In this paper, we describe an open source Python toolkit named Uncertainty Quantification 360 (UQ360) for the uncertainty quantification of AI models. The goal of this toolkit is twofold: firstly, to provide a broad range of capabilities to streamline, and hopefully foster the common practices of quantifying, evaluating, improving, and communicating uncertainty in the AI application development lifecycle; secondly, to disseminate the latest research and educational materials for uncertainty quantification in machine learning, and encourage further exploration of its utility and connections to other pillars of trustworthy AI such as fairness and explainability. Beyond the Python package (\url{https://github.com/IBM/UQ360}), we have developed an interactive experience (\url{http://uq360.mybluemix.net}) and guidance materials as educational tools to aid researchers and developers in producing and communicating high-quality uncertainties in an effective manner.