# Small Language Model (SLM)

> A model small enough to run on a laptop or a phone.

**Field:** Artificial Intelligence  
**Category:** LLMs  
**Also known as:** SLM, Compact Language Model

## Think of it like

Think of it like a small book that fits in your pocket. A big model is a whole shelf.

## Definition

A small language model has far fewer weights than the big ones. It most often has a few billion or less. It is small enough to run on a laptop or a phone, often with no internet. It knows less and reasons less well than a large model. But it is fast and cheap. Teams use small models for narrow jobs, like sorting messages or filling in forms.

## Example

Emma's notes app sums up each note right on her phone. A 3B model does the work, so her notes never leave the device.

## Why it matters

Small models keep data private. They also cost almost nothing per use. That makes some AI tools possible at all.

## Common mistake

Expecting the answers of a large model from a small one. It is good at a narrow job and weak at open questions.

## Three principles

- **Few billion weights.** Or fewer.
- **Runs on your device.** Often offline.
- **Fast and cheap.** Best at narrow jobs.

## Related terms

- [Large Language Model (LLM)](https://learn.benibix.com/term/ai/large-language-model)
- [Parameter Count](https://learn.benibix.com/term/ai/parameter-count)
- [Open-Weight Model](https://learn.benibix.com/term/ai/open-weight-model)
- [Phi](https://learn.benibix.com/term/ai/phi)
- [Model Selection (LLM)](https://learn.benibix.com/term/ai/model-selection)

## Taught in

- [Knowledge Distillation](https://learn.benibix.com/learn/ai-llms/knowledge-distillation): LLMs, 10/15
- [Model Selection and Testing](https://learn.benibix.com/learn/ai-working-with-ai/model-selection-and-testing): Prompt Engineering, 12/16
- [Gemma, Phi and gpt-oss](https://learn.benibix.com/learn/ai-models-providers/gemma-phi-and-gpt-oss): AI Models and Providers, 7/13
- [Predictive Text on Phones](https://learn.benibix.com/learn/ai-use-case/predictive-text-on-phones): AI Assistants and Chatbots, 12/13

![Four bars for models of 1, 8, 70 and 400 billion parameters. Under each, the memory the weights need at two bytes each: about 2, 16, 140 and 800 gigabytes.](https://learn.benibix.com/diagrams/llm-parameter-count.svg)

---

Source: https://learn.benibix.com/term/ai/small-language-model
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-18
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
