Sustainability & AI
Small language model (SLM)
A compact language model needing far less compute than frontier models, relevant to lower-footprint AI deployment.
Definition
A language model small enough to run with modest compute — on a workstation, device or single accelerator — trading peak capability for efficiency and deployability.
Quick reference
At a glance
- Subject
- Sustainability & AI
- Editorial status
- Editorial draft
- Definition status
- Established
- Last updated
- 21 August 2026
- Also known as
- small language model · compact language model
References
This source provides part of the technical or institutional basis for the definition.
This source supports the explanation of how the term is applied, measured or governed in practice.
Overview
What it means
SLMs pursue performance through training-data quality and design rather than sheer scale; Microsoft's Phi series popularised the approach with its 'Textbooks Are All You Need' recipe. They can run locally, keeping data in-house and avoiding hyperscale inference infrastructure.
How it is used
Use cases include on-device assistants, document triage inside secure environments, and lightweight analysis where a frontier model would be disproportionate — a pattern aligned with frugal-AI thinking.
Why it matters
SLMs are the practical counterweight to scale-at-all-costs AI: for many sustainability tasks they deliver most of the value at a fraction of the energy, cost and data-governance exposure.
Review
Help keep this definition useful and accurate.
Submitted reviews stay private until an editor decides whether to accept and attribute them.
Endorse this definition
Confirm what works
Endorse the current wording when it is accurate and useful. Suggested changes use the separate editorial form below.