Google’s Gemini 2.5 Flash Image Model Moves Into Production with New Aspect Ratios

Google has announced that its Gemini 2.5 Flash Image model, also known internally by the nickname Nano Banana, is now production-ready and supports a broader range of aspect ratios.

"Good news for developers: Gemini 2.5 Flash Image is now stable and ready for scaled production, in addition to coming with a new aspect ratio setting + the ability to specify image only output," Logan Kilpatrick, Group Product Manager at Google DeepMind, said.

The upgraded model enables developers and creators to generate and edit images with unprecedented flexibility. It now supports ten aspect ratios covering landscape, portrait, square, and flexible formats—ideal for use cases spanning cinematic, social media, and mobile layouts.

Alongside broader format support, Gemini 2.5 Flash Image improves image editing capabilities: users can blend multiple images, maintain character consistency across edits, perform prompt-based transformations, and leverage Google’s world knowledge to inform visual output.

The model continues to be accessible via the Gemini API, Google AI Studio, and Vertex AI for developers and enterprises.

"Gemini 2.5 Flash Image empowers users to seamlessly blend multiple images, maintain consistent characters for richer storytelling, perform targeted edits with natural language, and leverage Gemini's extensive world knowledge for image generation and modification," Google said in a blog post.

In addition, Google includes SynthID watermarking in the model outputs, an invisible tag to help trace AI-generated images back to source.

Google highlights early adopters as proof points: Cartwheel uses the model for precise character rendering, while Volley generates dynamic visuals in live gaming scenarios.

With Nano Banana now production-ready, Google is positioning Gemini 2.5 Flash Image not just as a research demo but as a foundation for real-world creative applications, content tools, and generative services that demand precision, consistency, and versatility in visuals.