Gemini Nano Banana 2.1

Google AI Studio unveiled Nano Banana 2.1, its most advanced image model, with notable improvements in reasoning, quality, and visual design, as reported on October 6, 2026. Users can upload their own images and videos to generate a custom AI avatar, and adjust camera angles, focus, texture, color, and style from any reference photo. The model supports 2K resolution, clear text rendering for logos and posters in multiple languages, and context-aware editing that combines two uploaded pictures into a single image. The release extends Nano Banana's figurine-style generation with upgraded features for converting favorite pictures into miniature figures, and adds retro-fashion trend transformations. Editing controls now include mask-based adjustments and better subject consistency, addressing prior model weaknesses in character preservation. The article frames the update as a Gemini-side rollout available through Google AI Studio, though no API endpoints, pricing, or benchmark numbers were disclosed in the source.