The modelling and analysis of composites, at the microscale and macroscale levels, saw a tremendous shift due to the growth of the computational power and mathematical representation of the reinforcement. This study was conducted to assess the importance of modelling fibre tow in the composite at the macroscale level compared to the homogenisation of a microscale unit cell model using SwiftComp GUI in ABAQUS. One microscale unit cell model and three macroscale laminates (CLCCA, CLSwC, and UT) models were created. The microscale model was subjected to homogenisation. The models were then subjected to tensile and shear loading in ABAQUS. The value of young's modulus in the longitudinal and transverse direction by SwiftComp was 12.91% and 5.41%, respectively higher than that by the CCA method. In contrast, Poisson's ratio was underpredicted by 2.35%. Due to the over-prediction of elastic moduli by SwiftComp GUI module, it is not recommended to homogenize the unit cell. The difference in the young's modulus and shear modulus is less than 7% for the UT model compared with the CLCCA model. In conclusion, the CL modelling approach with numerical methods such as CCA is favourable as it provides acceptable results with relative ease of meshing, application of periodic boundary conditions, and lower computational time. On the other hand, the UT model can study the fibre–matrix behaviour by specifying the contact property between the fibre and matrix instead of assuming the bond to be perfect. Hence, it is favourable to model tow and matrix separately when reviewing the failure mechanism on any unbalanced or unsymmetrical laminated composite.