What is a prompt?
A prompt is the input you give a language model: your question or instruction plus anything sent with it. That can include a system message, earlier turns of the conversation or pasted documents. Models read prompts as tokens, which are small chunks of text; in English, one token averages roughly three-quarters of a word. A short chat question may be a few dozen tokens, while a pasted report can run to thousands. Prompt length matters for local LLMs in two ways. Long prompts take longer to process before the first word of the answer appears. They also need extra memory to hold the context, which eats into your headroom. Hardware with stronger compute shortens the wait on long prompts. The Apple Mac mini (M5 Pro), for example, lists GPU Neural Accelerators for quicker prompt processing.