AI glossary
Local model
A local model is a language model that runs on hardware you control, a laptop, a server or your own cloud machine, rather than behind someone else's API.
You download open weights, often quantised, and serve them with an inference server that offers an API much like a hosted one, so your code barely changes.
The gains are privacy, control over versions and no charge per token. The costs are hardware, your own operations work, and usually a smaller model than the largest hosted ones.
Work out the sums before you choose: a machine that sits idle most of the day can cost more than paying per call.
Learn it hands-on
These free lessons teach it with commands to run in a lab and checked tasks. Read them whole with no account.