Google has introduced Gemini 2.0, which boasts a speed that is double that of its predecessor. What are the new features?

Google has unveiled Gemini 2.0, the latest version of its artificial intelligence technology. This new model features significant enhancements in performance, versatility, and usability, designed to tackle a wide range of tasks. It aims to revolutionize AI-driven interactions. CEO Sundar Pichai encapsulates the capabilities of Gemini 2.0 by stating, "If Gemini 1.0 was about organising and understanding information, Gemini 2.0 is about making it much more useful." According to the blog post, Google's Gemini 2.0 introduces a new category of agentic experiences, highlighting advancements in areas such as multimodal reasoning, long context comprehension, complex instruction execution and planning, compositional function-calling, native tool utilization, and reduced latency.

Gemini 2.0: Key Innovations

Gemini 2.0 boasts double the speed of its predecessor and incorporates sophisticated features in multimodal processing, as asserted by Google. It is capable of interpreting and generating outputs across various data formats, including text, audio, video, and images. A notable attribute is its capacity to manage a context window of up to 1 million tokens, allowing it to process and retain extensive information during prolonged conversations or projects.

In contrast to its predecessor, Gemini 1.0, which primarily focused on organizing and interpreting information, Gemini 2.0 introduces the concept of agentic AI.

Agentic AI refers to systems that can take initiative, make decisions, and perform tasks on behalf of users, all while being guided by human input and supervision. For instance, an agentic AI could autonomously arrange hotel bookings, suggest activities, secure dinner reservations, and develop a personalized itinerary based on the user's previous preferences.

The model's reduced latency enables it to achieve conversational speeds comparable to those of humans, making it particularly suitable for real-time interactions. Its seamless integration with tools such as Google Search, Lens, and Maps empowers Gemini 2.0 to efficiently manage complex inquiries and assist with practical tasks.

Availability of Google 2.0  

Beginning Wednesday, users of Gemini worldwide can access a chat-enabled version of 2.0 Flash by selecting it from the model menu on both desktop and mobile browsers. This version will also be made available shortly on the Gemini mobile application.

Demis Hassabis, CEO of Google DeepMind, an AI research laboratory dedicated to developing systems capable of reasoning, learning, and addressing complex challenges, announced that Gemini 2.0 will be incorporated into additional Google products next year.

“Our objective is to ensure that our models are delivered to users safely and promptly,” he stated.

Looking forward, the company intends to integrate Gemini 2.0 into Google AI Overviews to tackle more complex topics and multi-step inquiries, including advanced mathematical problems and coding tasks. This integration has commenced limited testing this week and will be expanded more broadly in early next year.

Gemini 2.0 Flash is currently available as an experimental model for developers through the Gemini API in Google AI Studio and Vertex AI, offering multimodal input and text output to all developers, while text-to-speech and native image generation features are accessible to early-access partners. General availability is expected in January, along with additional model sizes.