Google has unveiled Gemini, its “most advanced AI model” that will form the basis of the company’s products such as Bard and Pixel 8 Pro
Google has unveiled Gemini, its “most advanced and versatile artificial intelligence model to date,” which could be an important milestone in the development of AI.
Gemini, developed by Google DeepMind, is a multimodal model, meaning it can process and understand different types of information, including text, code, audio, images, and video. It is designed to be flexible and efficient, able to run on a variety of devices, from data centers to mobile phones. Gemini comes in three versions: Ultra, Pro and Nano, each optimized for different levels of complexity and tasks.
Here are the main features of Gemini that Google points out:
- High performance: The company notes that Gemini Ultra has shown exceptional results, outperforming human experts in the MMLU test and achieving top scores in various multimodal tasks. Its abilities allow it to process complex information from various industries.
- Next-generation capabilities: Unlike previous models that combined separate components for different modalities, Gemini is multimodal from the start. This allows it to more effectively understand and reason about a variety of inputs.
- Scalability and efficiency: Google says that Gemini is the company’s most scalable and reliable model for training and maintenance. It runs significantly faster on specially designed tensor processing units (TPUs), which are the centerpiece of Google’s AI products and services.
- Responsibility and safety: Gemini has undergone a comprehensive safety assessment, including bias and toxicity evaluations. Google says it is committed to responsible AI development, considering safety at every stage and working with external experts to eliminate potential risks.
- Availability: Gemini 1.0 is rolling out across a variety of Google products and platforms. Gemini Pro is already improving the Google Bard experience, and Gemini Nano is bringing new features to the Pixel 8 Pro. Developers and enterprise customers will soon have access to Gemini Pro through Google AI Studio and Google.
For the Bard chatbot, Gemini is being launched in two stages: starting today, Bard will use a specially customized version of Gemini Pro in English for more advanced reasoning, planning, understanding, and more. And early next year, the company promises to introduce Bard Advanced, which will provide users with access to the most advanced models and features, starting with Gemini Ultra. Apparently, it will be a paid service similar to ChatGPT Plus.