AI image generator

AI Image Generator with GPT Image 2 and Nano Banana Pro

The Orvlex AI image generator runs five models from one prompt box: GPT Image 2, Nano Banana Pro, Seedream 5.5, Seedream 4.0 and Z-Image Turbo. An image costs 3 to 10 credits (about $0.03 to $0.10) and goes up to 4K.

Three of the models edit your own photos with up to four reference images.

Updated November 12, 2026

Model

0/4000

Reference images

Aspect ratio

Resolution

Images

Set it up now. You sign in when you generate; new accounts get 30 free credits.

Real results made on Orvlex

Unedited · prompts shown

GPT Image 2: Event poster with readable text

GPT Image 2 · 2K · 3:4

Event poster with readable text

Prompt: “A minimalist poster for a jazz night titled "BLUE HOUR" in bold serif capitals, subtitle "Friday 9 PM · The Lantern Room", deep navy background, a single brass saxophone, clean grid layout with generous margins”

GPT Image 2: Product photo with a printed label

GPT Image 2 · 2K · 1:1

Product photo with a printed label

Prompt: “Product photo of a matte white ceramic coffee mug on a pale oak table, soft morning window light, a small kraft paper label on the mug that reads "ORVLEX ROASTERS", shallow depth of field”

Z-Image Turbo: Watercolor illustration from one sentence

Z-Image Turbo · 4:3

Watercolor illustration from one sentence

Prompt: “A cozy reading nook by a rainy window, warm lamp light, stacked books, a sleeping orange cat on a knitted blanket, watercolor illustration”

Z-Image Turbo: Isometric vector illustration

Z-Image Turbo · 1:1

Isometric vector illustration

Prompt: “Isometric illustration of a tiny glass greenhouse on a floating island, soft pastel colors, clean vector style, white background”

Seedream 4.0: Photoreal street scene at golden hour

Seedream 4.0 · 16:9

Photoreal street scene at golden hour

Prompt: “Photoreal street scene in Lisbon at golden hour, a yellow tram climbing a narrow cobbled hill, tiled facades, people in summer clothes, 35mm film look”

Seedream 4.0: Photoreal close-up portrait

Seedream 4.0 · 3:4

Photoreal close-up portrait

Prompt: “Close-up portrait of an elderly fisherman in a knitted cap, weathered skin, kind eyes, overcast harbor light, boats softly out of focus, photoreal”

Nano Banana Pro: Same product, new backgroundInput image for this Nano Banana Pro result

Nano Banana Pro · 1K · 1:1 · from the input image

Same product, new background

Prompt: “Place this exact mug on a snowy wooden windowsill with a pine branch and warm string lights in the background. Keep the mug, its shape and its label exactly the same.”

Nano Banana Pro: Top-down flat lay with hand-drawn map

Nano Banana Pro · 1K · 4:3

Top-down flat lay with hand-drawn map

Prompt: “Top-down photo of a hand-drawn travel map of Kyoto with small illustrated landmarks, a brass compass and a cup of matcha on natural linen”

Seedream 5.5: Natural skin and window light

Seedream 5.5 · 1.5K · 3:4

Natural skin and window light

Prompt: “Candid portrait of a young woman reading by a café window on a cloudy afternoon, soft natural window light on her face, real skin texture with freckles, loose strands of hair, a ceramic cup on the table, photoreal, 50mm lens”

Seedream 5.5: Small, dense text on a product label

Seedream 5.5 · 1.5K · 4:3

Small, dense text on a product label

Prompt: “Product photo of a glass jar of loose-leaf tea on a linen cloth, with a paper label that reads "HIGH MOUNTAIN OOLONG" in large letters and below it in small print "Roasted in small batches · Brew at 90°C for 3 minutes · Net wt 100 g · Harvested spring 2026", soft daylight, sharp focus on the label”

GPT Image 2: Café photo to professional headshotInput image for this GPT Image 2 result

GPT Image 2 · 1K · 3:4 · from the input image

Café photo to professional headshot

Prompt: “Turn the person in the photo into a professional corporate headshot. Keep the face, hair and identity exactly the same. Navy blazer over a plain white top, soft studio lighting, clean light-grey background, head and shoulders, looking at the camera with a calm, friendly expression, sharp focus, photoreal, suitable for a LinkedIn profile.”

GPT Image 2: Phone photo to studio product shotInput image for this GPT Image 2 result

GPT Image 2 · 1K · 1:1 · from the input image

Phone photo to studio product shot

Prompt: “Turn the product in the photo into a professional studio product photo. Keep the product, its shape, colours and any label or text exactly the same. Place it on a smooth stone surface with a soft beige backdrop, a small dried flower beside it, soft diffused light from the left, gentle shadow, sharp focus on the product, square composition, high-end e-commerce style.”

Nano Banana Pro: Same photo, soft glam makeupInput image for this Nano Banana Pro result

Nano Banana Pro · 1K · 3:4 · from the input image

Same photo, soft glam makeup

Prompt: “Apply a soft glam makeup look to the person in the photo. Keep the face, features, hair, clothes, pose, background and lighting exactly the same; change only the makeup. Even, dewy skin, warm brown eyeshadow blended softly, thin black eyeliner with a small wing, defined lashes, softly shaped brows, peach blush, nude-pink satin lipstick. Natural, photoreal skin texture.”

GPT Image 2: Empty room to Scandinavian living roomInput image for this GPT Image 2 result

GPT Image 2 · 1K · 4:3 · from the input image

Empty room to Scandinavian living room

Prompt: “Redesign the room in the photo in a Scandinavian interior style. Keep the walls, windows, doors, floor plan and camera angle exactly the same. Add a light oak floor, a soft grey linen sofa, a round white coffee table, a wool rug, a floor lamp, green plants and simple framed prints. Warm white walls, bright natural daylight, cosy and uncluttered, photoreal interior photography.”

GPT Image 2: Outdoor natural-light headshotInput image for this GPT Image 2 result

GPT Image 2 · 1K · 3:4 · from the input image

Outdoor natural-light headshot

Prompt: “Make a professional headshot of the person in the photo outdoors. Keep the face, hair and identity exactly the same. Smart casual light-blue shirt, soft late-afternoon sunlight, a blurred green park behind, head and shoulders, relaxed genuine smile, 85mm lens look with shallow depth of field, photoreal.”

GPT Image 2: Black and white editorial portraitInput image for this GPT Image 2 result

GPT Image 2 · 1K · 3:4 · from the input image

Black and white editorial portrait

Prompt: “Create a black and white editorial portrait of the person in the photo. Keep the face, hair and identity exactly the same. Black turtleneck, dramatic side light from a window, dark seamless background, head and shoulders, confident neutral expression, fine film grain, magazine style, photoreal.”

GPT Image 2: Gold necklace on velvet

GPT Image 2 · 1K · 1:1

Gold necklace on velvet

Prompt: “Create a luxury jewelry product photo. Place a delicate gold necklace with a small pendant on deep green velvet. Soft spotlight from above with tiny sparkles on the metal, dark moody background, macro lens, very shallow depth of field, sharp focus on the pendant, high-end jewelry advertising style.”

GPT Image 2: Perfume bottle with a water splash

GPT Image 2 · 1K · 1:1

Perfume bottle with a water splash

Prompt: “Create a perfume advertising photo. A clear glass perfume bottle with a gold cap stands on a wet black surface, a frozen splash of water around its base, soft rim light outlining the glass, dark blue gradient background, crisp reflections, premium fragrance campaign style.”

Edit a product photo without changing the product

Attach your photo and describe the new setting. Nano Banana Pro keeps the shape, color and label of the product while it rebuilds everything around it.

Made with Nano Banana Pro · 1K · 1:1

Nano Banana Pro: Same product, new backgroundInput image for this Nano Banana Pro result

Make posters and labels with text that reads right

Write the exact words in quotes. GPT Image 2 spells headlines, dates and labels correctly, so posters and packaging come out usable.

Made with GPT Image 2 · 2K · 3:4

GPT Image 2: Event poster with readable text

Create photoreal scenes without a camera

Describe the place, the light and the lens. Seedream 4.0 renders realistic streets, interiors and portraits at 2K.

Made with Seedream 4.0 · 16:9

Seedream 4.0: Photoreal street scene at golden hour

Draft illustrations in seconds for a few credits

Try ideas before you commit. Z-Image Turbo turns one sentence into a watercolor, vector or flat illustration for 4 credits.

Made with Z-Image Turbo · 4:3

Z-Image Turbo: Watercolor illustration from one sentence

How it works

Three steps, start to file

  1. 01

    Pick a model and describe the image

    Write what you want to see. To edit a photo, attach it as a reference.

  2. 02

    Check the price and generate

    The credit price is on the button before you click. Sign in at this step; new accounts get 30 free credits.

  3. 03

    Download, reuse or animate

    Results stay in your library. Reuse the prompt, or turn any image into a video.

Models

Pick the right image model

What you can do

More image features

01

Edit your own photos

Attach up to four reference images and describe the change: new background, new outfit, a different season. For a ready-made coastal illustration poster, use the GTA 6 AI photo generator.

02

Readable text

GPT Image 2 renders signs, labels and posters with spelled-out words instead of letter soup.

03

Up to 4K

Draft at 1K for a few credits, then render the keeper at 2K or 4K on GPT Image 2 or Nano Banana Pro.

04

Batches of four

Run up to four variations at once. Each is priced and refunded separately.

05

Image to video

Any image you make can become the first frame of a video with one click.

06

One library

Every result is saved with its prompt and settings, ready to download or reuse.

Guides

Image model prices, compared

FAQ

Common questions

Do I need an account or a card?

You can set up a generation without an account. To run it, sign in with email or Google. New accounts get 30 free credits after confirming their email, with no card needed.

What file do I get?

The image file the model returns (JPG, PNG or WebP) at the resolution you picked. Results made with free credits carry a small Orvlex mark.

How long does an image take?

Usually under a minute. Busy models can take a few minutes; the result appears here and in your library when it is ready.

Does it work on a phone?

Yes. The generator works in mobile browsers, and you can upload photos straight from your camera roll.

Which image model should I start with?

GPT Image 2 at 1K is the cheapest and handles text in images well. Use Nano Banana Pro when you edit photos of people and need faces to stay the same, and Seedream 4.0 for photoreal scenes.

Can I edit an existing photo?

Yes. GPT Image 2, Seedream 4.0 and Nano Banana Pro accept up to four reference images. Describe the change you want and attach the photo.

What resolutions are available?

GPT Image 2 and Nano Banana Pro go up to 4K. Seedream 4.0 renders at 2K and Z-Image Turbo at 1K. Higher resolutions cost a few more credits.

Can I make several images at once?

Choose 1 to 4 images per run. Each one is priced and refunded separately, so one failure does not cost you the others.

Make an image now

New accounts get 30 free credits after confirming their email. Plans and packs are on the pricing page.

Start creating