The paper introduces the Waymo Open Dataset, a large-scale, high-quality, and diverse multimodal camera-LiDAR dataset for autonomous driving research. The dataset includes 1150 scenes, each spanning 20 seconds, with synchronized and calibrated LiDAR and camera data captured in urban and suburban environments across San Francisco, Phoenix, and Mountain View. It is 15 times more diverse than the largest existing camera+LiDAR dataset based on geographical coverage. The dataset is exhaustively annotated with 2D and 3D bounding boxes, providing strong baselines for object detection and tracking tasks. The authors also study the impact of dataset size and generalization across different geographies on 3D detection methods. The dataset and code are publicly available, and the authors plan to add more labeled and unlabeled data to enable research on additional self-driving tasks.The paper introduces the Waymo Open Dataset, a large-scale, high-quality, and diverse multimodal camera-LiDAR dataset for autonomous driving research. The dataset includes 1150 scenes, each spanning 20 seconds, with synchronized and calibrated LiDAR and camera data captured in urban and suburban environments across San Francisco, Phoenix, and Mountain View. It is 15 times more diverse than the largest existing camera+LiDAR dataset based on geographical coverage. The dataset is exhaustively annotated with 2D and 3D bounding boxes, providing strong baselines for object detection and tracking tasks. The authors also study the impact of dataset size and generalization across different geographies on 3D detection methods. The dataset and code are publicly available, and the authors plan to add more labeled and unlabeled data to enable research on additional self-driving tasks.