Google's latest AI innovation, Gemini, emerges as a groundbreaking multimodal model. Gemini, which can adeptly process text, images, and videos, is being gradually integrated into Google's ecosystem, starting with the chatbot Bard. Available in over 170 regions, it will also be accessible to developers via Google Cloud's API.
Prospects include incorporating Gemini into Google's search engine, advertising platforms, and Chrome browser, with the most advanced version expected in 2024 after thorough safety evaluations. Gemini distinguishes itself by its training on a diverse set of data including images, video, and audio, offering three variants: Ultra, Nano, and Pro, each designed for different levels of capability and efficiency.
Google is enhancing its Bard chatbot with Gemini Pro, aiming to boost its reasoning and planning capabilities. A specialized variant of Gemini Pro is also being integrated into AlphaCode, a coding tool from Google DeepMind. The most advanced version, Gemini Ultra, will be incorporated into Bard and offered through a cloud API in 2024.