Agents3h ago
Gemini 3.5 Flash Adds Computer Use
Google DeepMind integrated computer use capabilities into Gemini 3.5 Flash, allowing the model to navigate desktops and interact with software. The system interprets screen pixels to execute clicks and keystrokes. This update moves the model from a text-based assistant to a functional agent. Developers can now automate complex, multi-step workflows across various third-party applications.