Gemini Image

Gemini Image, marketed under the name Nano Banana, is Google DeepMind's family of AI models for generating and editing images, built on Gemini. The lineup descends from earlier Google text-to-image work, including the original Imagen models, and now spans several tiers from a fast, low-cost version to a studio-quality version built on Gemini 3.

A user describes what they want in everyday language, optionally uploading reference images alongside written instructions, and the model generates a detailed image using deep language understanding to interpret the prompt. The conversation can continue afterward: a user can ask for specific edits, a new camera angle, a different pose, or a full change of artistic style, and the model updates the image while keeping the same subject and setting.

The model family includes Gemini 3 Pro Image (Nano Banana Pro) for studio-quality precision and control, Gemini 3.1 Flash Image (Nano Banana 2) for pro-level results at higher speed, and Gemini 3.1 Flash-Lite Image (Nano Banana 2 Lite) for the fastest, lowest-cost generation and editing. All of the models are natively multimodal and are built to follow real-world logic and detailed prompts.

Gemini Image is accessible through the Gemini app, Google AI Studio, Google Pics, and the Gemini API for developers building their own products, as well as through the Gemini Enterprise Agent Platform for building and governing agents. A prompt guide covers style, subject, setting, action and composition, plus techniques for rendering precise text, translating and localizing ideas, and producing images at a specific aspect ratio.

Use cases

• Generating a multi-image story from one reference photo
• Translating and localizing text shown inside an image
• Producing production-ready assets at a set aspect ratio