Google's New Milestone in AI development: Gemini

By: Eva Baxter

Google has introduced Gemini, its most sophisticated AI system till now, marking a significant evolution in its AI progression. As expressed by Google's CEO, Sundar Pichai, Gemini is developed to understand and operatively integrate text, images, audio, video, and other types of information.

This development in AI models represents one of the most extensive science and engineering efforts undertaken by the firm. Gemini outperformed all previous benchmarks, scoring higher than any other model on over 30 leading AI evaluation methods according to Google's evaluation.

Notably, Gemini Ultra, the largest rendition, is the first ever recorded AI system to surpass an average human performance on a language test spanning over 50 academic subjects. Gemini also displayed new strengths in computer programming. Employing a specialized version of Gemini, Google developed AlphaCode 2, an advanced code generation system which remarkably excels in solving competitive programming problems.

Gemini will be implemented immediately in several Google products, including the Bard AI conversational tool and Pixel smartphones. Certain developers will also gain early access to Gemini Ultra this year with the plan for wide availability on the developer platforms by 2023.