Google has announced the expansion of its Gemini 2.0 AI series with several new models aimed at improving user experience.

https://twitter.com/sundarpichai/status/1887169871697350775

On February 5, 2025, the tech giant introduced a range of innovative features for both mobile and web platforms, catering to users from free to paid subscriptions. One of the standout offerings is the Gemini 2.0 Flash Thinking model, which is now accessible to a broader audience, including an agentic version that can engage with popular Google applications like YouTube, Search, and Maps.

Gemini 2.0 Flash Thinking Model: An AI Model Centered on Reasoning

Initially launched in December 2024, the Gemini 2.0 Flash Thinking model is now available to all users of Gemini’s mobile and web applications. This AI model is tailored for advanced reasoning tasks, providing an experience similar to models like DeepSeek-R1 and OpenAI’s o1. Users can easily access this model through the new “model selector” feature in the interface. While the rollout is anticipated to be global, details regarding availability for free users and any usage restrictions remain uncertain.

Agentic Version Gemini 2.0 Flash Thinking Model: Ensures instinctive decision-making for App Integration

Alongside the standard Flash Thinking model, Google is introducing an agentic version of Gemini 2.0 Flash Thinking. This variant can interact with various Google applications, including YouTube, Google Search, and Google Maps. Users will have the ability to instruct Gemini to perform tasks directly within these apps, enhancing productivity and streamlining app management.

https://twitter.com/GoogleDeepMind/status/1887172464863506547

Introducing Gemini 2.0 Pro: The Ultimate AI for Power Users

For those using Gemini Advanced, Google is rolling out an experimental version of Gemini 2.0 Pro, which stands as the most sophisticated AI model in the series. This model is equipped to handle complex coding and mathematical challenges, featuring a context window of 2 million tokens and integrating tools like Google Search and code execution.