Tumour cell proliferation, metabolism and treatment response depend on the dynamic interaction of the tumour cells with other cellular components and physicochemical gradients present in the tumour microenvironment. Traditional experimental approaches used to investigate the dynamic tumour tissue face a number of limitations, such as lack of biological relevance for the tumour microenvironment and the difficulty to precisely control fluctuating internal conditions, for example in oxygen and nutrients. The arrival of advanced in vitro models represents an alternative approach for modelling the tumour microenvironment using cutting-edge technologies, such as microfabrication. Advanced model systems provide a promising platform for modelling the physiochemical conditions of the tumour microenvironment in a well-controlled manner. Amongst others, advanced in vitro models aim to recreate gradients of oxygen, nutrients and endogenous chemokines, and cell proliferation. Furthermore, the establishment of mechanical cues within such models, e.g., flow and extracellular matrix properties that influence cellular behaviour, are active research areas. These model systems aim to maintain tumour cells in an environment that resembles in vivo conditions. A prominent example of such a system is the microfluidic tumour-on-chip model, which aims to precisely control the local chemical and physical environment that surrounds the tumour cells. In addition, these models also have the potential to recapitulate environmental conditions in isolation or in combination. This enables the analysis of the dynamic interactions between different conditions and their potentially synergistic effects on tumour cells. In this review, we will discuss the various gradients present within the tumour microenvironment and the effects they exert on tumour cells. We will further highlight the challenges and limitations of traditional experimental models in modelling these gradients. We will outline recent achievements in advanced in vitro models with a particular focus on tumour-on-chip systems. We will also discuss the future of these models in cancer research and their contribution to developing more biologically relevant models for cancer research.