Google officially launches Gemini: its most powerful AI model to date

Google has launched Gemini, a series of large language models (LLMs) that represents the company's most ambitious and advanced AI project to date. This groundbreaking multimodal model is their most capable and versatile model yet, promising to completely transform the way people interact with technology and unlock new possibilities in various fields.

The first version, Gemini 1.0, has been optimized for three different sizes for various applications:

• Gemini Ultra: The largest and most powerful variant, suitable for complex reasoning tasks. It surpasses human-level performance in several benchmark tests.
• Gemini Pro: A model that adjusts scale and speed, suitable for commercial applications.
• Gemini Nano: A streamlined model designed to run efficiently on smartphones and other devices.

Google states that Gemini Ultra surpasses state-of-the-art results in 30 out of 32 key benchmark tests used for AI research. This includes surpassing human performance for the first time in tests of broad knowledge and problem-solving abilities.

Unlike earlier models that only dealt with text, Gemini was created from the start as a multimodal model, understanding text, images, audio, video, and other formats simultaneously. It can seamlessly reason about correlated concepts in ways that other AI systems currently cannot achieve.

Google plans to introduce Gemini's capabilities in many of its products in the coming months. This includes upgrades to search, Pixel phones, and the Bard conversational model. Meanwhile, developers and enterprise customers can access Gemini Pro starting from December 13th through Google AI Studio and Cloud Vertex AI. In early next year, Google will launch Bard Advanced, a new service that provides access to its best models, starting from Gemini Ultra.

Google emphasizes that security and regulation are top priorities in the ongoing development of Gemini, and it provides detailed explanations of extensive evaluations conducted by internal and external experts. However, the rapid advancement of artificial intelligence and commercial pressures have raised increasing concerns about potential downsides in society, and Google acknowledges its commitment to addressing these issues.

With Google now holding the title for the world's most powerful general-purpose model, the release of Gemini marks a defining moment in its AI journey. However, with OpenAI making impressive breakthroughs with its upcoming GPT-5 model, 2024 will be another exciting year in the field of AI.