An SLM is a reduced-size language model designed to run with limited resources: on a modest server, a local machine or even an embedded device. Less general-purpose than an LLM, it is faster, cheaper and more frugal, which makes it a sound choice for targeted tasks or offline deployments.
In practice at Gensai
This is the approach behind the “Monsters of the Oceans” installation: the entire AI chain runs locally on a Mac Mini, with no Internet connection.