Particles in flows are ubiquitous in natural and industrial processes, ranging from asbestos fibres, pollen, volcanic ash and aerosols to aquatic microorganisms, drug carriers and micro/nano robots. Their behaviour varies widely, yet accurately describing particle motion in fluids remains a major scientific and engineering challenge, especially for micron-sized and smaller particles, where experiments and fully resolved simulations are costly and often impractical. For large particle populations, Lagrangian point-particle tracking combined with state-of-the-art CFD is therefore the method of choice. However, existing approaches commonly rely on idealisations such as spherical or ellipsoidal shapes, homogeneous mass distributions, rigid material properties and isolated motion, neglecting interactions with other particles and nearby boundaries. These assumptions often fail to capture realistic particle dynamics. Therefore, we present efficient extensions of state-of-the-art Lagrangian point-particle tracking for dilute dispersed suspensions of non-ideal particles. We introduce a novel model for inhomogeneous particles, representing inhomogeneity by a mass inclusion embedded in the particle matrix. Non-spherical particles are described using the superellipsoidal surface equation and a novel surrogate model for their translational and rotational dynamics. A superellipsoid contact model based on the Lagrange multiplier technique and Newton–Raphson method is also proposed, including a tangential restitution model for arbitrary shapes that accounts for friction during collisions. Finally, an efficient model for soft, initially spherical or ellipsoidal particles is formulated within the Lagrangian point-particle framework. Based on pseudo-rigid-body theory with affine deformations, it avoids surface discretisation. All models were implemented in OpenFOAM and validated against literature benchmarks. The resulting methodology broadens Lagrangian point-particle simulations, enabling more realistic predictions of particle behaviour in arbitrary macroscopic flow fields. These advances support improved understanding and optimisation of particle-laden systems in environmental science, biomedicine and industrial processing.