Google launches new Gemini 2.0 Flash, AI model that can 'think' in response to DeepSeek & OpenAI

Google has officially launched its next-generation AI reasoning model, Gemini 2.0 Flash Thinking, aimed at enhancing how AI understands and responds to user prompts. This upgrade allows the AI to break down complex queries into manageable steps, making its reasoning process more transparent to users. By showcasing its thought process, assumptions, and line of reasoning, Gemini 2.0 aims to provide more reliable and insightful responses.

This release follows OpenAI’s launch of its o3 and o3-mini models last week. Although Google arrived a bit late to the reasoning AI race, the company asserts that Gemini 2.0 is currently the world's top-performing AI model.

Gemini 2.0 and its new features

Gemini 2.0 Flash Thinking is now available to users across web and mobile platforms. Originally introduced in December 2024, the model is Google’s answer to competitors like OpenAI’s reasoning series and DeepSeek’s R1. This latest iteration enhances its problem-solving capabilities by breaking down tasks into logical steps before providing answers. While this approach may take slightly longer to process, it results in greater accuracy and consistency.

More from Tech

Google drops pledge on never using AI for weapons, surveillance

Alphabet's Google seeks to revamp default search engine deals

Alongside the standard model, Google has also rolled out an experimental variant called 2.0 Flash Thinking with Apps. This version integrates with Google services such as YouTube, Search, and Google Maps, enabling users to interact with multiple platforms seamlessly. These integrations aim to offer a more personalised and dynamic experience, positioning Gemini as a powerful AI assistant for everyday tasks.

Both versions are being deployed globally, providing users with an AI that can "think" through prompts, making it easier to tackle complex queries.

Gemini 2.0 Pro: Advanced capabilities for power users

For users who require even more robust AI capabilities, Google has introduced Gemini 2.0 Pro Experimental. This version excels at coding, handling intricate prompts, and processing world knowledge with improved reasoning skills. The model is designed to interact with tools like Google Search, even executing code on behalf of users, making it a versatile tool for professionals.

One of the standout features of Gemini 2.0 Pro is its massive 2-million-token context window, which can process around 1.5 million words in a single session. This expansion significantly enhances the model’s ability to handle lengthy or complex tasks without losing track of context, setting a new standard for AI performance.

Gemini 2.0 Pro is now available to subscribers of the Gemini Advanced service via the Gemini app, offering unprecedented power and flexibility for users who demand high-level AI assistance.

Google’s vision for the future of AI

With the launch of Gemini 2.0 and its advanced variants, Google is doubling down on its vision of AI-powered productivity and personalisation. By improving reasoning capabilities and integrating with everyday tools, Gemini aims to become an essential assistant for a wide range of users. 

As the competition between tech giants like Google, OpenAI, and DeepSeek intensifies, users can expect rapid advancements in AI technology, making intelligent, adaptive assistance more accessible than ever.