Google has launched its AI model - Gemini

Today, Google has launched its AI model called Gemini.

Gemini, which Google DeepMind developed, can understand text, audio, video, images, and code. This model has been designed for easy use across various devices and platforms, from data centers to mobile devices.

Gemini has three versions of functionality: Nano (for compact devices), Ultra (the most potent version for complex tasks), and Pro (the optimal and middle option between Nano and Ultra). The Nano version will appear in the Pixel 8 Pro, and the Ultra version will be available only next year.

Before the release, Google tested Gemini and noted that the new AI model outperformed GPT-4 in 6 out of 8 tests. In addition, Google has specifically combined Gemini Pro with the Bard AI assistant so that the former model could improve understanding, reasoning, coding, planning, etc.

The English version of Gemini has been available from Bard since December 6. The company plans to introduce the Bard Advanced version and Gemini Ultra next year. For now, Gemini supports textual tooltips but will eventually support other modalities. Support for languages other than English is also planned.