Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.imagine.art/llms.txt

Use this file to discover all available pages before exploring further.

IMAGE MODELby Google DeepMindGemini 3 Pro Image

Nano Banana Pro

Google’s Gemini 3 Pro Image model — built for production-ready complex compositions. Supports up to 14 reference images, localized edits and camera transformations, and uses Google Search grounding for real-world accuracy. The professional tier in the Nano Banana lineup.

Identity Preservation
Up to 14
Resolution
Up to 4K
Context Window
65,536 tokens
Released
November 2025

Built for complex, production-grade work

Nano Banana Pro is powered by Gemini 3 Pro Image — Google’s most capable image generation model. Where Nano Banana 2 excels at speed and volume, Nano Banana Pro excels at depth: multi-subject identity preservation, precise localized edits, camera transformations, and highly detailed prompt adherence. It’s the right choice when your output needs to hold up to client scrutiny.
Final Plant Width 1000 Format Webp

Capabilities

Google Search grounding

Real-world grounding via Google Search produces accurate depictions of real products, brands, places, and entities without overprompting.

High-fidelity text rendering

Industry-leading text quality, including long passages, multilingual layouts, data visualizations, and typographic compositions.

Complex multi-image blending

Blend and composite multiple reference images while maintaining consistent lighting, perspective, and style across the final output.
Fluffy Monsters Width 1000 Format Webp

Specifications

FeatureDetails
Underlying modelGemini 3 Pro Image
ResolutionUp to 4K (flexible aspect ratios)
Reference imagesUp to 14
Context window65,536 input tokens
Web groundingYes (Google Search)
EditingLocalized, lighting, focus, camera
ReleasedNovember 20, 2025
Astronaut

How to use

1

Open the AI Image Generator

Go to the ImagineArt AI Image Generator.
2

Select the model

From the model dropdown, choose Nano Banana Pro.
3

Write your prompt

Be specific about subjects, lighting, and composition. Nano Banana Pro follows complex multi-part instructions reliably.
4

Attach reference images (optional)

For multi-subject consistency, upload your reference images. The model preserves identity across all subjects.
5

Generate and refine

Review the output and use localized edit instructions to refine specific areas without regenerating the whole image.
Nano Banana Pro WM Weather App

Prompting tips

  • Name your subjects explicitly — “A woman in a red blazer (Subject A) and a man in a navy suit (Subject B) shaking hands” gives the model clear anchors for identity preservation.
  • Use localized edit instructions — After generating, follow up with targeted edits: “Change the background to a blurred office interior” or “Shift the lighting to golden hour.”
  • Let Google grounding handle real-world references — For known brands, locations, or products, name them directly. Nano Banana Pro will render them accurately.

Example prompts

A product shoot featuring a cream-colored luxury skincare bottle (Subject A) and a matching moisturizer jar (Subject B) on a pale stone surface, dramatic side lighting, high-end editorial photography.
A tech CEO (Subject A) being interviewed on stage at a large conference, blurred audience in background, professional event photography, warm stage lighting.

Compare models

ModelBest forIdentity preservationSpeed
Nano Banana ProComplex compositions, productionUp to 5 subjectsFast
Nano Banana 2Rapid iteration, high volumeUp to 14 refs~4× faster
Nano BananaQuick edits, product swapsUp to 4 refsNear real-time
Nano Banana Pro is the recommended choice when final output needs to meet professional quality standards. Use Nano Banana 2 for early-stage exploration, then switch to Pro for the polished deliverable.