MeshLRM is a novel large reconstruction model (LRM) that can reconstruct high-quality 3D meshes from just four input images in less than one second. Unlike previous LRM approaches that focus on NeRF-based reconstruction, MeshLRM incorporates differentiable mesh extraction and rendering within the LRM framework. This allows for end-to-end mesh reconstruction by fine-tuning a pre-trained NeRF LRM with mesh rendering. The model's architecture is simplified to improve efficiency, and it uses a novel ray opacity loss to stabilize training and prevent floaters in the reconstructions. MeshLRM achieves state-of-the-art mesh reconstruction from sparse-view inputs and can be applied to downstream applications such as text-to-3D and single-image-to-3D generation. The key contributions of MeshLRM include a novel LRM-based framework for end-to-end few-shot mesh reconstruction, a novel ray opacity loss for stabilizing DiffMC-based training, and an efficient LRM architecture and training strategies for fast and high-quality reconstruction.MeshLRM is a novel large reconstruction model (LRM) that can reconstruct high-quality 3D meshes from just four input images in less than one second. Unlike previous LRM approaches that focus on NeRF-based reconstruction, MeshLRM incorporates differentiable mesh extraction and rendering within the LRM framework. This allows for end-to-end mesh reconstruction by fine-tuning a pre-trained NeRF LRM with mesh rendering. The model's architecture is simplified to improve efficiency, and it uses a novel ray opacity loss to stabilize training and prevent floaters in the reconstructions. MeshLRM achieves state-of-the-art mesh reconstruction from sparse-view inputs and can be applied to downstream applications such as text-to-3D and single-image-to-3D generation. The key contributions of MeshLRM include a novel LRM-based framework for end-to-end few-shot mesh reconstruction, a novel ray opacity loss for stabilizing DiffMC-based training, and an efficient LRM architecture and training strategies for fast and high-quality reconstruction.