On December 1st, Google announced the integration of its latest large model Gemini3 into search "AI mode," initially covering nearly 120 countries and regions, and is now available to AI Pro and Ultra subscription users. On the same day, the配套 new generation generative image model Nano Banana Pro was also launched, supporting 2K/4K resolution, accurate text rendering, and professional-level camera angles, depth of field, and lighting control, with prices set at $0.139 for 1080p and $0.24 for 4K.

Gemini3 adopts a native multimodal architecture, integrating text, images, audio, and video clips within a single Transformer, making the search interface no longer return only blue links, but instead immediately generate structured cards, timelines, and interactive tools. Google stated that the English version will be released first, with multilingual support rolling out in the coming weeks.

Nano Banana Pro is aimed at creative scenarios, offering an all-in-one experience of web search and image generation, such as "finding a recipe and generating learning cards," which can directly output mixed text and image materials. This model has been integrated into the Gemini app, NotebookLM, and developer APIs, and is expected to be available to more free users next month.