Loading model...
Loading model...
Best prompts, specs & benchmarks
OpenAI's next-gen image model with photorealistic quality and accurate text rendering.
API available
167
—
—
—
Pricing details for ChatGPT Image 2 haven’t been recorded yet.
No public benchmark results have been recorded for ChatGPT Image 2 yet.
We track AA Image Arena, LMArena T2I, GenEval and DPG-Bench for this kind of model — scores appear here as soon as a published result for ChatGPT Image 2 has been verified against its source.
Loading prompts...
3 questions answered
Advertisement

A supplement product image prompt that designs a complete, high-converting eCommerce listing image in one shot — hero bottle, headline and trust badges. Set your product name, benefits and packaging and the template renders a marketplace-ready visual: a photoreal bottle or jar as the hero, a strong headline, an icon-led features list and certification badges, composed like a top Amazon or marketplace listing. Packaging, style, palette, benefits layout, badges and background are all selectable, so the same system works for a premium dark jar, a clean white bottle, a natural botanical look or a bold colorful promo. What you can customize Brand Name and Product Name — the text rendered on the packaging and headline, exactly as typed Benefits and Trust Badges — three or four short icon-plus-label points that become the features UI, plus your real certifications and callouts such as GMP, ISO, capsule count or no additives Packaging / Bottle Type — six options including Premium Black Jar, Clean White Bottle, Amber Glass Bottle and Kraft Eco Pouch Design Style and Color Palette — five styles (Premium Clean, Natural Organic, Bold Promo, Medical Clinical, Luxury Dark) paired with six palettes from Charcoal & Silver to Black & Gold Luxe and Warm Earthy Background — six settings such as Light Studio Podium, Pure White, Natural Botanical, Dark Moody and Marble Surface Best for Supplement and wellness brands, dropshippers, Amazon and marketplace sellers, and agencies that need fast, professional listing images, hero shots and ad creative without a photo studio. Tips for the best results Name the brand and product exactly as they should appear, and keep each benefit to an icon label plus a few words rather than a sentence. List your real badges in the badges field and match the packaging, palette and background to your actual product so the mockup is believable. Keep every line concise so the typography stays crisp and correctly spelled; avoid long paragraphs and more than four benefits, which crowd the layout and degrade the text. Do not make unverified health or medical claims in the copy, and avoid clashing palette-and-style combinations. Text-in-design and small label text are demanding, so regenerate once if any line renders soft or misspelled.

An Instagram carousel template prompt that designs a full 3-slide set in one image — a premium hook, a structured content slide and a call-to-action outro. Write your topic and copy and the template renders a professional reference sheet: Slide 1 stops the scroll, Slide 2 delivers the value with icons and hierarchy, and Slide 3 drives the follow or save, every slide sharing the same palette, fonts and branding. Style, palette and iconography are all selectable, so the same structure can look like a bold modern deck, a minimal editorial layout, a playful gradient set or a luxury dark theme. What you can customize Slide 1 — Hook Headline — the short, punchy scroll-stopper on the opening slide Slide 2 — Content — the structure of the value slide: a numbered list, three tips, a before/after or a mini step-by-step Slide 3 — CTA / Outro — the closing call to action, such as follow, save or link in bio Design Style / Layout and Color Palette — six styles (Bold Modern, Minimal Editorial, Playful Gradient, Luxury Dark, Corporate Clean, Retro Y2K) and six palettes including Navy, Cream & Coral, Black & Gold Luxe and Pastel Gradient Iconography — five treatments from Minimalist Line Icons and Solid Filled Icons to 3D Glossy, Emoji Style or No Icons Best for Creators, marketers, coaches and brands who post carousels — educational posts, listicles, tips, launches and lead magnets — and designers who need fast, on-brand slide mockups without opening Canva or Figma. Tips for the best results Keep the hook headline short and punchy; it carries the whole first impression. In the content field describe the structure you want rather than a wall of text, and limit each slide to three or four points. State one clear CTA and match the palette and iconography to your brand. Because the design is text-heavy, keep each line concise so typography stays crisp and correctly spelled; long paragraphs compress the layout and degrade the text. Avoid clashing style and palette choices, and regenerate once if any line renders soft or misspelled.

An expressions grid prompt that turns one photo into a clean 4x4 sheet of 16 panels — the same person, one emotion per panel, identity locked throughout. Upload a reference photo and the template renders a consistent expression sheet: happy, sad, angry, surprised and a dozen more, each in its own panel with matching framing and lighting so only the face changes. It is the reference-sheet look creators use for character design, emotion charts, reaction packs and profile-grid posts, generated in a single image. What you can customize Reference Photo and Subject Description — upload one clear front-facing photo so the person is preserved identically across all 16 panels, and describe hair, clothing and features alongside or instead of it The 16 Expressions — exactly which emotions appear and in what order, one short label per panel Art Style — five options: Photorealistic Studio, Anime / Manga, 3D Pixar-Style, Black & White and Comic / Cartoon Panel Framing and Background — Front Headshot, Close-Up Face or Head & Shoulders 3/4, on Neutral Grey, Clean White, Soft Gradient or Dark Studio Panel Labels — show a small emotion label under each panel, or faces only Best for Character designers and artists who need expression references, streamers and creators making reaction-image sets or sticker packs, and anyone who wants a fun "all my moods" grid of themselves. Tips for the best results Attach one clear, well-lit front-facing photo so the identity stays locked across all 16 panels; likeness quality follows the reference. Make every expression visibly distinct and avoid vague or duplicate entries. Match style and background to the use: clean studio for references, anime for character sheets. Keep labels short, since long text is where models misspell, and avoid busy backgrounds or cluttered, fine-detail outfits that are hard to keep identical across panels. Consistency across 16 faces is demanding, so regenerate if any panel drifts off-model.

A mirror selfie AI prompt for OOTD fit checks — bedroom, hotel and fitting-room mirrors with phone-real texture, with or without your photo. The template recreates the outfit-check format exactly as it lives on Instagram and Pinterest: a full-length mirror, the phone visible in hand, a relaxed confident stance angled to show the fit, and the un-retouched realism of a photo taken in thirty seconds, mirror smudges included. It works two ways: attach a reference photo to keep your exact face and hairstyle while the outfit and scene restyle around you, or leave it empty and describe any subject. What you can customize Reference Photo (optional) — attach a clear photo to keep your face and hairstyle while the outfit and mirror scene change around you Subject & Outfit — the hero field; describe the fit head to toe, from lounge sets to full streetwear layers Mirror Setting — five options: Bedroom Mirror, Hotel Room with city windows, Boutique Fitting Room, Gym Mirror and vintage Hallway Mirror Lighting — four options: Soft Daylight, Warm Evening, y2k-style Phone Flash glare and Golden Hour through blinds Best for Fashion creators and everyday posters alike: daily OOTD posts, styling reel covers, wardrobe try-on carousels, resale listing photos with lifestyle energy, and Pinterest outfit boards. Tips for the best results Put the effort into the outfit description; it is the field that shapes the post most. Nano Banana Pro handles the reference-photo path and fabric texture best; ChatGPT Image 2 is a strong alternative. Describe only one full outfit per generation, since layered alternates confuse the drape. Skip studio-lighting requests that break the casual story. Leave poses out of your description entirely; the fixed fit-check stance is what makes the image read as authentic.

An outdoor girl photoshoot prompt that creates bright, aesthetic travel-style portraits — the candid Instagram look of a sunny day out, in one shot. The template stages a confident young woman in a scenic spot with natural light, a flattering casual outfit and a relaxed candid pose, rendered with realistic skin and vivid true-to-life color. Subject, outfit, pose and location are all adjustable, so the same wholesome aesthetic re-renders at a mountain viewpoint, a beach, a cafe rooftop, a flower field or a city street. What you can customize Subject Description — hair, features and vibe; this anchors the render Outfit — free text, from a cozy sweater to a flowy dress Pose — five candid options: Hands Tucking Hair, Leaning on Railing, Candid Walking, Looking Over Shoulder and Arms Open to View Location — six scenic spots: Mountain Viewpoint, Tropical Beach, Cafe Rooftop, Flower Field, City Street and Lakeside Model Photo — optional upload of a clear front-facing photo to carry a real likeness into the portrait Best for Girls who want scroll-stopping personal photos without a photoshoot, and creators and editors making aesthetic lifestyle content: Instagram posts, profile pictures, travel and fashion feeds, and personal branding. Tips for the best results Describe the girl concretely; hair, features and vibe do the most work. Pick an outfit and location that suit each other, such as a cozy sweater for a hill viewpoint or a flowy dress for a beach, and match the light to the place. Keep the pose natural and candid, which is what sells the look; avoid over-posed stiff stances. Avoid busy backgrounds that fight the subject and heavy beauty filters that erase natural skin texture. Skip revealing or explicit styling requests; the template is tuned for a tasteful, wholesome aesthetic.

An aesthetic room portrait prompt that seats a person in a moody, RGB-lit creator space — the styled content creator setup look in rich cinematic color. The template composes the whole scene: a relaxed subject in an armchair, a curated shelf wall of plants and figurines behind them, a glowing tube light washing the room in color, and a soft foreground framing, all with realistic skin and a warm-against-cool color contrast. Room style and accent-light color are the main creative knobs, so the same vibe re-renders as a navy shelf studio, a cozy plant den, a minimalist setup or a neon gamer room in any glow you choose. What you can customize Subject Description — the person and outfit; hair, glasses and clothing anchor the render Room Setting — five rooms: Navy Shelf Studio, Cozy Plant Den, Minimalist White, Neon Gamer Room and Warm Wood Library Accent Light Color — five glows: Magenta & Blue, Teal & Cyan, Warm Orange, Purple & Pink and Green & Lime Model Photo — optional upload of a clear front-facing photo to carry a real likeness into the portrait Best for Creators, streamers and anyone who wants a polished personal-brand portrait or channel visual — profile pictures, about-page shots, thumbnail bases and social posts — with the moody, aesthetic setup look and no studio required. Tips for the best results Describe the person and outfit concretely; hair, glasses and clothing are what the render holds onto. Match the accent-light color to the mood you want; magenta-pink and teal read the most "creator". Keep the pose relaxed and seated, which is what makes the scene feel candid, and attach a photo if you want a real likeness. Avoid over-cluttered shelf descriptions that turn to visual noise, and do not blow out the accent light so it washes the face flat. Skip busy patterned outfits that fight the colored ambient light; a simple top keeps the subject clean against the glow.

A Lord Hanuman AI image prompt that generates jaw-dropping, crisp devotional portraits of Bajrangbali across his divine aspects, moods, and epic settings. It is the most extensively customizable deity template on PromptsRush, with nine controls for the full sweep of Hanuman imagery. Every selector carries a complete, self-contained description, so the output stays crisp and coherent across more than two million combinations. Why every aspect stays iconographically correct The Divine Aspect selector holds ten complete manifestations: the mighty warrior with the golden gada, the Sanjeevani flight bearing the mountain in one palm, the humble Bhakt of Shri Ram, the five-faced Panchmukhi form, a colossal cosmic form, the burning-Lanka leap with tail aflame, the protective Sankat Mochan blessing, the meditating mountain yogi, a dark obsidian guardian at dusk, and the vanguard on the shores of Ram Setu. Because each is written out in full, the gada, sindoor, mane and sacred name राम appear where they belong for that form. What you can customize Divine Aspect — ten manifestations, from Mighty Warrior Bajrangbali and Sanjeevani Flight to Panchmukhi Hanuman and Ram Setu Vanguard Scene & Action — free text describing what he is doing Physique & Fur — six treatments: Golden-Brown Realism, Deep Reddish-Brown, Obsidian Black, Radiant Golden Light, Sindoor Vermilion and Silver-White Sage Attire & Ornaments — eight sets, from Crimson-Gold Royal and War Armor Kavach to Simple Ascetic rudraksha and the Ram-Naam Shawl Location / Background — twelve settings including Storm Mountain Peak, Ocean Cliff & Lightning, Fiery Sacred Sky, Burning Lanka at Night, Himalayan Sunrise and Ram Setu Shoreline Mood, Art Style, Light & Extras — ten expressions from Fierce Unshakable Resolve and Roaring War Cry to Serene Devotion and Meditative Bliss; six art styles from Epic Cinematic Photorealism to Classical Devotional Painting; eight light palettes; and a free-text field for sacred extras like a saffron banner or drifting embers Best for Devotional wallpapers, Hanuman Jayanti and Tuesday-post content, YouTube bhajan and Ramayana-story visuals, thumbnails, and art prints. Tips for the best results Keep one register per image: pair wrathful aspects with storm light and devotional aspects with dawn gold. Let the aspect lead; choose attire, location and palette that agree with it. Use Scene & Action for concrete posture and movement, and the extras field for one sacred detail. Let awe, not shock, carry the power; keep the description reverent.

A Lord Shiva AI image prompt that generates jaw-dropping, crisp devotional portraits of Mahadev across his many divine forms, moods, and settings. Its heart is the Divine Form selector, backed by six more controls for scene, setting, mood, art style, light and sacred extras. The result is epic, reverent imagery in which the crescent moon, tripundra, rudraksha, serpent and Ganga always land where they belong. Why every form stays iconographically correct The selector holds ten complete manifestations, each carrying its own full description: Mahadev in serene meditation on Kailash, the fiery Third Eye awakening, Rudra in the fury of the Tandava, Nataraja dancing within the ring of cosmic fire, Neelkanth with the blue throat after drinking the halahala, the ash-smeared Ascetic Yogi of the high Himalayas, gentle smiling Bholenath granting blessings, the formless radiant Jyotirlingam, the monumental black Adiyogi statue, and the cosmic protector wielding the trishul. Because each form is self-contained, the iconography never blurs between them. What you can customize Divine Form / Avatar — ten forms, from Rudra — Tandava Fury and Nataraja — Cosmic Dance to Bholenath — Gentle Blessing and Formless Jyotirlingam Scene & Action — free text describing the moment being depicted Setting / Background — ten epic settings including Mount Kailash, Storm-Churned Cosmic Ocean, Moonlit Ganges Ghat, Himalayan Cave Shrine and Monsoon Lightning Sky Mood & Expression — ten moods from Perfect Serenity and Transcendent Bliss to Blazing Wrath and Ecstatic Dance Trance Art Style, Light & Palette and Extra Sacred Elements — six styles (Epic Painterly Concept Art, Cinematic Photorealistic, Ethereal Luminous Fantasy, Dark Moody Concept Art, Classical Devotional Painting, Monumental Statue Photography), seven palettes such as Indigo Storm & Ember Gold, Sacred Golden Dawn and Diya Flame & Smoke, plus optional extras like Nandi, the trishul planted in rock, or drifting vibhuti ash Best for Devotional wallpapers and posts, Mahashivratri and Sawan content, YouTube bhajan visuals, and epic art prints. Tips for the best results Match mood, form and palette to one register per image: serene forms with soft light, wrathful forms with storm light. Write the Scene & Action concretely so the chosen form has something specific to do. Use the extras field sparingly; one sacred element reads more clearly than several. Let the description stay reverent; the power of this imagery comes from awe, not shock.

A little Krishna prompt that generates divine, consistent 3D scenes of the beloved child form of Krishna for YouTube devotional videos and festival art. It captures the modern animated-film treatment of Bal Krishna: luminous blue skin, sparkling eyes, curly hair under a tiny golden crown with its peacock feather, and every traditional ornament from the U-shaped tilak to the anklets. Six controls around the character shape the scene, styling, location, expression, lighting and props. Why the character stays consistent All of those features are pinned inside a fixed character lock, so he remains perfectly consistent across an entire channel of uploads. That is what separates a beloved recurring character from a series of lookalikes. What you can customize Scene & Action — free text describing what little Krishna is doing Dhoti & Styling — six variations, from Classic Golden-Yellow Silk pitambar to Festive Red & Gold, Peacock Blue & Silver and Flower-Adorned Pastel Location — ten settings including Carved Temple Courtyard, Yamuna Riverbank at Dawn, Vrindavan Forest Glade, Gokul Butter Kitchen, Holi Courtyard and Rain-Washed Palace Veranda Expression & Emotion — ten moods from Sweet Mischievous Smile and Guilty Butter-Thief to Wide-Eyed Innocence, Serene & Peaceful and Joyful Laughter Lighting & Mood and Extra Props — seven lighting options such as Radiant Divine Glow, Warm Diya Evening, Bright Holi Midday and Silver Moonlight, plus optional touches like an overturned pot of makhan Best for Devotional Shorts and story videos, Janmashtami and festival posts, bhajan visuals, thumbnails, and wallpapers. Tips for the best results Describe the action concretely; the scene field is where each frame gets its story. Pair location and lighting that belong together, such as a diya-lit evening in the temple courtyard or Holi's bright festival light in the Holi courtyard. Pick one expression that fits the moment rather than stacking moods. Keep the tone warm and reverent; the genre lives on innocence and charm, and the template is built to keep it there.

An AI Indian kid prompt that generates consistent, cinematic scenes of an adorable toddler for YouTube storytelling videos with full scenario control. It is built for the hugely popular AI-baby story genre: a chubby-cheeked little boy with big expressive eyes living small emotional moments on Indian streets, from pleading for chai with folded hands to flying kites on a rooftop. Six controls around the character set the scene, outfit, location, expression, lighting and props. Why the character stays consistent Like every serial-character template, its heart is a fixed character lock: the boy's face, hair, skin tone, proportions and his little silver bracelet are pinned precisely and never change. He stays the same recognizable child across an entire channel rather than morphing between uploads. What you can customize Scene & Action — free text describing the moment Outfit — six options: Mustard Cotton Kurta, Festive Kurta-Pajama with a Nehru jacket, School Uniform, Winter Sweater & Monkey Cap, Play Tee & Shorts and Yellow Raincoat Location — ten Indian settings from Roadside Chai Stall, Sabzi Market and Village Courtyard to Temple Steps, Mithai Shop, Railway Platform, Mustard Field, Rooftop Kite Festival and Village Mela Expression, Lighting & Extra Props — ten moods across the storytelling arc, from Pleading & Hopeful, Teary & Crying and Pouting & Sulky to Wide-Eyed Wonder, Mischievous Grin, Proud and Beaming Toothy Smile; seven lighting moods such as Golden Hour, Monsoon Grey, Warm Evening Bulbs and Diya & Fairy Lights; and optional props for the scene Best for Shorts and long-form storyboards, thumbnails, and channel art. Frame in 9:16 or 2:3 for Shorts. Tips for the best results Describe the action concretely; a specific gesture or object gives the frame its story. Keep one light source per scene and choose a lighting mood that matches the location. Pick the expression that carries the beat of the story rather than a neutral default. This template renders its own fixed fictional character: it is not designed to reproduce any real child, and works best kept that way.

An AI sad dog prompt that generates consistent, cinematic 3D scenes of an adorable teary-eyed puppy for YouTube storytelling videos. It produces a heart-melting golden puppy with huge glossy eyes, standing upright in tiny clothes, living through hardship and hope one frame at a time. The render style is modern animated feature film: detailed fur groom, subsurface scattering on the ears, cinematic depth of field and a 4k octane render finish, with no text anywhere in the image. Why the character stays consistent The template solves the problem that kills most of these channels: character drift. A fixed character lock describes the puppy's fur, eyes, proportions and toddler-like stance precisely, so he stays the same recognizable character across every scene of a video. What you can customize Scene & Action — free text for what he is doing, such as holding an empty wooden bowl and looking up hopefully. Outfit — six options including Denim Overalls, Yellow Raincoat, Ragged & Patched, Winter Scarf & Hat, Little Schoolkid and Cozy Pajamas. Location — ten settings from a Dusty Farmyard and Rainy Night Street to a Deep Forest, Old Barn, Snowy Field, Campfire at Night, Sunflower Meadow, Junkyard and Beach at Sunset. Expression & Emotion — ten states covering the full arc: Heartbroken, Sobbing, Scared & Trembling, Exhausted & Weak, Brave & Determined, Joyful Relief and Peaceful & Content. Lighting & Mood — seven looks such as Morning God-Rays, Overcast Rain, Cold Blue Dusk, Golden Sunset, Soft Moonlight, Warm Firelight and Harsh Midday Sun. Extra Props & Details — optional text for touches like a single tear or tiny pawprints in the dust. Best for The expression range is deliberate: a story needs the tearful beginning, the struggle and the happy ending, and all of them come from this one template. Use it for YouTube Shorts and long-form storyboards, thumbnails and channel art. Tips for the best results Frame in 9:16 or 2:3 for Shorts. Describe the action concretely — what he holds, where he looks. Pick one light source per scene. Avoid crowding scenes with other characters; the genre works because the little dog carries the frame alone.

A 3D floor plan prompt that renders photorealistic cutaway property visualizations for real estate listings, with full control over layout and style. It produces the premium dollhouse archviz look: walls cut at waist height, every room staged and visible at once, floating on a clean white background. This is the visualization listing portals, developers and interior designers pay studios to produce, rendered with soft even studio lighting, crisp shadows and fine material detail. What you can customize Layout & Rooms — the property described in plain language, room by room; any property works, from a studio flat to a 4-bedroom villa, penthouse or office suite. Design Style — eight interiors: Modern Minimalist, Scandinavian, Luxury Contemporary, Industrial Loft, Japandi, Warm Traditional, Coastal and Bohemian. Camera View — four angles: 45° Isometric Aerial, Straight Top-Down, Dollhouse Perspective and Dramatic Corner Close-Up. Materials & Palette — six combinations such as Light Oak & White, Dark Walnut & Charcoal, Marble & Gold and Polished Concrete & Black. Furnishing Level — from Fully Furnished & Staged through Lightly Furnished and Fixtures Only to Empty Shell. Extra Features — optional text for balconies, terraces, pools or gardens. Best for Property listings and brochures, Airbnb and rental presentations, renovation and staging proposals, architecture portfolio pieces, and interior design client mockups. Tips for the best results Describe the layout the way an estate agent would — name each room and how spaces connect, such as an open-plan kitchen flowing into the living room. Keep room counts realistic for a single floor; one level per image renders most cleanly. Generate multi-storey homes as one image per floor. Avoid asking for dimension labels or room names in the image itself. Models garble embedded text, which is why this template explicitly excludes it and leaves annotation to your listing software.