Small Language Model
Definition
A small language model requires comparatively little computing power and memory for its tasks.
A small language model is a relatively resource-efficient language model; “small” is context-dependent and not defined by a uniform parameter limit.
Simply explained
It is a more economical variant of a language model; there is no fixed size at which it is considered small.
Why it matters in practice
The term helps to select, develop, evaluate AI systems in a well-founded manner and realistically assess their technical risks.
Example from business
A company operates a compact model locally for narrowly defined classification tasks.
Limits and pitfalls
- Small does not automatically mean inexpensive, safe, or correct; quality and hardware requirements must be tested on the real workload.
Origin & history
With advances in distillation, quantization, and efficient training, powerful compact models for local and specialized tasks have become widespread.
Use in mid-sized companies
- Possible application: A company operates a compact model locally for narrowly defined classification tasks.
Frequently asked questions
What does Small Language Model (SLM) mean?
A small language model is a relatively resource-efficient language model; “small” is context-dependent and not defined by a uniform parameter limit. What should be considered? Small does not automatically mean inexpensive, safe, or correct; quality and hardware requirements must be tested on the real workload.
Sources
- Phi-3 Technical Report · Microsoft Research, Phi-3 Technical Report
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: