Small language model (SLM)

SLMs have far fewer parameters than the largest models, but they are perfectly adequate for many tasks such as classifying, summarising or extracting – especially after targeted fine-tuning.

Advantages are low cost, low latency and the fact that data does not have to leave your device or network. For broad world knowledge and demanding reasoning, large models remain superior.