SantageAI Glossary › Small Language Model (SLM)
AI Glossary

What is a Small Language Model (SLM)?

A small language model (SLM) is a language model with far fewer parameters than a frontier model, designed to run cheaply and quickly, often on local or edge devices.

What is the core idea behind small language models?

A small language model trades peak capability for speed, low cost, and the ability to run almost anywhere.

How do small language models differ from related concepts?

ConceptDifference
SLM vs LLMLarge language models maximise capability at high cost. Small language models optimise for speed, cost, and local deployment.
SLM vs Fine-Tuned ModelAn SLM is small by design. A fine-tuned model can be any size adapted to a task.
SLM vs Quantized ModelQuantization shrinks a model's memory use. An SLM is inherently smaller in parameters.

How do small language models work?

When should you use a small language model?

Why are small language models important?

Small language models are a major 2026 trend, because most tasks do not need a frontier model. They make AI cheaper, faster, private, and deployable on everyday hardware, complementing large models rather than replacing them.

How are small language models used in practice?

SLMs are used in mobile apps, edge devices, and routing systems that send only hard queries to larger models. Their limitation is that they cannot match frontier models on the most complex reasoning and knowledge tasks.

Frequently Asked Questions

What is the difference between an SLM and an LLM?
Size and purpose. Large language models maximise capability at higher cost, while small language models are built to run cheaply and quickly, often on local hardware, for tasks that do not need the frontier.
Can small language models run on a phone?
Yes. A key advantage of small language models is that they can run on phones, laptops, and edge devices without a cloud connection.