Google has officially launched Nano Banana 2.1, the latest iteration of its AI model designed for text-to-image generation and photo editing. The tool is now available across several platforms, including the Gemini app, Google Search, Google Ads, and developer environments like Google AI Studio, Flow, and Stitch.
The new model introduces improvements in visual design, mask-based editing, and subject consistency. In internal ELO testing, Nano Banana 2.1 achieved 1,050 points, surpassing the 990 points of its predecessor and the 935 points of the Pro version. The system can process up to 14 reference images, tracking four characters and ten objects simultaneously.
Technical capabilities include outputting images at 4K resolution with aspect ratios up to 8:1. Users can adjust the model's processing time and enable integration with Google Search and Google Image Search to ground generated results in real-world data.
The company has significantly reduced API costs for the new model. A standard 1K image now costs $0.0336, compared to $0.067 for the previous version. Similarly, 4K image generation is priced at $0.0756, down from $0.151. Additional discounts of 50% are available for bulk batch processing.
