Google unveils Gemini 2.0, a multimodal model focused on AI agents

Gemini 2.0 represents a milestone in the evolution of artificial intelligence models by integrating multimodal capabilities and functionalities designed for the Agentic AI era. With Gemini 2.0 Flash, the first release of this family, it surpasses its predecessors in speed and key benchmarks.

Key features include multimodal input and output, such as real-time image generation, multilingual text, and audio, as well as native tools like Google Search and code execution. These innovations aim to enhance the development of dynamic and interactive applications.

The model is now available to developers via the Gemini API on Google AI Studio and Vertex AI, with early access to advanced features like text-to-speech and image generation. General availability is expected in January, alongside additional model sizes.

Gemini 2.0 is not just aimed at developers but also at users of the Gemini app, who can experience an optimized assistant that combines contextual understanding, advanced planning, and native tool usage, paving the way for pioneering agentic experiences.

This launch also introduces experimental projects like Project Astra and Jules, exploring the future of AI agents in complex and personalized tasks. With a focus on responsibility and safety, Gemini 2.0 reinforces its commitment to building trustworthy technologies, prioritizing privacy and mitigating potential risks.

With this release, Gemini 2.0 opens a new chapter in artificial intelligence, moving closer to the safe and responsible development of AGI (Artificial General Intelligence).