Landscapes to social media posts, Gemini Nano Banana adds new aspect ratios

Google expands Gemini 2.5 Flash Image with new formats, faster editing, and better character consistency, now available for developers through AI Studio and Vertex AI

Google has rolled out the Gemini 2.5 Flash Image model update with 10 new aspect ratios, faster editing, and improved character consistency. The update also applies to the AI tool Nano Banana, which lets users blend images and generate new ones with text prompts. According to Google, the model supports a wide range of formats, from cinematic 21:9 landscapes to vertical 9:16 social media posts.

 

Gemini 2.5 Flash Image model is available in the Gemini AI app with the Nano Banana tool and is accessible for developers through the Gemini API on Gemini AI Studio and Vertex AI for enterprises.

 

Gemini 2.5 Flash Image model: What’s new

According to Google, Gemini 2.5 Flash Image allows users to generate images from text, blend multiple visuals, and edit specific regions using simple language instructions. It can also maintain consistent characters across different images, a feature designed to support storytelling, gaming, and animation workflows.

 

Google also said that the model now supports a range of formats from cinematic landscapes to vertical social media posts. Supported ratios include:

• Landscape: 21:9, 16:9, 4:3, 3:2
• Square: 1:1
• Portrait: 9:16, 3:4, 2:3
• Flexible: 5:4, 4:5

Google tips for using Gemini Nano Banana 

Google has shared four tips to help users get the best results from the new feature.

• Subject consistency: Users can create a series of images that maintains character consistency irrespective of the changes and edits — whether it’s changing outfits, lighting, or backgrounds. Even when the surroundings are altered, the person or object remains the same.

• Precise edits: With Nano Banana, users can make pixel-perfect edits to specific parts of an image without affecting the rest. They can change an object’s colour, fix text on a sign, or adjust a logo using simple natural language prompts. The model also remembers previous edits, allowing for a smooth, step-by-step creative process.

• Natural prompts: Users can rely on Nano Banana’s conversational understanding to perform complex visual tasks with ease. This means that the user can use a simple and natural tone rather than complex prompts to make desired edits. This can be used for turning sketches into realistic scenes, restoring old photos, or merging multiple images into a single creative composition.

• Build an app in Canvas or AI Studio: Using platforms like AI Studio and Canvas, developers can design interactive experiences that let users edit or transform their own images. For example: “PictureMe,” an app that reimagines user photos in various styles, from classic portraits to cartoon-inspired looks.