Google has launched Nano Banana 2, an upgraded AI image generation model officially branded as Gemini 3.1 Flash Image. The new version focuses on speed while delivering more realistic visuals and is now the default image model across all modes of the Gemini app, including Fast, Thinking, and Pro. This update builds on the success of the original Nano Banana, released in August 2025, and the later Pro version, which emphasized higher image fidelity.
Nano Banana 2 supports image generation from 512 pixels up to 4K in multiple aspect ratios. It can maintain visual consistency for up to five characters and accurately render up to 14 objects within a single workflow, enabling more coherent visual storytelling. The model also handles complex prompts with greater precision and improves text generation within images, producing clear, readable typography that can even be automatically translated and localized. Sharper details, richer textures, and more dynamic lighting further enhance output quality.
The rollout extends beyond the Gemini app. Nano Banana 2 becomes the default image model in Google Search via Google Lens and AI Mode across 141 countries and is integrated into the Flow video editing tool. Users on Google AI Pro and Ultra plans retain access to Nano Banana Pro for specialized tasks. For developers, the model is available in preview through the Gemini API, Gemini CLI, Vertex AI, AI Studio and Antigravity. All generated images include a SynthID watermark and support C2PA Content Credentials, reinforcing transparency and verification of AI-generated media.