AI agent

An AI agent is a system in which a model plans and carries out multi-step tasks by calling tools, such as running code, searching the web or editing files.

  • Models and architecture
  • Updated

Instead of just replying, an agent works in a loop: it decides what to do, uses a tool, reads the result and continues until the task is finished. Coding assistants that edit files and run tests are the most common example.

Agents put more strain on a model than chat does. They need reliable tool calling, a long context and the ability to keep making progress without repeating the same step. Many new open models, such as GLM-5.3 and Kimi K3.1 Pro, advertise agentic strengths; look at independent results where you can and try a shortlist on your own tasks.

Related terms

  • Reasoning model A reasoning model works through a problem step by step in a 'thinking' phase, visible or hidden, before it gives its final answer.
  • Context window The context window is the maximum amount of text, measured in tokens, that a model can take into account at one time, including your prompt and its own answer.
  • Benchmark A benchmark is a standardized test used to compare models, such as a set of maths, coding or knowledge questions; scores are useful but easy to over-read.
  • Agent memory Agent memory lets an AI agent keep facts, preferences and past experience between sessions instead of starting from a blank context.

All terms