Google Nano Banana 2 is here with Flash-speed image generation and Pro-grade features: Here’s how to use it in Gemini

Nano Banana 2

• Google unveils Nano Banana 2 for faster, high-quality image creation
• Model offers improved world knowledge and visual fidelity
• Nano Banana 2 now integrated into Gemini app and Google products

Did our AI summary help?

Google has introduced Nano Banana 2, its latest image generation model, combining the advanced capabilities of Nano Banana Pro with the speed of Gemini Flash. The new model, also referred to as Gemini 3.1 Flash Image, is rolling out across Google products starting today.

Nano Banana 2 aims to bridge the gap between high-quality visual output and rapid generation. According to Google DeepMind, the model delivers improved world knowledge, stronger instruction following and enhanced visual fidelity while maintaining fast response times.

What’s new in Nano Banana 2

One of the key upgrades is access to advanced world knowledge, allowing the model to generate more accurate subject representations. It can also use real-time information and web-based image grounding to improve outputs such as infographics and data visualisations.

The model improves text rendering and translation within images, enabling clearer marketing mockups, greeting cards and localized creative assets. It also introduces better subject consistency, maintaining resemblance for up to five characters and preserving up to 14 objects within a workflow.

Nano Banana 2 supports production-ready specifications, including multiple aspect ratios and resolutions from 512px up to 4K. Google says visual fidelity has also improved, with sharper details, richer textures and more balanced lighting.

How to use Nano Banana 2 in Gemini

Nano Banana 2 is now integrated directly into the Gemini app, where it replaces Nano Banana Pro across Fast, Thinking and Pro modes. Google AI Pro and Ultra subscribers can still access Nano Banana Pro by regenerating images via the three-dot menu inside the app.

Users can generate images by entering descriptive prompts or selecting style templates within the Gemini interface. The new templates feature allows users to pick predefined artistic styles before generating an image.

The model is also available in AI Studio, the Gemini API, Google Cloud’s Vertex AI, Flow, Google Ads, and in AI Mode within Search and Lens across mobile and desktop.

Provenance and verification

Google is continuing to embed SynthID watermarking and C2PA Content Credentials into generated media. The company says its SynthID verification feature has been used over 20 million times to identify AI-generated images, video and audio.