Build with Nano Banana 2, our best image generation and editing model

Google has launched Nano Banana 2 (Gemini 3.1 Flash Image), which enhances image generation and editing with higher fidelity and speed, providing an excellent price-performance ratio for scalable visual creation. Key improvements include leveraging Gemini's world knowledge for visually grounded images via web search, exemplified by the "Window Seat" app. The model also features upgraded text rendering for crisp, accurate text within images, supporting in-image localization for multiple languages, as demonstrated by the "Global Ad Localizer" demo. Creative control is enhanced through native support for new aspect ratios (4:1, 1:4, 8:1, 1:8), a new 512px resolution tier for efficiency, improved instruction following for complex prompts, and configurable thinking levels to control the model's reasoning before rendering. Nano Banana 2 is available now via the Gemini API in Google AI Studio (requiring a paid API key), for enterprise deployment on Vertex AI, and in Google Antigravity and Firebase.

(Source：Gemini)