Nano Banana | Google Gemini

Nano Banana is image generation powered by Google Gemini. Write a description of what you want to see, and the AI generates it — people, places, products, scenes, artwork, anything you can describe in words. Three model tiers let you choose between speed and quality depending on what you need.

Nano Banana is the fast option for quick iterations and testing ideas. Nano Banana Pro balances speed with higher quality output. Nano Banana 2 is the top tier with a unique thinking mode — the AI actually plans the composition, lighting, and coherence before it starts generating, producing more intentional and polished results.

The reference image system is what makes this tool powerful for consistent character work. Upload up to 14 reference photos of your character, label them, and mention the labels in your prompt. The AI uses those references to keep the same face, body, and style across every generation — essential for building a recognizable AI persona across hundreds of posts.

Resolution scales from 1K for fast drafts up to 4K for maximum detail — capturing skin texture, individual hair strands, and catchlight in the eyes. Seven aspect ratios cover every platform format from square Instagram posts to vertical Stories and wide YouTube thumbnails. Generate up to 9 variations per prompt to explore different interpretations before committing.

Available Models

Nano Banana 2 Recommended

gemini-3.1-flash-image-preview

Fast, 4K, 14 refs, thinking mode. Breakthrough micro-detail and text rendering.

Nano Banana Pro Premium

gemini-3-pro-image-preview

Professional model with advanced reasoning. Best instruction-following and compositions.

Nano Banana Budget

gemini-2.5-flash-image

Speed and efficiency optimized. High-volume, low-latency tasks.

Best results

📝

Write Scenes, Not Keywords

Gemini excels at understanding natural language. Write a narrative paragraph describing the scene — what is happening, the environment, the mood, the lighting. This produces far better results than a list of keywords separated by commas.

📷

Use Photography Terms for Realism

For photorealistic results, mention camera specifics — shot on 85mm lens, golden hour lighting, shallow depth of field, visible skin pores. These terms tell the AI to simulate real photography rather than illustration.

🖼️

3 to 6 Reference Images for Characters

Upload 3 to 6 clear reference photos of your character at 1024px or higher. Include different angles and lighting conditions. Quality plateaus around 6 images — more than that rarely improves consistency.

🧠

Enable Thinking Mode on Nano Banana 2

Thinking mode makes the AI plan composition and lighting before generating. It takes slightly longer but produces more intentional, coherent results — especially for complex scenes with multiple elements.

🎯

4K for Final Content, 1K for Testing

Use 1K resolution while iterating on your prompt — it generates faster and costs less. Switch to 4K only when you are happy with the composition and want maximum detail for publishing.

💬

Provide Context and Intent

Tell the AI what the image is for. A professional Instagram photo for a luxury fashion brand produces very different results from a photo of a woman — context shapes composition, lighting, and overall quality.

Gallery

@furiousteam

Guides

Nano Banana

🖼️ Image

✨ Gemini 5 min read

Nano Banana — Technical Guide

Generate AI images powered by Google Gemini — three model tiers from fast drafts to 4K maximum detail, with reference images for character consistency and a thinking mode that plans the composition before creating

🍌

Try Nano Banana

No subscription required. Pay only for what you create.

Start Creating →

🍌Nano Banana