Gemini 2.0 Flash: Google opens the era of agentic AI in 2025

Google developers are working on the next version of Gemini. This new generative artificial intelligence model is intended to compete with other giants in the sector, such as OpenAI with the famous ChatGPT. According to the CEO of Google, Gemini 2.0 should open a new era in the field of AI, facilitating the daily tasks of users of the service.

Gemini 2.0 Flash will be released in January 2025

Google announced the arrival for the general public of Gemini 2.0 Flash in January 2025. More than improving its reasoning capabilities, this new model is designed for the “agentic” era and multiplies its capabilities to be used by the famous AI agents that we are hearing more and more about lately. Gemini 2.0 will thus serve as a multimodal basis to allow Google to deploy its current projects:

• Project Astra: improved understanding in different languages, Google Lens and Maps search, extended memory and significant reduction in latency in audio conversation;
• Mariner Project: autonomous navigation on the user's computer using a plugin for Google Chrome;
• Jules: personalized agents for developers in the IT sector;

Google intends to open the agentic era

In addition to the targeted and supervised development of some of its projects, Google now has a common guideline with that of OpenAI: the launch and diversification of AI agents. The California giant thus indicates that it is working on gaming agents, who would provide assistance to users stuck in a video game, for example. This functionality is reminiscent of Google's interest in video games, the firm having recently communicated around Genie 2, its new AI model capable of generate 3D and playable virtual environments.

If Google is moving in the direction of AI agents, the contours of these projects at this stage remain rather vague. Thus, only the future will tell us if this trend continues over time or if the evolution of technology causes it to take another path.