Gemini is Google's advanced multimodal AI model capable of understanding text, images, videos, and audio, designed to perform complex tasks across various domains.
Developed by Google and Alphabet, Gemini saw significant contributions from Google DeepMind, with a collaborative effort across multiple Google teams, including Google Research.
Gemini is released in three sizes—Nano for on-device tasks on smartphones like the Google Pixel 8, Pro for data centers powering AI chatbots like Bard, and Ultra for highly complex tasks, exceeding benchmarks in language model research.
Gemini is currently available on Google products in Nano and Pro sizes, integrated into Pixel 8 and Bard. Over time, Google plans to incorporate Gemini into its various services like Search, Ads, and Chrome.
Developers and enterprise customers can access Gemini Pro through the Gemini API in Google's AI Studio and Cloud Vertex AI starting from December 13. Android developers can access Gemini Nano via AICore on an early preview basis.
Gemini stands out from other AI models, such as GPT-4, due to its native multimodal capabilities. While GPT-4 relies on plugins for multimodal tasks, Gemini seamlessly operates across text, code, audio, image, and video.
Gemini Ultra, the most capable version, is currently undergoing testing and is designed for highly complex tasks. Google aims to release it after completing the testing phase.
Google plans to gradually integrate Gemini into more services beyond its initial release on Pixel 8 and Bard, indicating a broader implementation across Google's ecosystem.