The Case for Using Small Language Models

"Until now, the AI revolution has been largely measured by size: the bigger the model, the bolder the claims. However, as we move closer to truly autonomous and pervasive AI systems, a new trend is emerging that suggests a surprising shift: smaller may actually be better. According to a recent study by NVIDIA researchers, small language models (SLMs), rather than their larger counterparts large language models (LLMs), could become the true backbone of the next generation of intelligent enterprises. The age of "bigger is better" may be giving way to "smaller is smarter.""

"Ajay Kumar is an Associate Professor of Information Systems & Business Analytics at EMLYON Business School, France. Ajay has held postdoctoral fellow positions at the Massachusetts Institute of Technology, and Harvard University. Presently, he is affiliated as a visiting fellow with the Said Business School at the University of Oxford."

"Thomas H. Davenport is the President's Distinguished Professor of Information Technology and faculty director of the Metropoulos Institute for Technology and Entrepreneurship at Babson College, a visiting scholar at the MIT Initiative on the Digital Economy, and a senior adviser to Deloitte's Chief Data and Analytics Officer Program."

AI development historically prioritized larger models, equating size with capability. Emerging evidence indicates that small language models (SLMs) may offer superior suitability for autonomous, pervasive systems within enterprises. SLMs can serve as efficient, scalable backbones for next-generation intelligent operations, potentially lowering resource demands while maintaining or improving task-specific performance. NVIDIA research highlights this shift toward pragmatism in model selection, emphasizing operational efficiency and integration across enterprise workflows. The trend suggests a strategic pivot from maximizing model size to optimizing model fit, deployment costs, and real-world autonomy for business applications.

Read at Harvard Business Review