Google Gemini, which is Google’s answer to OpenAI’s GPT-4, has launched officially as the most capable and the largest AI model from the search giant. Sundar Pichai (CEO of Google and Alphabet) and Demis Hassabis ( CEO and Co-Founder of Google DeepMind) announced the launch through an official blog post on The Keyword.
According to the blog, the Gemini 1.0 is optimised for three different sizes namely Ultra, Pro, and Nano. The Gemini Pro is already available in Bard and Gemini Nano is rolling out to Pixel 8 Pro in December Feature Drop. Let’s take a look at what Google Gemini is and the different sizes being offered in the AI model.
Gemini 1.0 is the Most Advanced Multimodal AI Model From Google
Google has announced that Gemini is the most advanced and largest AI model from the brand and is the realisation of the vision with which the Google DeepMind was established earlier this year. The Gemini 1.0 is the Google’s reply to ChatGPT 4. According to Demis Hassabis, CEO and Co-Founder of Google DeepMind, Gemini is designed “to understand the world the way we do and absorb any type of input and output.”
To elaborate, Google Gemini 1.0 has been designed from the start to be multimodal. It means Gemini can understand and generalise multiple data points including text, images, audio, video, and codes, all at the same time. It means that the newest AI model is designed to parse nuances better and answer complicated questions more accurately.
Google shared various benchmarking results where Gemini Ultra has been able to surpass GPT 4. But what is Gemini Ultra? Google has launched Gemini 1.0 in three optimisation sizes viz. Gemini Ultra, Gemini Pro, and Gemini Nano.
Google Gemini Available in Bard and Pixel 8 Pro
Seeing some qs on what Gemini *is* (beyond the zodiac :). Best way to understand Gemini’s underlying amazing capabilities is to see them in action, take a look ⬇️ pic.twitter.com/OiCZSsOnCc
— Sundar Pichai (@sundarpichai) December 6, 2023
Google explains that Ultra is meant for highly complex tasks, Pro is designed to scale across a wide range of tasks, and Nano is meant for on-device tasks.
Google says that the Gemini Ultra can work across languages and process complex information. It can understand, explain, and generate high-quality codes in languages such as C++, Python, and Java, among others. Gemini Ultra will be made available early next year through an all-new Bard Advanced offering. Google is currently running more tests to ensure Gemini Ultra is safe and secure for developers, businesses, etc. Google is positioning Bard Advanced as an option offering early access to its most advanced AI models and capabilities.
Speaking of Bard, it is now using what Google is calling a fine-tuned version of Gemini Pro. According to the company, Bard should now be able to offer “more advanced reasoning, planning, understanding, and more.” Users can now test text-based prompts in English in Bard with Gemini Pro starting today in 170 regions.
Google also announced the Pixel Feature Drop December 2023, which brings the Gemini Nano to the Pixel 8 Pro. In the blog post, Google mentions that the Pixel 8 Pro is the first smartphone engineered to run Gemini Nano. As per Google, Gemini Nano is powering features like Summarize in the Recorder app and Smart Reply in Gboard in WhatsApp with more messaging apps coming next year.
Google also said that Gemini will be coming to more products and services like Chrome, Search, Ads, and Duet AI in the coming months. The company will make the Gemini API available to developers and enterprise customers in Google AI Studio or Google Cloud Vertex AI starting December 13. We can expect more apps and services using Gemini 1.0 in the coming days.
