Language model
Any system that assigns probabilities to sequences of words, and so can predict or generate text.
A language model gives a probability to every possible next word or piece of word, given the text so far. Early language models counted how often short runs of words appeared together, the n-gram method used in speech recognition and machine translation from the 1980s. Neural language models replaced counting with learned representations, first with recurrent networks and from 2017 with transformers. The same idea, scaled up, is what powers today's chatbots. A model that is small by today's standards is still a language model, just not a large one.