Agent
A model that works towards a goal over many steps, choosing and using tools and reacting to the results.
Also: AI agent, agentic AI
An agent runs in a loop: it plans, calls a tool, reads the result, and decides what to do next until the task is done or it gives up. Agents can browse the web, write and test code, fill in forms, or operate a computer through screenshots. The ReAct paper of 2022, which interleaved reasoning and actions, was an early template. Agents are powerful but fail in new ways, such as compounding small errors over long tasks or acting on malicious instructions hidden in a web page, so many designs ask a person to confirm important steps.