Google launches Nano Banana 2 for AI image creation

Google introduces Nano Banana 2, an image generation and editing model that combines some of the advanced capabilities of Nano Banana Pro with greater speed. Its technical name is Gemini 3.1 Flash Image, and it is starting to roll out across several Google products.

The main difference is the balance between speed and control. Nano Banana 2 is designed to help you generate images, fix details, and try multiple versions without waiting as long, while retaining features that were previously reserved for the Pro model.

What Nano Banana 2 can do

The model uses Gemini’s knowledge and, in some products, recent information and images from Google Search. This helps it represent specific topics more accurately and create visual materials that require context.

For example, it can turn notes into a diagram, create an infographic about the water cycle, or prepare a data visualization. It can also generate readable text within an image and translate or adapt it into another language, which is especially useful for posters, cards, and marketing designs.

Its capabilities include:

• Preserving the likeness of up to five characters within the same workflow.
• Maintaining the fidelity of up to 14 objects across different images or scenes.
• Following complex instructions more accurately.
• Creating images in different formats and resolutions, from 512 pixels to 4K.
• Generating results with more defined lighting, textures, and details.

In practice, this makes tasks such as creating a storyboard without changing the characters’ appearance, preparing a social media campaign in multiple sizes, or adapting a design for different countries easier.

Where it will be available

Nano Banana 2 is gradually being added to several Google services:

• Gemini: replaces Nano Banana Pro in the Fast, Thinking, and Pro models. Google AI Pro and Ultra subscribers retain access to Nano Banana Pro for specialized work and can regenerate images from the three-dot menu.
• Google Search: arrives in AI Mode and Lens, both in the Google app and in mobile and desktop browsers. Availability is expanding to 141 new countries and territories and eight additional languages.
• AI Studio and Gemini API: available in preview, including within Google Antigravity.
• Google Cloud: available in preview through the Gemini API in Vertex AI.
• Flow: becomes the default image generation model and is available to all Flow users without consuming credits.
• Ads: used to provide suggestions while creating campaigns.

You can also try it through the new templates in the Gemini app.

How AI-generated images will be identified

Google maintains two systems that provide information about content origins. SynthID embeds a digital signal that helps detect images, videos, and audio generated by its models. Since its launch in November, the SynthID verification feature in Gemini has been used more than 20 million times in different languages.

The company is also preparing to bring C2PA credential verification to Gemini. This standard can indicate not only whether AI was used, but also how the content was created or modified.

For you, the most important change is that generating images from detailed instructions, with readable text and consistent characters, is no longer as limited by speed or access to Pro models. The difference between Nano Banana 2 and Nano Banana Pro will mainly come down to the maximum level of fidelity and factual accuracy each task requires.