Reclined seating configurations, relevant to the future of Autonomous Driving Systems is likely to challenge the current state-of-the-art restraint systems. Human body models (HBM) offer an attractive tool to support the design process, however their validity in the reclined scenario remains questionable. The goal of this study is to compare the response of selected HBMs in the frontal, reclined scenario, while utilizing a new prototype restraint system. A sled model with a generic seat, 50 deg seatback recline angle and a prototype 3-point belt system was used in this study. Four different male HBMs were compared, the Global Human Body Model Consortium (GHBMC) simplified occupant model (GHBMC-S), the GHBMC detailed model (GHBMC-D), Total Human Model for Safety SAFER (THUMS-S) model, and THUMS-v5 model. All HBMs showed good pelvis engagement, except GHBMC-D that submarined under the lap belt. Additionally, large differences were observed in pelvis and lumbar spine response between GHBMC and THUMS family models. Since no relevant PMHS data is currently available, it is impossible to evaluate the biofidelity of these models in the reclined scenarios. Evaluating the relative biofidelity of these models can only be accomplished with experimental data capturing detailed 3D skeletal kinematics and all the boundary forces necessary for model evaluation.