Google unveiled a new image generator called Nano Banana 2

Google has announced an updated image generation model, Nano Banana 2 (Gemini 3.1 Flash Image), which will replace Nano Banana Pro in the Gemini app.

Subscribers to Google AI Pro and Ultra will retain access to the Pro version. The new model will also roll out to AI Mode and Lens in Search, appear in preview in AI Studio and via the Gemini API, on the Google Antigravity developer platform, in Vertex AI, and will become the default model in the Flow video generation service.

The model builds on Gemini’s “world knowledge” and uses Search data and images to generate more accurate visuals. Google highlighted the model’s ability to clearly generate and translate text. Compared to its predecessor, Nano Banana 2 shows improved consistency: it can keep the appearance of up to five characters and the look of up to 14 objects unchanged, enabling the creation of source frames for videos with preserved detail.

The model follows instructions more precisely and handles complex prompts, “capturing the specific nuances of your idea.” Users can adjust aspect ratios and resolution from 512px up to 4K.