Google Unveils Gemini 2.0: A Leap Towards Smarter AI Agents

Google has introduced Gemini 2.0, the latest version of its flagship AI model, ushering in a new era of innovation in virtual agents. Announced on December 11, this upgrade builds on the Gemini family’s legacy, aiming to advance the capabilities of multimodal AI and establish a stronger foothold in the competitive AI technology landscape.

Gemini 2.0, launched a year after the debut of the original Gemini model, reflects Google’s deep commitment to redefining AI through its unified Google DeepMind division, created by merging DeepMind and Google Brain in April 2023. As Google CEO Sundar Pichai explained, “If Gemini 1.0 was about organising and understanding information, Gemini 2.0 is about making it much more useful.”

Key Features of Gemini 2.0

The model offers significant upgrades and is accessible to developers via the Gemini API in Google AI Studio and Vertex AI. Developers can experiment with the new Gemini 2.0 Flash model, designed to deliver fast, low-latency performance for various applications. Key features include:

• Multimodal Output Support: Combine text with natively generated images.
• Customisable Text-to-Speech (TTS): Available in multiple languages, tailored for natural communication.
• Enhanced Integration: Supports tools like Google Search, code execution, and user-defined third-party functions.
• New Multimodal Live API: Enables real-time audio and video streaming inputs, offering dynamic interactions and simultaneous tool utilization.

While multimodal input and text output are available to all developers, features like text-to-speech and native image generation are currently exclusive to early-access partners. These capabilities will roll out widely by January, along with additional model sizes.

Gemini 2.0 for Consumers

Consumers can look forward to a chat-optimized version of Gemini 2.0 through the Gemini AI chatbot, which will soon be available on desktop and mobile platforms, with mobile apps following shortly.

A premium feature called Deep Research will transform the chatbot into an advanced research assistant, capable of analyzing complex topics and creating detailed reports. This feature will be part of Gemini Advanced, the chatbot's premium tier.

Projects Powered by Gemini 2.0

Google is leveraging Gemini 2.0 in exciting new ventures, including:

• Project Astra: A universal AI assistant.
• Project Mariner: An AI extension for Chrome, designed to perform actions.
• Jules: An AI-powered coding assistant currently under development.

As AI innovation accelerates, Gemini 2.0 sets the stage for developers and consumers alike, creating new opportunities for interactivity, productivity, and creativity in the agentic era.