Mineral grains serve as the basic units of rocks. Understanding rock materials at the grain-scale level allows us to unveil the underlying fundamental mechanisms, thereby facilitating the development and innovation of engineering solutions. This is especially significant amidst the current energy and sustainability transition, where many geomechanics applications require a shift towards more scientific, environmentally friendly, and sustainable approaches. The Discrete Element Method (DEM) is a numerical tool that has been widely used to understand the micro-mechanisms associated with rock deformation and cracking. Here, through reviewing the past 50 years of research, we present a comprehensive state-of-the-art review of mechanical and coupled hydro-mechanical (H-M) DEM models tailored to elucidate the grain-scale behaviour of crystalline rocks. We first examine the logic, principles, and capabilities of existing DEM models and conceptual grain-based models (GBMs) and identify crucial aspects of the grain-scale behaviour of crystalline rocks. We also assess the existing coupled hydro-mechanical models and their adaptability, and then introduce a capable Hydro-GBM model specifically developed for grain-scale simulation of crystalline rocks. Finally, we discuss the selection of DEM micro parameters, which is an important and longstanding challenge for DEM, and accordingly, we provide some strategies that could alleviate the challenge. It is found that the widely used DEM parallel bonded contact model would lead to significant deficiencies in capturing the real grain-scale cracking behaviours of crystalline rocks. The previous GBM models also suffer from limitations in reproducing the grain-scale cracks and cracking modes (tension and shear cracks). Fluid-driven grain-scale cracking is at an early stage and the Hydro-GBM model appears to be an appealing tool. This review also emphases that reconsiderations of past micro-parameter selections of DEM models are necessary. In the review, we also argue that modelling of rock grain-scale H-M behaviour should be based on appropriate considerations of: (a) the solid phase (mechanical), (b) the fluid flow (hydro), and (c) micro parameters for both the mechanical and hydro aspects.