Google introduced a slate of new artificial intelligence models and agents today as part of a major upgrade to its Gemini project.
Gemini, a set of large language models, powers Google’s AI efforts like its Gemini chatbot and NotebookLM. Gemini 2.0 can generate images and audio, and operate online tools.
“Over the last year, we have been investing in developing more agentic models, meaning they can understand more about the world around you, think multiple steps ahead, and take action on your behalf, with your supervision,” said Google CEO Sundar Pichai.
“With new advances in multimodality — like native image and audio output — and native tool use, [Gemini 2.0] will enable us to build new AI agents that bring us closer to our vision of a universal assistant,” Pichai said.
The Gemini 2.0 Flash model will be widely available from today, in an experimental form. Audio and image generation will roll out to the public in January.
According to Google, Gemini 2.0 Flash can operate at twice the speed of Gemini 1.5 Pro.
The company has introduced three new AI agents — Project Astra, Project Mariner, and Jules — as part of Gemini 2.0.
Project Astra describes the physical world, Project Mariner navigates the internet through the Chrome browser, and Jules assists with coding in GitHub.
Gemini 2.0 uses Google’s Trillium tensor processing unit, which was also released to the public today.
Share prices at Google’s parent company Alphabet (NASDAQ: GOOG) closed at US$196.71 after the announcement, surging from the previous day’s $186.53. Alphabet’s market cap is US$2.4 billion.