Gemini: Google Elevates Its AI Game With A New Human-Like Language Model

Google has unveiled its latest large language model called Gemini. The model is designed to behave in human-like ways and is touted as outperforming other models in tasks such as understanding, summarising, reasoning, coding, and planning.

In a demonstration, Google illustrated how Gemini possesses the ability to perceive the environment like a human eye, analyse information in real-time, and provide recommendations for the subsequent steps.

Gemini extends beyond a singular AI model. It comprises a lighter version known as Gemini Nano, intended for native offline operation on Google Android devices.

Additionally, there exists a more robust iteration named Gemini Pro, set to drive numerous Google AI services and serve as the foundation for Bard.

Lastly, there is an advanced model referred to as Gemini Ultra, representing the most potent LLM developed by Google to date, primarily tailored for data centres and enterprise applications.

While the Nano and the Pro version are already available, the Ultra version will be released early next year.