Purpose of review: It is now fully clear that information on the molecular underpinnings of tumors of the central nervous system (CNS) can be used for a more robust characterization of at least selected neoplasms. During a meeting organized in Haarlem, The Netherlands, in May 2014, about 30 neuropathologists discussed how exactly molecular information could be incorporated in the routine classification of CNS tumors. Recent findings: This meeting laid the groundwork for an update of the WHO CNS tumor classification that integrates histopathological and molecular findings. Furthermore, a layered diagnostic approach was proposed that not only allows for integration of relevant molecular information in the pathological diagnosis, but also retains the option for rendering a diagnosis based on histopathological analysis alone. An integrated morphological and molecular definition of CNS tumors brings new challenges as well. For example, criteria for grading within molecularly defined categories of diffuse gliomas will require modification, and some tests used in clinical practice for the detection of molecular features, may provide false positive or false negative results. Summary: The evolving paradigm shift represents a major leap forward in the diagnosis of CNS tumors that will contribute substantially to optimizing interobserver reproducibility and clinico-pathological predictions.