The field of 3D vision-language understanding is rapidly advancing, with a focus on improving the alignment between 3D point clouds and natural language descriptions. Researchers are exploring new methods to capture fine-grained alignments, leveraging pre-trained language models and introducing innovative modules for temporal reasoning and cross-modal fusion. These advancements are enabling more accurate and robust 3D scene understanding, object detection, and segmentation. Notable papers in this area include Capturing Fine-Grained Alignments Improves 3D Affordance Detection, which proposes a novel method for affordance detection in 3D point clouds, and GroundFlow: A Plug-in Module for Temporal Reasoning on 3D Point Cloud Sequential Grounding, which introduces a module for temporal reasoning on 3D point cloud sequential grounding. Additionally, papers like Segment Any 3D-Part in a Scene from a Sentence and SAM4D: Segment Anything in Camera and LiDAR Streams are pushing the boundaries of 3D-part scene understanding and multi-modal foundation models for promptable segmentation.