Industry analysis reports and market research publications

Google DeepMind Introduces Gemini 2.0: Enhanced AI with Agentic Capabilities

Dec 12, 2024, 02:00 PM

Google has unveiled Gemini 2.0, its most advanced AI model to date, which introduces significant enhancements in speed, capabilities, and multimodal outputs. Gemini 2.0 Flash, an experimental version of the model developed by Google DeepMind, is now available to developers and trusted testers, with general availability planned for January. This model supports multimodal inputs like images, video, and audio, and can generate text, images, and audio outputs. It also integrates with tools such as Google Search, code execution, and third-party functions. Key features include improved latency, better dialogue capabilities, and the ability to conduct deep research, providing comprehensive reports on complex topics. Google is also exploring agentic AI experiences with prototypes like Project Astra, Project Mariner, and Jules, aiming to create AI agents that can perform tasks on behalf of users. These advancements are part of Google's broader strategy to integrate AI more deeply into its products, including Search, Gemini app, and other services, with a focus on safety and responsibility in AI development. Additionally, the Multimodal Live API and streaming capabilities are being introduced to enhance real-time interactions.