Hugging Face has made a significant advancement in the field of small language models (SLMs) with the release of SmolLM2, a compact 1.7 billion parameter language model. SmolLM2, which sets a new benchmark for performance among models of its size, is designed for on-device use, handling texts up to 20 pages, generating concise summaries, and supporting interactive follow-up questions. In addition, Hugging Face introduced SmolTools, a suite of lightweight AI-powered tools built with LLaMA.cpp and small language models, showcasing the capabilities of these new models. SmolLM2 is available in three sizes: 135M, 360M, and 1.7B. Other companies are also accelerating their SLM efforts, with Meta open-sourcing the MobileLLM foundation model for on-device scenarios. The rise of small language models is driven by their efficiency and ability to perform tasks typically handled by larger models, making them an attractive option for businesses aiming to operationalize responsible AI.