Introducing Gemini: our largest and most capable AI model

Google, Alphabet and DeepMind CEOs introduce Gemini, the largest, most capable AI model thus far. It enables significant advancements in scientific discovery, human progress and improves lives greatly. Gemini 1.0 is optimised for different sizes for use across products and services like Search, Ads, Chrome and Duet AI. The model recognises text, images, audio and more simultaneously, excelling at explaining complex reasoning in subjects like maths and physics. Gemini is being released in Google products like Bard and Pixel 8 Pro and will soon be accessible to developers and enterprise customers for their advantage. The development and deployment of Gemini strictly adhere to robust safety policies across products and Google's AI principles.

Opens on Google Blog