Tissue engineering enables the production of three-dimensional microtissues which mimic naturally occurring conditions in special tissues. These 3D culture systems are particularly suitable for application in regenerative medicine or experimental pharmacology and toxicology. Therefore, it is important to analyse the cells in their 3D microenvironment with regard to viability and differentiation. Tetrazolium assays (WST-8 and MTS) are still the methods of choice for estimating the number of living, metabolically active cells, with WST-8 being cell-impermeable compared to MTS. In contrast to these methods, the ATP assay is an endpoint method based on the luciferase-induced reaction of ATP with luciferin after cell lysis. We compared three methodologically different proliferation/toxicity assays (MTS, WST-8, ATP) in monolayer (2D) and 3D culture systems to improve the technically challenging determination of the number of viable cells. Chondrocytes were isolated from human articular cartilage. Three different test systems (MTS, WST-8, ATP) were applied to monolayer cells (2D, varying cell numbers) and spheroids (3D, different sizes) in 96-well plates. The intracellular ATP concentration was determined by luciferase-induced reaction of ATP with luciferin using a luminometer. Formazan formation was measured spectrophotometrically after different incubation periods. Evaluation was performed by phase contrast microscopy (toxicity), correlation of cell count and ATP concentration or absorption signal (Gompertz function) and propidium iodide (PI) staining to proof the cell lysis of all cells in spheroids. In 2D culture, all three assays showed a good correlation between the number of seeded cells and the ATP concentration or absorption data, whereas the MTS-assay showed the lowest specificity. In 3D culture, the spheroid sizes were directly related to the number of cells seeded. The absorption data of the WST-8 and MTS assay correlated only for certain spheroid size ranges, whereas the MTS-assay showed again the lowest specificity. Only the measured intracellular ATP content showed a linear correlation with all spheroid sizes ranging from 100-1000 μm. The WST-8 assay revealed the second-best sensitivity which allows the measurement of spheroids larger than 240 μm. Phase contrast observation of monolayer cells showed toxic effects of MTS after 6 h incubation and no signs of toxicity of WST-8. Staining with propidium iodide showed complete lysis of all cells in a spheroid in the ATP assay. Among tetrazolium-based assays, WST-8 is preferable to MTS because of its non-toxicity and better sensitivity. When determining the number of viable cells in the 2D system, caution is advised when using the ATP assay because of its two-phase slope of the correlation graph concerning cell number and intracellular ATP. In 3D systems of human chondrocytes, the ATP-assay is superior to the other two test systems, as the correlation graph between cell number and intracellular ATP is biphasic. Since differentiation processes or other metabolic events can influence the results of proliferation and toxicity assays (determination of viable cells), this should be taken into account when using these test systems.