In practice
You declare the functions with name, parameters, and description; the model picks when to call them. It is the building block of every agent. Validate arguments carefully: the model sometimes invents parameters or forgets them.
Related terms
Seen in the wild
12 entries mentioning it- HighLM Studio 0.3.17 becomes an MCP host: external tools for local models
- HighOpenAI o3 and o4-mini: reasoning models learn to use tools
- MediumHugging Face smolagents: agents that write code instead of JSON
- HighOpenAI o1 arrives in the API with function calling, Structured Outputs and vision
- HighModel Context Protocol: the open standard to connect LLMs and data
- HighOpen WebUI: Tools and Functions turn the frontend into a plugin platform
- MediumCohere Command R+: an enterprise-focused model built for RAG and tool use
- HighClaude 2.1: 200K context and fewer false statements
- HighFunction calling: GPT learns to speak JSON
- MediumChatGPT Plugins: the model starts using external tools
- HighToolformer: the LLM that learns to use tools on its own
- LandmarkReAct: the framework that unites reasoning and acting in LLMs