Nano Banana Pro Generator

Google’s Gemini 3 Pro Image model, online: infographics and diagrams, posters with legible text in many languages, product mockups and edits that combine up to 8 of your photos, in up to 4K. No Google AI plan, from 5 credits a picture.

Make a Gemini 3 Pro image online

Model

Prompt

0/20,000

Format

Aspect ratio

Size

2K · Auto · No watermark

A battle robot covered in cables and armour striding out of a wall of fire, made with Google’s Gemini 3 Pro Image modelExample

What is Nano Banana Pro?

Gemini 3 Pro Image is Google DeepMind’s image generation and editing model built on Gemini 3 Pro, released on November 20, 2025, as the Pro member of Google’s Nano Banana family. Google calls it its reasoning-driven model for professional work: complex graphic design, high-fidelity product mockups and factual data visualizations with accurate text. On Melies AI it runs in your browser, with up to 8 images per edit and no Google AI plan.

Google’s launch post names four strengths. The model turns notes and facts into infographics and diagrams. It renders legible text in many languages, from a short tagline to a long paragraph, and can translate the text in a picture. It blends up to 14 images while keeping up to 5 people recognisable. And it changes camera angle, focus, colour grading and lighting, such as day to night, on request.

This site sends your prompt and images only. Google’s own API can also ground a picture in Google Search results, such as live weather or sports scores; that option is not switched on here. In Google’s October 2026 tests, the newer Nano Banana 2.1 scored higher than the Pro model on overall preference, so try both on the same prompt.

Pick Auto or one of 10 shapes from 21:9 to 9:16, and 1K, 2K or 4K. A square is 1024 × 1024, 2048 × 2048 or 4096 × 4096 pixels, and a 16:9 picture at 4K is 5504 × 3072. Prompts run up to 20,000 characters. Save as PNG or JPEG. Google embeds an invisible SynthID watermark in every picture; no visible watermark is added here.

Largest size
4K
Reference images here
8
Aspect ratios + Auto
10
Prompt characters
20,000
Released by Google
2025
Credits · one 2K image
5

By Google · run through a third-party API · not affiliated with Google · Last updated

What Google’s Pro image model does well

Five jobs Google built Gemini 3 Pro Image for, from data graphics to campaign mockups.

An infographic titled Inside the Transformer LLM that follows a prompt through tokenization, attention layers and a softmax to the output token, with readable labels
01

Infographics and diagrams from your notes

Paste notes, a recipe or a list of facts and ask for an infographic, a step-by-step diagram or an explainer. The model lays out headings, labels and icons and keeps the text readable. Check facts in the result, since this site does not connect it to Google Search.

A Japanese product page for an anime figure, with Japanese specs, an English caption under each detail close-up and a stylised logo, all legible
02

Legible text in many languages

Google built the Pro model for text inside pictures: taglines, poster titles, menus and longer paragraphs, in a range of fonts, textures and calligraphy. Upload a design and ask for the same text in another language to localise a campaign.

Four studio portraits beside the same four people in a desert fashion editorial at golden hour, each face, hairstyle and pair of glasses unchanged
03

Up to 8 images, people kept consistent

Combine people, products and places from several photos into one scene, such as a fashion editorial with the same faces in a new location. Google’s own API takes up to 14 images and 5 people; this site takes up to 8 images per edit.

The same cobbled café street in midday sun and, after the edit, at night with warm street lamps, string lights on the awning and blurred walkers behind
04

Change light, focus and camera angle

Ask in words for studio-style changes to a photo: turn day into night, add a bokeh background, move the focus onto one subject, change the camera angle or apply a colour grade. Name the photo @Image1 and describe one change at a time.

A marker sketch of a sage green TIDE water bottle with a bamboo cap and carry loop, beside the same bottle as a photo-real studio shot
05

Product mockups that keep your brand

Turn a sketch into a product shot, a blueprint into a photo-real building, or apply one look to a set of mockups so the branding stays the same on every piece. Render the final at 2K or 4K for print.

The Pro model vs Nano Banana 2.1

Both Google models are in the same image generator. The figures come from Google’s docs and its October 2026 model card.

ProNano Banana 2.1
Google modelGemini 3 Pro Imagegemini-nano-banana-2.1
Built onGemini 3 ProGemini 3.6 Flash
ReleasedNovember 20, 2025October 6, 2026
Google pitches it forComplex design, product mockups, data graphicsFast, efficient everyday generation and editing
Sizes1K, 2K, 4K1K, 2K, 4K
Aspect ratios10, from 21:9 to 9:1614, from 8:1 to 1:8
Reference images here810
Overall preference (Elo)9351050 (thinking on)
Price here5 credits (1K, 2K), 6 (4K)1, 2 or 3 credits

How to use Google’s Pro image model

  1. 01

    Pick a tab

    Text to image to start from words, or Image edit to start from up to 8 of your own pictures.

  2. 02

    Write the prompt

    Describe the subject, layout, light and camera, and put any text in quotes with its language. In an edit, name your images @Image1, @Image2, say what to change and what to keep.

  3. 03

    Choose shape and size

    Auto or one of 10 shapes, and 1K, 2K or 4K. The price is on the button: 5 credits up to 2K, 6 at 4K.

  4. 04

    Generate and download

    Press Generate and save as PNG or JPEG. The picture stays under “My images” for 14 days.

Gemini 3 Pro Image pricing per picture

One price per picture: 5 credits at 1K or 2K and 6 at 4K. Edits cost the same however many images you upload. Packs from $9.90, no subscription, failed images refunded.

Model1K2K4K
Pro5 credits5 credits6 credits

Gemini 3 Pro Image FAQ

Something else? Email support@meliesai.com.

Not on its own. In the Gemini app, Google AI Pro, Plus and Ultra subscribers can redo a picture with the Pro model, and free users get an earlier Nano Banana. On Melies AI a Pro picture costs 5 credits at 1K or 2K. The 4 free sign-up credits are not enough for one, but they cover Nano Banana 2.1 pictures. Packs start at $9.90.

A Gemini 3 Pro Image picture costs 5 credits at 1K or 2K and 6 credits at 4K. Edits cost the same, whether you upload 1 image or 8. The $9.90 pack has 100 credits, enough for 20 pictures at 2K. There is no subscription, and failed pictures are refunded.

It depends on the job. In Google’s October 2026 human tests, Nano Banana 2 scored higher on overall preference and infographic design, while the Pro model scored higher on keeping characters and products consistent in edits and on infographic factuality. The newer Nano Banana 2.1 beat both on every measure Google listed.

Start with Nano Banana 2.1 for most work: Google’s tests score it higher, it adds 8:1 and 1:8 panoramas, and it costs 1 to 3 credits here. Use Google’s Pro model when you want its look for a design, mockup or text-heavy poster. Both are in the same generator, so run one prompt on each and compare.

Yes, they are one model. Gemini 3 Pro Image is the official name in Google’s API, and Google uses the Pro Nano Banana name in the Gemini app and its launch post. Google DeepMind released it on November 20, 2025, built on Gemini 3 Pro. The model code is gemini-3-pro-image.

In the Gemini app, yes. Google says Google AI Pro, Plus and Ultra subscribers get its newest model, Nano Banana 2.1, and can regenerate a picture with the Pro model through “Redo with Pro”. In Google AI Studio and the Gemini API it is billed per use instead. This site needs no Google plan: you pay per picture with credits.

Open the Image edit tab, add 1 to 8 photos and describe the change. Name each photo @Image1, @Image2 in the order you added them, say what to change, such as day to night, and say what must stay the same, such as the people. Pick Auto or a shape and a size, then press Generate.

Up to 8 images per edit on this site, as JPG, PNG or WebP. Google’s own API takes up to 14 and keeps up to 5 people consistent. Images are numbered in the order you add them and called @Image1, @Image2 and so on in the prompt. The price stays the same however many you add.

Three sizes, 1K, 2K and 4K, where a square is 1024, 2048 or 4096 pixels a side. 10 shapes plus Auto: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 5:4, 4:5, 21:9. At 4K a 16:9 picture is 5504 × 3072 and a 21:9 banner is 6336 × 2688. For 8:1 or 1:8 panoramas, use Nano Banana 2.1.

Not here. Google’s API can ground a Gemini 3 Pro Image picture in Google Search results, for live weather, sports scores or recent facts. This site sends your prompt and images only, with grounding switched off. For an infographic, put the facts in your prompt yourself and check the result.

No visible watermark is added to pictures made on Melies AI. Google embeds an invisible SynthID watermark in every picture its tools make, which the Gemini app can detect. In the Gemini app itself, Google adds a visible sparkle mark on free and Google AI Pro pictures. Each picture stays under “My images” for 14 days.

Yes. Google offers it as gemini-3-pro-image in the Gemini API and Google AI Studio, and in Google Cloud. That route needs your own Google account, billing and code. Melies AI has no API of its own: it is a website where you make pictures in the browser and pay with credits.

Yes. Pictures you generate are yours to use, including in ads, client work and products, as long as your prompt and uploaded images do not infringe anyone else’s rights and follow our content standards. Edits of real people need their consent.

No. Melies AI is an independent platform that gives you access to Google’s Gemini 3 Pro Image model through a third-party API provider. It is not affiliated with, endorsed by or sponsored by Google. Google’s own places to use it are the Gemini app, Google AI Studio and Google Flow.

Make your first Pro picture

Write a prompt or upload your photos, pick a shape and a size up to 4K, and download the result.

Create an image