Nano Banana 2.1
More natural images, clearer designs and more consistent subjects. Google’s updated model for efficient image generation and editing.
Official Nano Banana 2.1 showcases
Google DeepMind text-to-image examples with original English prompts and source attribution. These were not generated on this site.
What improves in Nano Banana 2.1?
Creative improvements described in Google’s documentation and demonstrations.
Visual design and typography
Improves text rendering, infographic layouts and prompt adherence for posters, packaging and content graphics. Check small text, spelling and numbers.
Multiple references, consistent subjects
Google supports up to 14 reference images and emphasizes character and object consistency across edits, useful for exploring series and product scenes.
Mask and sketch guided edits
Official demonstrations show changing environments while preserving subjects with mask or sketch guidance. Complex blending can still introduce artifacts.
More natural high resolution images
Google lists 1K, 2K and 4K output and improved tiling on wide formats, alongside better materials, lighting and fine detail.
Nano Banana 2.1 questions
Create with Nano Banana 2.1
Enter a prompt or add references, then review settings and credits before submitting.