What is a small language model (SLM)?

An SLM, or small language model, is a language model built with far fewer parameters than a frontier LLM — think millions up to a few billion, instead of hundreds of billions or more. That size difference is the whole point: a smaller model can run fast and cheap, often directly on a phone, laptop, or a single in-house server, instead of needing a warehouse of GPUs in the cloud.

Why smaller is a feature, not just a limitation

Running locally means your data never has to leave the device, which is a real privacy win for anything sensitive. It also tends to be cheaper per query at scale, faster to respond since there's no round trip to a data center, and it keeps working even offline — on a plane, on a factory floor, or anywhere without reliable internet.

The honest tradeoff: SLMs generally know less and reason less well than the biggest frontier models. Ask one something outside its training, or something that needs a long chain of multi-step logic, and it's more likely to stumble.

That's why SLMs work best pointed at a narrow, well-defined job — say, answering questions about one company's product line, or classifying support tickets — rather than treated as an everything-assistant. Use the right-sized tool: a frontier LLM for open-ended, complex work, and an SLM for focused, repeatable tasks where speed, cost, and privacy matter more than raw brainpower.