More on News
Google has finally launched Gemini AI after a few days of delay from the original launch date. Google’s much-awaited Gemini AI is said to be the largest and most advanced large language model. Dubbed Gemini 1.0, this model, as described by Google CEO Sundar Pichai and Demis Hassabis, CEO of Google’s AI unit DeepMind, is the first of “new generations of AI models.”
Gemini stands out as a comprehensive multimodal AI model comprising fundamental models. The model has been pre-trained and fine-tuned across several modalities; it seamlessly comprehends and reasons about diverse inputs, outperforming existing multimodal models. The model is well-versed in programming and is twice as fast as its predecessor, positioning its performance above 85% of previous competition participants.
Google Gemini AI is not a single language model but has three different models- Gemini Nano, Gemini Pro, and Gemini Ultra. Each model has its unique application.
Gemini Ultra is the largest and high-performing variant, tailored for complex tasks, making it particularly suitable for deployment in data centers and enterprise applications. On the other hand, while not as large as the Ultra, Gemini Pro possesses robust capabilities that can be scaled across a diverse array of tasks and will work as the powerhouse for several Google AI services. Then, Gemini Nano is a more compact model that can operate natively and offline on Android devices. It is highly efficient and is ideal for specific tasks and mobile devices.
Bard, now upgraded to Gemini Pro and powered by Google’s LaMDA model, brings users advanced reasoning, planning, and comprehension capabilities. Starting December 6, individuals can access the latest and upgraded Bard in the same 170 countries as the previous version. An even more advanced iteration, named “Bard Advanced,” is currently in development and is scheduled for release next year. This advanced version will be driven by Google’s largest and most advanced model, Gemini Ultra.
Developers and enterprise users can experiment with Gemini Pro using Google Generative AI Studio or Vertex AI in Google Cloud from December 13.
Meanwhile, Google’s latest smartphone, the Pixel 8 Pro, is set to get advanced AI functionalities offered by Gemini Nano. The first is an auto-summarization feature in the Recorder app, while the second is the introduction of Smart Reply. Notably, these features are designed to function offline. However, the Smart Reply currently supports only English and is compatible with WhatsApp. There are plans to integrate a Gemini-powered Bard into Google Assistant in the coming year.
For people who wish to explore Gemini Ultra, it won’t be available to the public until at least 2024, as additional red-team testing is required. As per Pichai, Gemini will eventually find its way into Google Search, Chrome, and various other applications as part of ongoing developments.