Generative Image Creation: Google's New AI Model Nano Banana 2 Now Taps into Google Search
[16:07 Fri,27.February 2026 by Rudi Schmidts]
Google has released a new AI image generation model called "Nano Banana 2." It is intended to combine the speed of the predecessor model with the advanced capabilities of the Pro version. However, the most significant innovation is access to current internet data.
Google Nano Banana 2
For this purpose, Nano Banana 2 can now access not only the Gemini knowledge base but also real-time information and images from Google Search—which should result in more precise subjects and a deeper contextual understanding. This is also expected to lead to improvements in creating infographics, converting notes into diagrams, and generating data visualizations.
The rendering of text in images is also said to be even clearer and legible now, and it can even be translated directly within the image and adapted to regional contexts. Furthermore, there will be more control options for users because the new model now understands finer nuances in text instructions (prompts) and can implement them more precisely. Additionally, up to five different characters can now be consistently depicted in one image, while detail fidelity is maintained for up to 14 individual objects. This facilitates the creation of storyboards and other narrative formats.
Google Nano Banana 2 keeps characters and objects consistent
According to Google, the results also benefit from improved lighting, richer textures, and overall sharper details. Common image formats and resolutions from 512 pixels up to 4K are supported.
Nano Banana 2 is available immediately in numerous Google services. In the Gemini Assistant, it replaces the Pro version and is available in the Fast, Thinking, and Pro models. Also noteworthy: Paying subscribers of Google AI Pro and Ultra can still use the predecessor model, Nano Banana Pro, if they wish.
Furthermore, the new image AI is accessible via the Gemini API, as a preview version in AI Studio, and through the Google Cloud platforms Vertex AI and the Gemini CLI. It is also now the standard model in Google's AI video tool "Flow" and available to subscribers at no extra cost. Image generation has even been integrated into Google Ads.
To maintain traceability of AI-generated images, Google relies on its proprietary SynthID technology for Nano Banana 2. This is now combined with the industry-recognized C2PA (Content Credentials) standard to expand labeling. The Gemini Assistant is also expected to receive C2PA support shortly.
;)
;)