# Inference

> Inference is the process of running a trained model to produce an output — for example generating a response or a prediction from an input. It's distinct from training, and its cost and latency are key production concerns.

*Source: https://www.lazlosoftwaresolution.com/glossary/inference*

Optimising inference (batching, caching, quantisation, the right hardware) is central to making AI features fast and affordable at scale.