On Wednesday, Google presented Gemini, its most powerful language model to date. According to Google, Gemini performs better on massive multi-task language understanding (MMLU) tasks than human experts and OpenAI’s GPT-4 language model.

• With Gemini, Google has introduced its best performing language model.
• According to Google, Gemini outperforms human experts and OpenAI’s GPT-4 language model on massive multi-task language understanding tasks.
• The latest AI model Gemini is installed on the Pixel 8 Pro and the Google Bard.

Google CEO Sundar Pichai announced the launch of the latest artificial intelligence model, Gemini 1.0 – the most powerful and versatile in Google’s portfolio – on the X platform. This first model in the Gemini series is multimodal from the start. Three variants of Gemini have been developed: Ultra, Pro and Nano. The technical director added that Gemini Ultra performs better than currently available models in 30 out of 32 common academic tests. It is the first model to achieve a score of 90.0% on the MMLU test, exceeding the performance of human experts.

Gemini will appear alongside Google Bard

Google announced on Wednesday that the Gemini Pro language model is now powering the AI chatbot Bard, improving its capabilities in the areas of understanding, summarising, reasoning, coding and planning. Google also announced that Bard with Gemini Pro outperforms ChatGPT, which uses GPT-3.5 (free version), in six out of eight industry-standard tests.

Sissie Hsiao, vice president of Google, announced in a post announcing Bard’s new features that starting today, the chatbot will use, but is not limited to, a specially tuned version of Gemini Pro in English for more advanced thinking and planning. In addition, Bard Advanced will launch early next year, providing first access to the most advanced models and features, starting with Gemini Ultra.

Pixel 8 Pro equipped with Gemini AI functions

The Mountain View, California-based company also announced that the Pixel 8 Pro device’s features are now powered by Gemini AI Nano, a more efficient version of the language model. The Gemini model powers the Pixel 8 Pro’s new features, which include artificial intelligence-based chat responses and summaries of recorded conversations.