Google on Wednesday unveiled its most advanced artificial intelligence model till date, Gemini 2.0, saying that it brings the tech giant closer to its vision of a “universal assistant”.
Google CEO Sundar Pichai said that the ‘most capable model yet’ marks the start of the firm’s next era, the “new agentic era”.
“With new advances in multimodality — like native image and audio output — and native tool use, it will enable us to build new AI agents that bring us closer to our vision of a universal assistant,” Pichai said.
Google had introduced its first AI model, Gemini 1.0, in December 2023, which drove big advances with multimodality and context to understand multi-media information, from videos to codes, everything.
The release of Gemini 2.0 also led Google shares to skyrocket by more than four per cent on Wall Street, a day after tech giant’s stock already gained 3.5 per cent following the release of its ‘mind-boggling’ quantum chip Willow.
Silicon Valley’s latest trend, an AI ‘agent’, is a digital helper meant to sense surroundings, make decisions, and even take actions to accomplish specific goals.
The Google and Alphabet CEO also said that Gemini 2.0 Flash experimental model will be available to all Gemini users.
The tech giant also launched a new feature, ‘Deep Research’, which uses “advanced reasoning and long context capabilities to act as a research assistant, exploring complex topics and compelling reports” on the user’s behalf.
Gemini 2.0 is developed on custom hardware like Trillium, Google’s six-generation TPU (Tensor Processing Unit), which is now made generally available to all customers.
TPU Trillium 100 per cent powered the training and inference of Gemini 2.0, Google said.
“If Gemini 1.0 was about organizing and understanding information, Gemini 2.0 is about making it much more useful. I can’t wait to see what this next era brings,” Sundar Pichai said.
The tech giant said that Gemini 2.0 Flash will be made generally available in January, along with more model sizes.
Google will expand Gemini 2.0 to more of its products in early 2025.
The company also hinted at a new version of Project Astra, a smartphone digital assistant similar to Apple’s Siri, saying that it now has the ability to converse in multiple languages.
“With Gemini 2.0, Project Astra can use Google Search, Lens and Maps, making it more useful as an assistant in your everyday life,” Google said.
