This scalable language-model tool is used to build language models from large amounts of data. It supports modified absolute discounting and Kneser-Ney smoothing. The tool has been used successfully to build a seven-gram language model on 40 billion words within eight hours.