What is Small Language Models (SLM)?
Small language models (SLMs) are AI models trained on a smaller volume of parameters than large language models, typically optimised for a narrower set of tasks or a specific domain. They generally require less compute, are cheaper to run and easier to deploy on private infrastructure, which makes them well-suited to regulated environments where data sovereignty and cost control matter. SLMs trade some general capability for greater efficiency, speed and specialism.
Small Language Models (SLM) in financial services
Financial services firms consider SLMs where data control, cost and predictable performance on a defined task matter more than broad general knowledge — for example, running within a firm's own environment on a specific type of review.