Google Unveils Nano Banana 2.1 Image Model With Enhanced Character Consistency and 50% Cost Cut

Google has launched Nano Banana 2.1, a new image generation model that significantly improves character consistency while cutting costs in half, becoming available across its Gemini ecosystem starting October 6, 2026. The model, part of the Gemini 3 series and built on Gemini 3.6 Flash architecture, represents a focused refinement of Google's image generation capabilities rather than a generational leap. It now processes up to 14 reference images per request, maintaining visual consistency for up to four distinct characters and ten objects—a substantial improvement over its predecessor.