TLDR: Google has officially launched its next-generation AI image generation model, previously codenamed ‘Nano-Banana,’ as Gemini 2.5 Flash Image. Available within the Gemini app for both paid and unpaid subscribers, and also via the Gemini API, Google AI Studio, and Vertex AI, this new model, developed by Google’s DeepMind, focuses on significantly improving character and scene consistency in AI-generated images. It also introduces advanced multi-turn editing capabilities and incorporates visible and invisible watermarks for safety.
Google has officially confirmed the widespread availability of its highly anticipated next-generation AI image generation model, now formally known as Gemini 2.5 Flash Image. This advanced tool, which had garnered significant attention under its mysterious codename ‘Nano-Banana,’ is now accessible to all Gemini app users, including both paid and unpaid subscribers. Furthermore, its capabilities are extended to developers and enterprises through the Gemini API, Google AI Studio, and Vertex AI, marking a significant step forward in generative AI technology.
Developed by Google’s DeepMind department, Gemini 2.5 Flash Image is designed to address one of the most persistent challenges in AI image generation: maintaining consistency across characters and scenes. Google highlighted this focus, stating, ‘Our latest update is designed to make photos of your friends, family, and even your pets look consistently like themselves, whether you’re trying out a 60’s beehive haircut or putting a tutu on your chihuahua.’ This means users can now expect more coherent and realistic alterations to existing images, ensuring that subjects retain their likeness even through various stylistic changes or scenarios.
The new model builds upon Google’s existing AI image generation features in the Gemini app, which were introduced earlier this year. Its enhanced capabilities include multi-turn editing, allowing users to provide an initial image and a prompt, then iteratively refine the output with subsequent instructions. For instance, a user could provide an image of an empty room, instruct the AI to paint it a different color, and then further ask it to add furniture like a bookshelf, a couch, and a rug, all while maintaining the integrity and consistency of the original scene. The AI can also combine elements from different images or adapt design styles, such as applying the texture and color of flower petals to a pair of boots.
Early previews of the ‘Nano-Banana’ model had already impressed testers with its ability to create images that closely resemble real camera captures, leading to its top-rated status in LMArena rankings. Google has acknowledged the enthusiastic reception, teasing that ‘People have been going bananas over it already in early previews.’
Recognizing the potential for misuse with such powerful image editing tools, Google has integrated robust safety features into Gemini 2.5 Flash Image. All images generated or edited within the Gemini app will feature a visible ‘AI’ watermark. Additionally, an invisible SynthID digital watermark will be embedded in all outputs, serving as a clear indicator that the images were created using artificial intelligence. These measures aim to prevent malicious activities and ensure transparency regarding the origin of the images.
Also Read:
- Adobe Integrates Google’s Gemini 2.5 Flash Image Model to Enhance Generative AI Capabilities in Firefly and Express
- Oracle and Google Cloud Forge Alliance to Bring Gemini AI Models to Enterprise Customers
The launch of Gemini 2.5 Flash Image underscores Google’s continued commitment to advancing AI capabilities, making sophisticated image generation and editing more accessible and reliable for a wide range of users and applications.
