Enterprises are shifting from large language models (LLMs) to smaller AI models due to high compute costs, model hallucinations, and the need for domain-specific expertise. Small models are less resource-intensive, cost-effective, and suitable for specialized industries like healthcare. Analysts predict a boost in small language model adoption, with some enterprises exploring models with 1-10 billion parameters. While offering advantages in cost and sustainability, small models may not match the versatility of LLMs, necessitating careful case matching for effectiveness.
https://www.ciodive.com/news/small-language-models-AI-LLMs/740281/