During the last ten years there has been many attempts to provide a Computer Algebra System (CAS) for Python that have important applications in code generation tools, for example. In most cases, one of the following approaches has been proposed: wrap existing CAS libraries to Python, create Python interfaces to existing CAS programs, or implement pure Python CAS from scratch. In this talk I will discuss pros and cons of these approaches as well as try to give an overview of what is the current state with CAS-s for Python. Finally, a pure Python package, sympycore, will be introduced as sufficiently efficient and robust implementation of a CAS for Python. For example, the sympycore speed is comparable with the speed of many CAS-s that are implemented using a compiled language.