Gemini is an artificial intelligence model released by Google DeepMind (an AI lab under Alphabet, Google's parent company) on December 6, 2023. It is capable of recognizing five types of information: text, images, audio, video, and code. It can also understand and generate high-quality code in mainstream programming languages such as Python, Java, and C++, and comes with a comprehensive security assessment. The first version, Gemini 1.0, includes three different scales of models: Gemini Ultra for "highly complex tasks," Gemini Nano for multiple tasks, and Gemini Pro for "specific tasks on end-user devices."