This Tuesday, Google launched its new language model Gemini 3 Pro which has found a very favorable reception, but it is not the only news that the chatbot receives this week. The company is also deploying an LLM for the creation and editing of images within the chatbot and in other applicationsupdating the informally known as Nano Banana (Gemini 2.5 Flash Image on Gemini), introduced in September and which already offered better and faster results than ChatGPT.
Now it comes Gemini 3 Pro Image (Nano Banana Pro)a next-generation model that Google says uses Gemini’s reasoning capabilities and real-world knowledge to produce more accurate images than ever before. Furthermore, you can rely on the Search from Google to generate images based on real-time information such as weather, sporting events and other up-to-date data.
Google has also tried to improve text rendering within images. According to the company, Nano Banana Pro can produce more detailed text on mockups or posters, with a greater variety of textures, fonts and calligraphy styles. Thanks to the multilingual support of the Gemini ecosystem, it is possible generate text in multiple languages within the same image. One of the new features of Gemini 3 Pro was the expansion of the supported languages, which now include Basque, Catalan and Galician.
The consistency It is another of the model’s strong points. Google highlights that Nano Banana Pro allows combine up14 images maintaining visual coherence and resemblance of up to five people.
The model also introduces advanced creative controls. Users can select, adjust and transform any part of an image using localized editing. It also allows you to modify camera angles, change focus, apply color correction, adjust scene lighting, and more. Likewise, it is possible to generate images in a wide variety of aspect ratios and resolutions2K and 4K.
It should be noted that images created or edited with the Nano Banana Pro model will have embedded C2PA metadata. This should make it easier to detect generative AI or deepfakes in search results and social media feeds, once the industry organizes about it. TikTok announced this week that it would also use C2PA metadata to add invisible watermarks to AI-generated content.
Free deployment in Gemini, Google search engine and NotebookLM
The deployment of the model has already begun in several Google products and services. For the general public, it is coming to the Gemini app worldwide. Free users will have a limited number of daily generations, while Google AI Plus, Pro and Ultra subscribers will enjoy higher usage quotas. Ultra mode users in flow They can also use it.
In NotebookLMNano Banana Pro is available globally to subscribers. The model is also enabled in the AI Mode of Google Search in the US. Coming soon to Google Ads, Google Slides, Vids, Gemini API, Google AI Studio and to the new development environment Google Antigravity IDE.
How to use Nano Banana Pro on Gemini
After the addition of Gemini 3 Pro to the chatbot, users have two language model options to choose from. It remains Gemini 2.5 Flashwhich corresponds to the option Fast in the selector.
Gemini 3 Pro, on the other hand, appears with the name Reasoning. To use Nano Banana Pro you have to select it and then, in Toolschoose Create imagesthe option shown next to the banana icon. If you use the same option Create images in Fastyou will be using the text-to-image model above, Nano Banana.