Google Gemini 3.5 Flash: The Era of AI Agents Has Begun

Google has unveiled its new Gemini 3.5 Flash AI model. This model is considered the company's most powerful solution for programming and working with autonomous agents. The new technology goes beyond simply answering questions, possessing the ability to plan complex tasks, write code, and even build operating systems from scratch. This is reported by Techcrunch.com reports .

DeepMind Chief Technologist Koray Kavukcuoglu noted that Gemini 3.5 Flash outperforms its predecessor, the Gemini 3.1 Pro model, in almost all metrics. The model's main advantage is its speed: it operates four times faster than other advanced models, and its optimized version is 12 times faster. This speed is crucial for complex processes where multiple agents must work simultaneously.

During the presentation, Google engineers demonstrated how agents, using the Antigravity platform, independently created software components and integrated them into a single system. Antigravity 2.0 is an environment specifically designed for agents, providing a convenient space for AI to perform its tasks.

Currently, Gemini 3.5 Flash is being actively tested by financial organizations and data analysis teams. Although the model is capable of operating autonomously for several hours, it requests human intervention for critical decision-making or stages requiring authorization. Together with the upcoming Gemini 3.5 Pro model, they will form an even more efficient system.