Google has chosen a global stage to present Nano Banana Pro, the new version of its image generation and editing model based on Gemini 3 Pro . After the hype caused in September by the first release, which went viral thanks to its hyper-realistic 3D figurine-style miniatures, the company is relaunching a more mature and capable system, designed to offer much broader creative control. But the most surprising detail is another: anyone can try it for free , simply by opening the Gemini app and selecting “Create image” in “Thinking” mode.
The leap in quality is primarily in the ability to handle complex content. Google emphasizes how Nano Banana Pro can combine up to 14 images and integrate up to five people into the same visual result, maintaining proportions, lighting, and stylistic consistency. This feature is designed for users who want to create complex compositions, from mood boards to narrative collages, without having to resort to professional software.
Another element that sets this version apart from the past is text rendering. The model can write directly on images —from short slogans to entire paragraphs—with much more legible results than the average graphics generator. Google hopes this feature will pave the way for the rapid creation of posters, invitations, and multilingual graphics, leveraging the advanced global knowledge built into Gemini 3 Pro. Currently, however, the ability to generate text within images in Search AI Mode is limited to US users.
On the post-production front, the Nano Banana Pro introduces features typical of editing programs . It’s possible to select specific portions of the image and edit them locally, correcting perspective, colors, or lighting, even simulating day-to-night transitions or adding effects like photographic bokeh. The model also supports resolutions up to 4K, making it interesting for those producing content for the web or print.
Google has also included C2PA metadata in image generation, a standardized system that allows for the automatic identification of content created or modified using AI. This move promotes transparency, especially at a time when deepfakes are attracting widespread attention. TikTok will also adopt the same technology to invisibly mark artificially generated content.
As for access, Google has set a free usage threshold, expandable for those with an AI Plus, Pro, or Ultra subscription . The model is also available in Search AI Mode for US users with Pro or Ultra subscriptions and, globally, via NotebookLM.
