Deep research, flash thinking and more: 5 things to know about Gemini 2.0 | Mint

Google CEO Sundar Pichai has recently announced the launch of Gemini 2.0, the latest and most advanced AI model developed by Google and its parent company, Alphabet. The model promises to redefine how information is accessed, processed, and utilised across multiple platforms. Here is a look at five key features of Gemini 2.0:

1. Enhanced Multimodality

Gemini 2.0 introduces native image and audio outputs alongside its existing capabilities to process text, video, images, audio, and code. This makes it a natively multimodal model, enabling seamless communication and interaction across multiple formats.

2. Deep Research Feature

A standout feature is Deep Research, a tool designed to act as a virtual research assistant. It utilises advanced reasoning and long-context understanding to explore complex topics and compile detailed reports for users. This feature is now available in Gemini Advanced.

3. Flash Thinking Mode

The new experimental Flash Thinking Mode is built to simulate the “thinking process” of the model during response generation. This enhances the model’s reasoning capabilities, particularly useful for tackling advanced topics like mathematical equations or step-by-step problem-solving. Developers can access this feature through the Gemini API or Google AI Studio.

4. Integration with AI-Powered Search

Google’s Search platform has already transformed with AI Overviews, and Gemini 2.0 brings further enhancements. The model’s advanced reasoning will soon handle multimodal queries, complex topics, coding questions, and even advanced mathematics. Testing has begun, with broader rollouts planned for early next year.

5. Powered by Google’s TPUs

Gemini 2.0 is underpinned by Google’s sixth-generation Tensor Processing Units (TPUs), named Trillium. These TPUs, now generally available for customers, powered all training and inference for the model, showcasing Google’s commitment to full-stack AI innovation.

Looking Ahead

Gemini 2.0 builds on the success of its predecessor by not just organising and understanding information but also making it significantly more actionable. Pichai expressed excitement about how these advancements will shape the future of AI, especially as the model integrates into Google’s ecosystem, including its seven products used by over two billion people worldwide.