Google unveils Gemini 2.5 Flash Image with advanced generation and editing capabilities

Google has announced Gemini 2.5 Flash Image — code-named “nano-banana” — its latest image generation and editing model. This release introduces features that let users blend multiple source images, ensure character consistency for storytelling or branding, perform targeted edits using natural language, and leverage world knowledge for creative image tasks.

These capabilities mean users can maintain the same character across various environments, generate unified brand assets, and showcase products from different angles while preserving key features. Image blending also enables merging of separate images, such as incorporating new objects into scenes or restyling rooms, all through a single prompt.

Following the new feature set, Gemini 2.5 Flash Image enables detailed transformations. Users can blur backgrounds, remove individuals or objects, alter poses, or colorize images simply by describing the desired change. These edits are localized and do not require pixel-level selections.

While the model is currently in preview, developers can already access it via the Gemini API and Google AI Studio, with enterprise integration through Vertex AI. To support transparency, all images produced with Gemini 2.5 Flash Image will include an invisible SynthID digital watermark to identify AI-generated or edited content.