Quiet Nodes
← Wiki Quiet Nodes

Vocabulary

SLM

Small language model: a lighter model that can run locally or cheaply for focused jobs, instead of waking a giant every time.

Also called: small language model · small models

SLMs are models small enough to run on a laptop GPU, a homelab box, or a cheap cloud instance. They will not match the strongest frontier systems on every benchmark. They often win on privacy, latency, cost, and “good enough for this chore.”

The Death Star toast problem is using a planet-scale model for a tiny task. An SLM is the toaster.

Cups that use this term