Gemini 1.0 is currently being deployed across various products and platforms such as Bard which has started to employ an optimised iteration of Gemini Pro to enhance capabilities in advanced reasoning, planning, understanding, and more
Google unveils its general AI model dubbed Gemini. (Credit: Outreach Pete/Wikimedia Commons)
Search engine major Google has unveiled the company’s general artificial intelligence (AI) model, dubbed Gemini, in a move to make AI more helpful for people.
The new AI model, which Google calls its largest and most capable, is built to be multimodal.
It is pre-trained from the beginning on different modalities, enabling it to generalise and seamlessly comprehend. The model can operate across and integrate various types of information including text, code, audio, image, and video.
The first version of Gemini is trained at scale on Google’s AI-optimised infrastructure using in-house designed tensor processing units (TPUs) v4 and v5e.
According to Google, the new model will efficiently operate on all systems ranging from data centres to mobile devices. Gemini’s advanced capabilities will also improve the way developers and enterprise customers develop and scale with AI.
Google has optimised Gemini 1.0 for three different sizes.
Gemini Ultra is Google’s largest and most capable model for highly complex tasks while Gemini Pro can be used for scaling across a wide range of tasks. Gemini Nano is said to be the most efficient model for on-device tasks.
Google and Alphabet CEO Sundar Pichai said: “Now, we’re taking the next step on our journey with Gemini, our most capable and general model yet, with state-of-the-art performance across many leading benchmarks.
“Our first version, Gemini 1.0, is optimised for different sizes: Ultra, Pro and Nano. These are the first models of the Gemini era and the first realisation of the vision we had when we formed Google DeepMind earlier this year.
“This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company.”
Gemini is capable of extracting insights from hundreds of thousands of documents through reading, filtering, and understanding information. This is expected to help in delivering new breakthroughs at digital speeds in several fields from science to finance.
Besides, the first version of the new AI model will understand, explain and generate high-quality code in programming languages including Python, Java, C++, and Go.
Gemini 1.0 is currently being deployed across various products and platforms. Beginning today, Bard will employ an optimised iteration of Gemini Pro, enhancing capabilities in advanced reasoning, planning, understanding, and more.
In the upcoming months, Gemini is set to expand its presence in additional products and services such as Search, Ads, Chrome, and Duet AI.
Google is extending Gemini to the Pixel series. The Pixel 8 Pro stands as the first smartphone designed to run Gemini Nano, powering features like ‘Summarize’ in the Recorder app. It will be introduced in Smart Reply on Gboard, initially with WhatsApp, and additional messaging apps are expected to follow in the coming year.