A Gaussian Mixture Model is a model that assumes the data comes from a mixture of a finite number of Gaussian distributions. The goal of modeling is to estimate the parameters of such Gaussian components, namely their means and covariance matrices. That is commonly achieved using an algorithm known as Expectation-Maximization, which can be thought of as a generalized version of K-Means clustering that takes into account the assumption that samples from each cluster are distributed as a Gaussian.