Google introduces Gemini, its latest language model, available in three sizes tailored to specific purposes. This versatile AI model is engineered to comprehend and navigate diverse forms of information.
Last night, Google unveiled its latest language model (LLM) named Gemini, a versatile AI designed to comprehend and integrate various forms of information, such as text, code, audio, images, and video. Claimed to be Google’s most capable and general-purpose AI model to date, Gemini is not just a single model but a family of AI models, including three sizes: Gemini Ultra, Gemini Pro, and Gemini Nano.
Gemini Nano, optimized for mobile devices like the Pixel 8, is a lightweight AI model enabling seamless on-device tasks, even offline. It excels in suggesting replies and summarizing text, providing powerful AI capabilities without relying on external servers.
Gemini Pro, the advanced variant, powers Google AI services and serves as the foundation for Bard. Running on Google’s data centers, it ensures fast response times and a profound understanding of complex queries.
Gemini Ultra, the most powerful LLM, excels at tackling complex tasks, ideal for data centers and enterprise applications. Despite not being widely available yet, it surpasses other LLMs on standard tests and is expected to be released after completing its testing phase.
Gemini, starting with Gemini Nano, is now available on Pixel 8 Pro, enhancing features like Summarize in the Recorder app and Smart Reply on Gboard. The expansion includes integration into various Google products and services, such as Search, Ads, Chrome, and Duet AI.
Developers and enterprise users can access Gemini Pro through the Gemini API in Google AI Studio or Google Cloud Vertex AI starting December 13th. Android developers can use Gemini Nano via AICore on Pixel 8 Pro devices with Android 14. Gemini Ultra is undergoing trust and safety evaluations, with early access granted to select customers, developers, partners, and safety experts. It will be available to a broader audience early next year.
Bard will receive a customized version of Gemini Pro, enhancing its reasoning, planning, and understanding capabilities. Google plans to introduce Bard Advanced, offering early access to advanced models and capabilities, beginning with Gemini Ultra in the coming year.
