Autonomous driving has made steady progress over the last decade, but it is unclear how close we are to truly deploying autonomous vehicles at scale. In this workshop we will explore a crucial stepping stone on the way towards full integration of autonomous vehicles: how to verify progress in the field of robotic driving. Traditionally, datasets have been proprietary and mainly focused on the perception pipeline. While both paradigms are slowly shifting, our workshop will discuss ways to accelerate the efforts towards diverse, open datasets while highlighting recent developments in benchmarking tools, datasets, and research. The workshop will contain keynote addresses from speakers with a wide range of expertise related to autonomous vehicles, including industry practitioners, academic researchers, regulators, and insurance providers. Additionally, we will examine several existing benchmarks from the aspect of the developers and the participants. In all cases, we will cover benchmarking in three areas of the autonomous driving problem: 1) perception, 2) planning and control, and 3) system integration. Finally we will conduct panel debates with expert moderators and synthesize our conclusions into a report that will be released to the public.