a massive crescent moon sits low on the horizon, partially obscured by wispy clouds. Golden sunset light filters through the clouds from behind the moon. Wide-angle 24mm establishing shot capturing the full scope of the floating archipelago. Style: Makoto Shinkai's atmospheric backgrounds from "Your Name" meets Thomas Cole's Hudson River School landscapes. Dreamy palette — soft pink blossoms, warm golden light, cool blue-violet shadows, misty white clouds. 21:9 ultra-wide aspect ratio. No buildings or structures except the stone bridges, no characters visible, the scale should feel vast and awe-inspiring.

What it creates: A cinematic fantasy landscape perfect for desktop wallpapers, game concept art, or editorial illustrations about imagination and exploration. The 21:9 ultra-wide ratio makes it ideal for website banners.

Customization tips: Replace cherry blossoms with autumn maples for a warmer palette, or ice crystals for a winter version. Change the floating islands to floating ships, floating libraries, or floating gardens. The Makoto Shinkai reference consistently produces that ethereal atmospheric quality with luminous skies.

Best model: Nano Banana 2 (Flash). Landscapes without people or complex character interactions are Flash’s strength. The wide-angle composition and atmospheric effects render beautifully at Flash speed.

Prompt 17: Enchanted Object

The prompt:

A close-up photograph of an ancient leather-bound spellbook lying open on a stone altar, with luminous golden runes rising from the pages and floating upward like embers. The runes cast a warm glow on the surrounding stone surfaces. Small tendrils of emerald-green magical mist curl around the book's edges. The pages are aged parchment with hand-drawn botanical illustrations visible beneath the floating runes. A single black raven feather serves as a bookmark. Dramatic chiaroscuro lighting — the book's glow is the primary light source in an otherwise dark chamber. Shot at 50mm macro f/2.8. Style: a prop photograph from a Peter Jackson film combined with the material detail of a medieval illuminated manuscript. Dark magical palette — aged parchment gold, emerald mist, warm amber glow, deep stone gray, raven black. 3:2 aspect ratio. No human hands visible, no modern materials, the floating runes should have slight motion blur suggesting movement.

What it creates: An atmospheric magical object study that feels like a prop from a high-budget fantasy film. The self-illuminating book creates a natural focal point and dramatic lighting without needing external light sources.

Customization tips: Replace the spellbook with any magical object — a glowing crystal on a pedestal, an enchanted sword in a stone, a potion bottle with swirling contents. The framework is: detailed object + self-illumination effect + atmospheric secondary effects (mist, particles, glow) + dark environment where the object is the light source.

Best model: Nano Banana Pro. The self-illumination, the precise interaction between the glowing runes and the stone surfaces, and the motion blur on the floating elements all benefit from Pro’s advanced light rendering.

Prompt 18: Storybook Character

The prompt:

A full-body illustration of a tiny fairy clockmaker working at a workbench made from a matchbox, using tweezers to adjust the gears inside a ladybug-sized pocket watch. She has translucent dragonfly wings with a subtle iridescent sheen, wears brass goggles pushed up on her forehead, and a leather work apron with tiny tool pockets. Around her workbench: spools of spider-silk wire, thimble cups of oil, and a candle in a bottlecap holder casting warm light. The scene sits on a real wooden windowsill with raindrops visible on the glass behind. Tilt-shift perspective making the scene feel miniature. Style: Arthur Rackham's fairy illustrations meets steampunk aesthetic meets Pixar character design. Warm workshop palette — brass, amber candlelight, translucent wing pastels, weathered leather brown. 1:1 aspect ratio. The fairy should have realistic proportions (not chibi), visible mechanical detail in the tiny watch, natural interaction between tiny and human-scale objects.

What it creates: A richly detailed character illustration that combines fairy-tale charm with steampunk craft elements. The scale contrast between the tiny fairy and the real windowsill creates visual interest and storytelling depth.

Customization tips: Change the fairy’s profession for different stories — a fairy botanist growing plants in acorn pots, a fairy cartographer mapping a garden, a fairy astronomer with a thimble telescope. The tilt-shift perspective instruction is the key to making miniature scenes feel both tiny and real. Swap “steampunk” for “art nouveau” or “cottage-core” for different aesthetic directions.

Best model: Nano Banana Pro. The extreme detail at miniature scale, the translucent wing rendering, and the precise scale relationships all demand Pro’s rendering precision.

Category 5: Abstract and Conceptual (3 Prompts)

Abstract prompts are where Nano Banana can produce genuinely unique art that doesn’t exist anywhere else. These are ideal for backgrounds, textures, social media graphics, and artistic expression. The key is being specific about the abstract qualities you want — “abstract art” alone produces nothing interesting, but precise descriptions of color, movement, and texture produce stunning results.

Prompt 19: Fluid Dynamics Abstract

The prompt:

An abstract macro photograph of colliding ink drops in water — deep indigo and liquid gold inks meeting and creating turbulent fractal patterns as they mix. The ink tendrils form organic, almost neural-network-like branching structures. Captured at the exact moment of maximum turbulence before the colors fully blend. Pure white background behind the water, lit from behind to create translucency in the thinner ink tendrils. Shot at 100mm macro f/8 with flash freezing the motion. Style: high-end abstract photography by Alberto Seveso meets scientific fluid dynamics visualization. Limited palette — deep indigo, liquid gold, pure white, with thin areas showing purple where colors partially mix. 1:1 aspect ratio. No bubbles, no surface ripples, no container edges visible — just the ink interaction isolated in space.

What it creates: A mesmerizing abstract image that works as a desktop wallpaper, phone background, social media graphic, or fine art print. The neural-network branching pattern gives it relevance for AI and technology contexts while remaining purely abstract.

Customization tips: Change the ink colors to match any brand palette. Replace “indigo and gold” with “crimson and silver” or “emerald and copper.” The Alberto Seveso reference is particularly effective for ink-in-water shots — swap it for “Fabian Oefner” for more explosive, energetic patterns. Adding “at the exact moment of” is a powerful instruction that tells the model to capture a specific phase of a dynamic process.

Best model: Nano Banana 2 (Flash). Abstract fluid dynamics without faces or complex objects is a strength of Flash, and the quick generation lets you run multiple color combinations rapidly.

Prompt 20: Geometric Growth Pattern

The prompt:

An abstract 3D render of a Fibonacci spiral constructed from hundreds of translucent glass cubes that gradually transition in color from deep sapphire blue at the center to warm rose gold at the outer edge. Each cube is slightly different in size, following the mathematical progression. The spiral floats against a soft gradient background transitioning from dark navy at the corners to warm charcoal at the center. Subtle caustic light patterns appear on the background surface where light passes through the glass cubes. Studio lighting from above with a slight warm rim light from behind. Style: teamLab digital installation art meets Olafur Eliasson's geometric light sculptures. Jewel-tone gradient palette — sapphire, amethyst, warm rose, rose gold, against dark navy. 16:9 aspect ratio. Each cube should be distinct and countable — not a blur, perfect mathematical spacing, no random elements breaking the pattern.

What it creates: A mathematically precise abstract composition that appeals to audiences interested in design, mathematics, science, and technology. The caustic light effects through glass cubes add a layer of photorealistic detail to an otherwise geometric concept.

Customization tips: Replace the Fibonacci spiral with other mathematical forms — a Voronoi diagram, a fractal tree, a Penrose tiling, or a golden ratio rectangle subdivision. Change the material from glass cubes to metallic spheres, wooden blocks, or liquid droplets. The key instruction is “each should be distinct and countable” — this prevents Nano Banana from blurring repeating elements into indistinct patterns.

Best model: Nano Banana Pro. The mathematical precision, caustic light rendering through translucent materials, and the color gradient across hundreds of individual elements all benefit from Pro’s detail capabilities.

Prompt 21: Organic Texture Study

The prompt:

An extreme macro photograph of a cross-section of a geode — crystalline amethyst formations growing inward from a rough volcanic rock exterior. The crystals catch light at different angles, creating prismatic rainbow micro-reflections. A thin layer of chalcedony (smooth, milky blue-white) forms the transition zone between the rough exterior and the crystal interior. The depth of field is razor-thin — only a narrow band of crystals is in focus, with the foreground and background crystals dissolving into a soft sparkle. Lit by a single fiber-optic spotlight from the upper left to maximize crystal facet reflections. Shot at 100mm macro f/2.8. Style: National Geographic macro photography meets James Turrell's light installations. Natural mineral palette — deep violet amethyst, milky blue chalcedony, rough brown exterior, prismatic rainbow highlights. 3:2 aspect ratio. Maximum crystalline detail in the focus band, no artificial color enhancement — the colors should look naturally mineral.

What it creates: A stunning macro texture study that works as fine art, scientific illustration, or premium background imagery. The narrow depth of field creates a dreamy quality while the focused band shows incredible crystalline detail.

Customization tips: Replace the geode with other natural textures — a cross-section of petrified wood, a macro of a butterfly wing’s scales, frost crystals on a window, or the surface of a soap bubble. The framework is: extreme macro + natural subject + specific focus band + single dramatic light source + scientific reference for authenticity.

Best model: Nano Banana 2 (Flash). Macro texture shots without people or complex compositions render beautifully on Flash, and the faster generation time lets you experiment with different lighting angles quickly.

Category 6: Architecture and Interior Design (3 Prompts)

Architecture prompts benefit enormously from Nano Banana’s spatial reasoning capabilities. The model can track up to 14 objects within a single scene, making it excellent for interior compositions with multiple furniture pieces, lighting sources, and decorative elements. These prompts use architectural photography terminology that the model understands from its training data.

Prompt 22: Modern Interior Living Space

The prompt:

An interior architecture photograph of a Japandi-style living room — a low walnut platform sofa with linen cushions in oatmeal and sage, a round stone coffee table with a single ikebana arrangement, and a floor-to-ceiling bookshelf in light ash wood with asymmetrically arranged books and ceramics. Large sliding shoji-inspired panels filter soft natural light from the left wall. The floor is polished concrete with a natural jute area rug. One wall features a subtle limewash texture in warm clay. Shot at 24mm f/8 for full room sharpness, camera height at seated eye level. Style: Kinfolk Home magazine meets Axel Vervoordt's wabi-sabi interiors. Calming palette — oatmeal, sage, walnut, warm clay, concrete gray, natural jute. 16:9 aspect ratio. No people, no pets, no electronics visible, no clutter — every object should feel intentionally placed.

What it creates: A serene, beautifully composed interior that could pass for a professional architectural photograph. The Japandi style (Japanese minimalism + Scandinavian warmth) is highly trending in 2026 and resonates with design-conscious audiences.

Customization tips: Swap “Japandi” for “mid-century modern,” “industrial loft,” “coastal grandmother,” or any interior style. Change the Axel Vervoordt reference to “Kelly Wearstler” for maximalist glamour or “John Pawson” for extreme minimalism. The key is specifying “camera height at seated eye level” — this perspective feels more natural and immersive than the default overhead angles AI models tend to produce.

Best model: Nano Banana 2 (Flash). Interior scenes with clearly defined objects and natural lighting are Flash’s territory. The 4-6 second generation time means you can quickly iterate on furniture arrangement and styling.

Prompt 23: Exterior Architectural Detail

The prompt:

An architectural detail photograph of a contemporary concrete and glass pavilion emerging from a hillside covered in wild grasses and native wildflowers. The building features a dramatic cantilevered overhang extending over a natural stream below. Floor-to-ceiling glass reveals a minimalist interior with a single Isamu Noguchi paper lantern glowing warmly inside. Shot at golden hour with the warm sunset light catching the raw concrete texture while the glass reflects both the landscape and the glowing interior simultaneously. Wide shot at 35mm f/5.6 from a slightly low angle to emphasize the cantilever. Style: architectural photography by Iwan Baan for Dezeen meets the landscape integration of Tadao Ando's work. Natural material palette — raw concrete gray, wild grass green, golden sunset, warm interior glow, clear glass. 16:9 aspect ratio. No cars, no people, no fences — the building should feel like it grew naturally from the landscape.

What it creates: A stunning architectural photograph that shows a building in perfect harmony with its landscape. The dual reflection in the glass (landscape + interior glow) creates the kind of complex visual layering that makes architectural photography compelling.

Customization tips: Change the building style and landscape — a desert adobe with Joshua trees, a seaside concrete house with ocean views, a forest cabin with surrounding pines. The cantilever element adds drama to any structure. Replace the Tadao Ando reference with “Frank Lloyd Wright” for more organic integration or “Zaha Hadid” for futuristic curves.

Best model: Nano Banana Pro. The simultaneous reflection and transparency in the glass, the precise concrete texture catching golden light, and the landscape-architecture integration all benefit from Pro’s rendering precision.

Prompt 24: Atmospheric Interior Detail

The prompt:

A moody interior detail photograph of a reading corner in an old English country house — a worn leather Chesterfield armchair beside a tall window with heavy linen curtains partially drawn. A stack of antique books with cracked leather spines sits on a small mahogany side table next to a crystal whiskey glass (empty, with residual amber tone). Dust motes float in a single shaft of afternoon sunlight cutting diagonally across the frame. The wallpaper is a faded William Morris floral pattern in muted greens and golds. Shot at 50mm f/2.0 with the armchair in focus and the window gently blown out. Style: interiors photography for The World of Interiors magazine meets a still from a Merchant Ivory period film. Rich heritage palette — aged leather brown, faded William Morris green and gold, linen cream, warm shaft-of-light amber, crystal clear. 4:3 aspect ratio. No person sitting in the chair, the impression that someone just left — a slight indent in the leather seat, a reading bookmark visible in the top book.

What it creates: An evocative, story-rich interior scene that conveys history, comfort, and quiet luxury. The “someone just left” instruction creates narrative tension — the empty chair tells a story that a occupied one wouldn’t.

Customization tips: Replace the country house aesthetic with a 1970s modernist apartment, a Japanese ryokan room, or a Moroccan riad courtyard. The “dust motes in a shaft of light” instruction works in any interior and adds atmospheric depth. The “someone just left” concept can be applied to any space — a recently played piano, a garden chair with a hat left behind, a desk with an open letter.

Best model: Nano Banana Pro. The dust motes in the light shaft, the faded wallpaper detail, the leather texture with the subtle indent, and the precise crystal refraction all require Pro’s material rendering capabilities.

Category 7: Infographic-Style Compositions (3 Prompts)

One of Nano Banana 2’s breakthrough capabilities is accurate text rendering within images. According to Google’s developer documentation, the model can render text and structured data accurately when you describe layout logic, hierarchy, and explicit labels. These prompts leverage that capability for infographic-style visuals that combine information design with aesthetic appeal.

Prompt 25: Comparison Infographic

The prompt:

A clean, modern infographic comparing two items side by side. Left column header: "NANO BANANA 2" in bold sans-serif. Right column header: "NANO BANANA PRO" in bold sans-serif. Below each header, 5 comparison rows with icons: Speed (clock icon) — "4-6 sec" vs "10-20 sec"; Quality (star icon) — "95%" vs "100%"; Cost (dollar icon) — "$" vs "$$$"; Best For (target icon) — "Social Media, Blogs" vs "Print, Brand Assets"; Resolution (screen icon) — "Up to 4K" vs "Up to 4K". Use a vertical divider line between columns. Background: clean white with subtle light gray grid lines. Style: Apple Keynote presentation slide meets Information is Beautiful by David McCandless. Minimal palette — white background, dark charcoal text, electric blue for left column accents, warm coral for right column accents. 16:9 aspect ratio. Clean typography, generous whitespace, no decorative borders, no gradients on the background.

What it creates: A shareable comparison graphic perfect for social media, blog posts, or presentations. The explicit text instructions leverage Nano Banana 2’s text-rendering capabilities.

Customization tips: Replace the comparison items with any two things — tools, services, plans, features. Keep the number of rows to 5-6 for readability. The icon descriptions (clock, star, dollar, target, screen) help the model generate simple recognizable symbols. For a three-way comparison, change to a three-column layout but reduce to 4 rows to maintain clarity.

Best model: Nano Banana 2 (Flash). Text rendering is actually a strength of the Flash model (Nano Banana 2 outperforms Pro for text accuracy in many scenarios), and the structured layout plays to Flash’s speed advantage.

Prompt 26: Process Flow Diagram

The prompt:

A horizontal 5-step process flow diagram on a clean white background. Each step is a rounded rectangle with a large step number (1-5) and a short label inside: Step 1 "Write Prompt" → Step 2 "Choose Model" → Step 3 "Generate" → Step 4 "Refine" → Step 5 "Export". Connecting arrows between each step in a gentle curve. Each step has a small relevant icon above it: a pencil, a toggle switch, a sparkle burst, a slider control, and a download arrow. A subtle gradient shadow beneath each rectangle gives depth. Style: Stripe's documentation design system meets Linear app interface design. Clean SaaS palette — white background, soft rounded rectangles in light gray with colored left borders (gradient from blue step 1 to green step 5), dark charcoal text, muted colored icons matching their step's border. 16:9 aspect ratio. Maximum readability, consistent spacing between all elements, no decorative flourishes.

What it creates: A professional process diagram suitable for tutorials, documentation, blog posts, or presentations. The Stripe documentation reference produces that clean, modern SaaS aesthetic.

Customization tips: Change the steps to match any process — onboarding flows, recipe steps, prompt writing workflows, project management stages. Keep steps to 5-7 for horizontal layouts. For vertical processes (like funnels), change to a top-to-bottom layout with downward arrows. The “colored left borders with gradient” instruction creates visual progression that helps readers follow the flow.

Best model: Nano Banana 2 (Flash). Clean diagrams with text and simple shapes are ideal for Flash — the text rendering is strong and the speed lets you iterate on layout quickly.

Prompt 27: Data Visualization Art

The prompt:

An artistic data visualization showing the growth of AI image generation from 2022 to 2026 — represented as a tree growing from a small sapling (left/2022) to a massive, elaborate banyan tree (right/2026). Each major branch is labeled with a model name at its branching point: "DALL-E 2" (2022), "Midjourney v5" (2023), "Stable Diffusion XL" (2023), "DALL-E 3" (2024), "Midjourney v6" (2024), "Flux" (2024), "Nano Banana" (2025), "Nano Banana Pro" (2025), "Nano Banana 2" (2026). The branches grow progressively thicker and more elaborate toward the right. Leaves on each branch represent user adoption — sparse leaves in 2022, dense canopy by 2026. Background: a soft warm gradient from dawn (left) to full daylight (right) representing the timeline. Style: Giorgia Lupi's data humanism visualization meets botanical illustration by Ernst Haeckel. Organic palette — bark brown, leaf greens (lighter for earlier years, deeper for recent), warm sky gradient. 21:9 ultra-wide aspect ratio. Each label must be clearly readable, the tree should follow realistic growth patterns, the timeline progression should feel natural left-to-right.

What it creates: A unique data visualization that transforms timeline data into an organic, visually striking illustration. This blends information design with art in a way that performs well on social media and in presentations where standard charts feel boring.

Customization tips: Replace the tree metaphor with a river (tributaries joining a main stream), a city skyline (buildings getting taller), or a mountain range (peaks getting higher). The key is mapping your data dimension to a physical dimension the model understands. Change the model names to any timeline data — company milestones, technology evolution, historical events.

Best model: Nano Banana Pro. The combination of readable text labels, organic tree structure, and timeline-accurate branching points creates a complex composition that benefits from Pro’s spatial reasoning and text-rendering precision.

Category 8: Dark, Cinematic, and Dramatic (3 Prompts)

Cinematic prompts use film-industry terminology that Nano Banana understands deeply from its training data. Specifying a director’s visual style, a film stock look, or a specific color grading approach produces dramatically different results from generic “dark and moody” instructions. These prompts are designed for editorial content, creative projects, and social media graphics that need to evoke strong emotion.

Prompt 28: Neo-Noir Urban Scene

The prompt:

A cinematic neo-noir photograph of a rain-soaked Tokyo alley at night. Neon signs in Japanese kanji reflect in puddles on the wet asphalt, creating a mirror world of distorted light below. A lone figure in a long dark coat walks away from the camera, their silhouette backlit by a distant warm-toned ramen shop lantern. Vertical neon signs in electric pink, cyan, and amber line both sides of the narrow alley. Steam rises from a grate, catching the neon light and creating colored fog at ankle level. Shot at 35mm f/2.0 from a low angle, emphasizing the reflections and the corridor-like depth. Style: Ridley Scott's Blade Runner meets Wong Kar-wai's "In the Mood for Love" color grading. Saturated neon palette against deep black shadows — electric pink, cyan, warm amber, wet asphalt black, steam white. 9:16 vertical aspect ratio. The figure should be small in the frame (1/5 of image height), no face visible, the environment is the star.

What it creates: A moody, atmospheric urban scene that works as a phone wallpaper, Instagram story, editorial illustration, or concept art. The Blade Runner + Wong Kar-wai combination produces a specific neo-noir aesthetic that’s visually distinctive.

Customization tips: Move the setting from Tokyo to Hong Kong, Seoul, or a fictional cyberpunk city. Remove the human figure entirely for a pure environment shot. Change the neon colors to match a brand palette. Replace “Blade Runner” with “Drive by Nicolas Winding Refn” for a more restrained, cool-toned look, or “Enter the Void by Gaspar Noe” for more extreme neon saturation.

Best model: Nano Banana 2 (Flash). Urban nightscapes with neon reflections are a sweet spot for Flash. The model handles the wet surface reflections and neon glow well, and the silhouette figure avoids the facial detail challenges where Pro would be needed.

Prompt 29: Dramatic Nature Power

The prompt:

A dramatic nature photograph capturing the exact moment a lightning bolt strikes a lone ancient oak tree on a hilltop during a violent thunderstorm. The lightning illuminates the tree's twisted branches in stark white-blue light while the sky behind churns with layered storm clouds in deep purple and charcoal. Rain is frozen mid-fall by the flash, creating a curtain of silver droplets. The foreground shows wind-whipped tall grass bending dramatically to the right. A second, distant lightning bolt forks across the sky in the background. Shot at 24mm wide-angle f/4, long exposure feel with the lightning sharp and the clouds showing subtle motion streaks. Style: National Geographic storm photography by Mike Hollingshead meets Romantic-era landscape painting by Caspar David Friedrich. Dramatic storm palette — electric white-blue lightning, deep purple sky, charcoal clouds, silver rain, dark green grass, ancient bark brown. 16:9 aspect ratio. The tree should feel ancient and massive (at least 200 years old), gnarled and weathered, not a generic tree shape. Maximum drama in the sky.

What it creates: A breathtaking dramatic nature image that conveys raw power and the sublime beauty of extreme weather. The frozen rain and dual lightning bolts create a sense of captured-moment immediacy.

Customization tips: Replace lightning with other dramatic weather — a tornado approaching a farmstead, massive waves crashing against a lighthouse, an avalanche descending a mountain face. The “exact moment” instruction is critical for dynamic weather shots. Replace Caspar David Friedrich with “Ansel Adams” for black-and-white dramatic landscapes or “Albert Bierstadt” for golden-lit American West grandeur.

Best model: Nano Banana 2 (Flash). Landscape scenes without people or complex character interactions, even dramatic ones, render excellently on Flash. The speed advantage lets you generate multiple variations to find the most dramatic lightning composition.

Prompt 30: Cinematic Sci-Fi Establishing Shot

The prompt:

A cinematic establishing shot of a massive abandoned space station orbiting a gas giant planet — the station's curved hull is covered in centuries of micrometeorite pitting and faded corporate logos, with several sections trailing debris where hull breaches have occurred. Warm amber light from the gas giant's atmosphere illuminates the station from below while cold starlight provides rim lighting on the upper surfaces. Through a large shattered viewport, you can see the interior: floating debris, dormant control panels, and a single emergency light still blinking red. The gas giant fills 60% of the background with swirling bands of burnt orange, cream, and deep rust. Shot as if from a nearby spacecraft window, slight vignetting at the edges. Style: concept art by Syd Mead meets the production design of "Alien" (1979) by Ron Cobb and Chris Foss. Retrofuturist sci-fi palette — weathered white hull with rust stains, warm amber gas-giant light, cold blue starlight, emergency red blink, deep space black. 21:9 ultra-wide cinematic aspect ratio. The station should feel massive (include tiny recognizable scale references like antenna arrays and docking ports), aged and abandoned but not destroyed — dignity in decay.

What it creates: A grand-scale science fiction scene that evokes classic hard sci-fi aesthetics. The dual lighting (warm from below, cold from above) creates dramatic dimension, while the visible interior through the broken viewport adds narrative depth — who were here, and why did they leave?

Customization tips: Replace the space station with a derelict generation ship, a lunar mining colony, or an underwater research facility on Europa. The “dignity in decay” instruction prevents the common AI mistake of making abandoned structures look generically trashed. Swap the Syd Mead reference for “Simon Stalenhag” for a more grounded, near-future industrial feel, or “Moebius (Jean Giraud)” for a more illustrative, comic-art approach.

Best model: Nano Banana Pro. The complex spatial composition (exterior hull + visible interior through viewport + background planet), the precise lighting interaction between two color-temperature sources, and the micro-detail of the weathered hull all demand Pro’s rendering capabilities.

Style Keywords That Actually Work with Nano Banana

After generating hundreds of images across the Nano Banana model family, certain keywords consistently produce better results than others. Here’s a curated reference of the most effective style keywords organized by category, based on testing across both Flash and Pro models:

Photography Style Keywords

Lens and camera terms: “shot on 85mm f/1.4” (portrait compression + bokeh), “wide-angle 24mm” (environmental/architectural), “100mm macro” (detail/texture), “135mm telephoto” (intimate compression), “tilt-shift” (miniature effect). These terms activate specific visual qualities because Nano Banana was trained on millions of photographs tagged with EXIF data.

Lighting terms: “Rembrandt lighting” (dramatic triangle under eye), “butterfly lighting” (even, glamorous), “split lighting” (half-face shadow), “golden hour” (warm directional), “blue hour” (cool ambient), “practical lighting” (light sources visible in scene). Specifying the exact lighting position (“45 degrees from camera right, slightly above”) gives even more control.

Film and processing terms: “Kodak Portra 400” (warm skin tones, pastel palette), “Fujifilm Pro 400H” (cool greens, soft contrast), “Kodachrome” (rich saturated vintage), “cross-processed” (shifted color cast), “cinematic color grading” (teal-orange split toning). Film stock references are one of the most reliable ways to control Nano Banana’s overall color palette.

Art and Illustration Style Keywords

Named artist references: “Studio Ghibli” (warm hand-painted anime), “Makoto Shinkai” (luminous atmospheric), “James Jean” (intricate organic illustration), “Moebius” (clean-line sci-fi), “Arthur Rackham” (detailed fairy-tale), “Alphonse Mucha” (Art Nouveau decorative). Named artists are the most powerful style controls — they compress dozens of visual attributes into a single reference.

Art movement terms: “Art Deco” (geometric glamour), “Art Nouveau” (organic flowing lines), “Bauhaus” (geometric minimalism), “Impressionist” (visible brushstrokes, light play), “Dutch Golden Age” (rich chiaroscuro still life), “Pre-Raphaelite” (vivid, detailed, romantic). These work best when combined with a subject that matches the movement’s historical strengths.

Digital and modern terms: “volumetric lighting” (god rays, atmospheric depth), “subsurface scattering” (translucent skin/materials), “caustic lighting” (light through water/glass), “ambient occlusion” (soft shadow in crevices), “ray-traced reflections” (precise mirror effects). These 3D rendering terms activate specific visual effects in the model’s output.

Aspect Ratio Recommendations by Use Case

Choosing the right aspect ratio before you prompt saves regeneration time and ensures your image fits its destination without awkward cropping. Here’s a practical guide based on where you’ll actually use the images:

• 1:1 (Square): Instagram feed posts, profile pictures, product thumbnails. The most versatile ratio — works everywhere but excels nowhere specifically.
• 4:5 (Vertical rectangle): Instagram feed posts (takes up more screen space than 1:1), Pinterest pins, portrait photography. This is the maximum vertical ratio Instagram allows in-feed, making it ideal for maximum engagement.
• 9:16 (Tall vertical): Instagram Stories, TikTok, YouTube Shorts, phone wallpapers. Essential for vertical-first mobile content.
• 16:9 (Widescreen): Blog hero images, YouTube thumbnails, Twitter/X posts, presentation slides, desktop wallpapers. The standard widescreen ratio for web content.
• 3:2 (Classic photo): Editorial photography, print use, photography portfolios. Matches traditional 35mm film proportions and feels “photographic.”
• 21:9 (Ultra-wide): Cinematic scenes, website banners, panoramic landscapes, email headers. Creates a dramatic, film-like composition.
• 4:3 (Standard): Traditional display ratio, good for detailed scenes that need more vertical space than 16:9 but less than 1:1.

Nano Banana 2 supports all these ratios plus extreme formats like 1:4, 4:1, 1:8, and 8:1 — useful for bookmark designs, timeline graphics, and vertical scrolling content. Always specify the aspect ratio in your prompt rather than relying on the platform’s default setting.

Common Mistakes (and How to Fix Them)

After analyzing thousands of Nano Banana generations, these are the mistakes that consistently produce poor results — and the specific fixes for each:

Mistake 1: Style Stacking Without Hierarchy

The problem: Combining “watercolor + cyberpunk + Pixar + oil painting” in one prompt creates visual noise. The model tries to satisfy all style references simultaneously and produces a muddy compromise that matches none of them.

The fix: Use a maximum of 2 style references, and make them complementary rather than contradictory. “Studio Ghibli meets Wes Anderson” works because both share warm palettes and deliberate composition. “Gritty realism meets Disney animation” fails because the foundational aesthetics clash. If you must use multiple references, specify hierarchy: “primarily in the style of X, with the color palette of Y.”

Mistake 2: Ignoring the Exclusion Layer

The problem: Without explicit exclusions, the model sometimes adds extra limbs, warped text, messy logos, or unwanted background elements. Many users write detailed positive instructions but forget to specify what they don’t want.

The fix: Always end your prompt with 3-5 specific exclusions. “No text overlays, no watermarks, no extra fingers, no distracting background elements” is a solid default set. For portraits, add “no asymmetric eyes, no plastic skin texture.” For architecture, add “no people, no cars, no anachronistic elements.” Think about what commonly goes wrong in the type of image you’re generating.

Mistake 3: Vague Lighting Instructions

The problem: “Good lighting” or “well-lit” tells the model nothing useful. The default when no lighting is specified is flat, even illumination — which is why so many AI images look like they were shot under fluorescent office lights.

The fix: Specify at minimum: light source position (left, right, above, behind), light quality (hard/soft, focused/diffused), and color temperature (warm amber, cool blue, neutral white). “Warm golden hour side lighting from camera left with long shadows extending to the right” gives the model everything it needs to create dramatic, professional-quality lighting.

Mistake 4: Overloading a Single Prompt

The problem: Trying to fit 15 characters, a complex narrative, and a detailed environment into one prompt exceeds the model’s compositional capacity. Google’s documentation says the model tracks up to 5 characters and 14 objects — exceeding these limits causes elements to merge, disappear, or render incorrectly.

The fix: Respect the limits. For complex scenes, break them into multiple generations and composite in post-production. For a single generation, keep to 3 characters maximum and 10 significant objects. Use Claude or another AI assistant to help you simplify and prioritize your prompt elements before generating.

Mistake 5: Not Iterating Systematically

The problem: Rewriting the entire prompt from scratch each time you don’t like the result. This makes it impossible to identify which element caused the issue.

The fix: Keep 80-90% of your prompt identical between iterations and change only one or two attributes at a time. If the lighting is wrong, change only the lighting section. If the composition feels off, change only the camera/framing section. This systematic approach converges on your desired result much faster than starting over, and Nano Banana 2’s 4-6 second generation time makes rapid iteration practical.

Level Up Your Prompt Skills with the CRAFT Framework

The prompts in this guide follow principles from the CRAFT Prompt Engineering Framework — our structured approach to writing prompts that consistently produce professional results across any AI model, not just Nano Banana. CRAFT stands for Context, Role, Action, Format, and Tone — and when applied to image generation, it translates to Environment, Style Reference, Subject Action, Composition, and Mood.

If you want to go beyond copy-pasting prompts and learn to write your own from scratch — for images, text, code, and more — the AI Agent Starter Kit ($19) includes the complete CRAFT framework with 50+ templates, real-world examples across every major AI platform, and a prompt engineering cheat sheet you can reference while you work. It’s built specifically for beginners who want to stop guessing and start engineering.

Get More AI Guides Delivered Free

Want prompt templates, tool comparisons, and AI tutorials like this one delivered to your inbox every week? Join 5,000+ readers who get our free Weekly AI Intel Report — no spam, no fluff, just the practical AI knowledge that matters for beginners.

Download the free Weekly AI Intel Report on Gumroad

Related Articles

• The Complete Nano Banana AI Guide: Everything You Need to Know
• AI Image Generation Guide: From Beginner to Advanced
• Best AI Tools for Beginners (2026)
• Best Claude Prompts: 50 Examples That Actually Work
• How to Write AI Prompts: The Complete Beginner’s Guide

Frequently Asked Questions

What is the difference between Nano Banana 2 and Nano Banana Pro?

Nano Banana 2 is built on Google’s Gemini 3.1 Flash Image architecture and generates images in 4-6 seconds at roughly 95% of Pro’s quality — making it ideal for social media content, blog images, and rapid iteration. Nano Banana Pro is built on Gemini 3 Pro Image and takes 10-20 seconds per generation but delivers the highest fidelity available — richer textures, more natural lighting, and superior spatial composition. Choose Flash for volume and speed, Pro for hero images and print-ready assets where every detail matters. Both support resolutions up to 4K and the same set of aspect ratios, though Flash adds extreme formats like 1:8 and 8:1 that Pro does not support.

How do I access Nano Banana for free?

You can access Nano Banana through several free channels. Google AI Studio (aistudio.google.com) offers free access to Nano Banana 2 with generous rate limits for individual users. The Gemini app (gemini.google.com) includes image generation capabilities in its free tier. Flowith offers a free Starter plan with 300 one-time credits that includes access to Nano Banana alongside 40+ other AI models. Third-party platforms like Leonardo.ai and SeaArt also provide free-tier access to Nano Banana models with daily generation limits. For serious use, Flowith’s Professional plan at $19.90/month provides 22,000 monthly credits across all models.

Why do my Nano Banana prompts produce generic-looking images?

Generic results almost always come from prompts that are too short or too vague. Nano Banana 2’s reasoning layer — built on Gemini 3.1 — is designed to parse complex, multi-layered instructions including spatial relationships, lighting logic, and compositional intent. When you write “a beautiful landscape,” you’re bypassing 90% of the model’s capabilities. The fix is structural specificity: use the 7-part formula (Subject + Action + Environment + Camera + Lighting + Style + Exclusions) and include concrete details in each section. “A weathered lighthouse on a rocky Maine coastline during a nor’easter, shot at 35mm f/5.6 from sea level, dramatic storm lighting from the left, style of Andrew Wyeth’s coastal paintings, no people, no boats” will always outperform “a lighthouse in a storm.” The prompts in this guide demonstrate this principle across all 30 examples.

Can Nano Banana generate accurate text within images?

Yes — text rendering is one of Nano Banana 2’s significant improvements over earlier models and most competitors. The model can render titles, labels, and short text blocks accurately when you describe the layout logic, text hierarchy, and explicit content. The key is thinking like a graphic designer in your prompt: specify where the text appears, what font style it should approximate (sans-serif, serif, handwritten), the exact words to render, and the text’s visual relationship to other elements. For best results, keep text to 20 words or fewer per image, specify exact capitalization, and include text styling in your prompt (“bold white sans-serif text in the upper third of the frame reading ‘AI TOOLS 2026′”). Nano Banana 2 actually outperforms Pro for text accuracy in many scenarios, as confirmed by Google’s own comparison data.

How many elements can Nano Banana handle in a single prompt?

According to Google’s technical documentation, Nano Banana 2 can track up to 5 characters and 14 distinct objects within a single generation while maintaining subject consistency. In practice, this means you can create complex scenes — a room with multiple furniture pieces, a street scene with a few people, or a product arrangement with several items — without elements merging or disappearing. However, exceeding these limits degrades quality quickly: characters may share features, objects may fuse together, or elements may simply not appear. For the best results, keep your prompts to 3 characters maximum and 10 significant objects. If your vision requires more, break the scene into multiple generations and use image editing tools to composite the final result. Every prompt in this guide stays within these limits, which is one reason they produce reliable results.

Stay Updated

AI image generation is evolving fast — Nano Banana went from its first release to 200 million edits in under a year. To keep up with new models, updated prompt techniques, and the latest tools, subscribe to the Beginners in AI newsletter for weekly updates delivered straight to your inbox.