The new Veo Video is live.Contact us
Text to image and image to image, six models

AI Image Generator with Nano Banana, Nano Banana Pro and more

Describe a picture or upload a photo, pick a model, and get a finished image in the shape you need. Nano Banana is ready by default; switch to Nano Banana Pro for readable text and 4K files, or to a lighter model for quick drafts. Any image you make can also become the first frame of a video.

Try Agent modeNEW

Describe your idea in a conversation and keep refining it.

Try Agent
Number of outputs

Production preview

Result
Examples

Three AI images, three models

Each image was generated with the model named on its card, from the prompt shown. Copy a prompt, or send it to the generator above and pick the same model.

  • Nano Banana 2 text to image: a red-and-white lighthouse on a cliff at dusk with a keeper in a yellow raincoat carrying a lantern
    Nano Banana 2 · text to image · 1K

    A film still made to open a video

    Prompt

    Cinematic wide shot of a red-and-white lighthouse on a rocky cliff at dusk. Waves break on the rocks below. A lone keeper in a yellow raincoat walks up the cliff path carrying a glowing lantern. Stormy purple sky with the first stars appearing. Photorealistic film still, natural light, no text.

    • Mode: Text to image
    • Size: 16:9
    • Resolution: 1K
  • GPT Image 2.5 Flare text to image: a white running shoe with a mint sole splashing above a puddle on a neon-lit street, headline Run the City
    GPT Image 2.5 Flare · text to image · 1K

    An ad banner with the headline in the picture

    Prompt

    Advertising banner photograph: a white running sneaker with a mint green sole floating above a puddle on a city street at night, the water splash frozen mid-air, neon shop lights reflected in the wet asphalt. Bold white headline on the left side reading "RUN THE CITY". Photorealistic, high-end sports campaign look. No other text, no logos.

    • Mode: Text to image
    • Size: 16:9
    • Resolution: 1K
  • Nano Banana 2 Lite text to image: a watercolor of a rainy cobblestone street with a glowing bookshop window and a black cat under the awning
    Nano Banana 2 Lite · text to image

    A watercolor scene as a quick draft

    Prompt

    Full-bleed wide watercolor illustration filling the entire 16:9 frame edge to edge, no paper border, no margins. A narrow old-town street with a small bookshop on a rainy evening. Warm light glows from the shop windows, puddles on the cobblestones reflect the light, a black cat sits dry under the shop awning, a person with an umbrella in the distance. Soft loose brushwork, muted blues and warm ambers. No text, no signs with letters.

    • Mode: Text to image
    • Size: 16:9

Text to image or image to image: how the AI image generator works

An AI image generator turns a written description into a picture, and here it works two ways. Text to image starts from your prompt alone: describe the subject, the setting, the light and the style, choose an aspect ratio, and the model draws it. Image to image starts from photos you upload — a product shot, a portrait, a sketch — and your prompt says what to change: replace the background, restyle the picture, combine several photos into one scene, or put the same character somewhere new. Every model here does both. They differ in resolution, how well they handle text inside the image, how many photos they accept, and what each image costs.

Models

Which AI image model to use: Nano Banana, Nano Banana Pro, GPT Image 2.5 Flare or P-Image

Start with Nano Banana for everyday pictures and photo edits. Move to Nano Banana Pro when the image needs readable text or a 2K–4K file, to Nano Banana 2 for all-round work at higher resolution, and to Nano Banana 2 Lite or P-Image when you want many quick drafts.

DefaultNano BananaEveryday images and photo editsGenerateNano Banana 2All-round work up to 4KNano Banana 2 LiteFast 1K draftsNano Banana ProReadable text, products, complex layoutsGPT Image 2.5 FlareSocial posts, ads and product shotsP-ImageThe lowest-cost drafts
Resolution1K1K, 2K, 4K1K1K, 2K, 4K1K, 2K, 4KFixed size
Aspect ratios10 plus auto14 plus auto14 plus auto10 plus auto8 plus auto7
Photos you can uploadUp to 10Up to 14Up to 10Up to 8Up to 16Up to 10
Credits per image48 (1K)420 (1K)3 (1K)1 (text to image)
From prompt to picture

How to use the AI image generator

  1. 1

    Pick a model and a starting point

    Nano Banana is selected by default; switch models in the generator. Choose text to image to start from a description, or image to image to start from your photos.

  2. 2

    Describe the picture

    Say what is in it, where it is, how it is lit and what style it should have. Put any words that must appear in the image in quotes. For an edit, say what to change and what to keep.

  3. 3

    Set the shape and size

    Pick 1:1 for a post or product tile, 9:16 for a story, 16:9 for a banner or a video frame. On models with a resolution choice, use 1K to test and 2K or 4K for the final file.

  4. 4

    Generate and refine

    Make up to four takes at once and keep the best. To adjust one, upload it as a starting image and describe the next change.

Turn an AI image into the first frame of a video

A picture you make here can open a video. Generate it in the shape the video needs — 16:9 for YouTube and the web, 9:16 for Shorts, Reels and TikTok — then open the AI video generator, choose image to video and upload it as the start frame. Veo 3.1 or MiniMax H3 Max animates it from there, with sound generated in the same clip. The lighthouse above was made with Nano Banana 2 at 16:9 for exactly this.

Model guides

Nano Banana and Nano Banana Pro in detail

Each page covers what the model is good at, example prompts, and how it compares.

FAQ

AI Image Generator FAQ

Make your first AI image

Type a prompt or upload a photo — Nano Banana is ready to go.