Cross-laminated timber (CLT) is increasingly being used in medium-rise timber buildings for a number of reasons, such as rapidity of construction, cost effectiveness and robustness. Like for other building materials, verification of the loadbearing performance in fire conditions is an important issue. Experimental fire tests have been performed on loaded CLT wall elements at research institutes in Sweden and Italy. In total, three large-scale and four medium-scale tests have been carried out. The aim was to gain information about initially protected and unprotected elements, to be used for classification and also for validation of calculation models. In the test series, reference tests at normal temperature were included to obtain information (e.g. stiffness, strength) about the material tested in fire conditions. In addition, model-scale fire tests were performed to investigate the loss in stiffness resulting from fire exposure and the effect of different protection types. Loaded fire tests varied in the range of 41.8 min to 120 min, depending on the CLT structure, the level of load, and the type of protection. Data on temperature within specimens and residual crosssections were collected. Charring rates evaluated from experimental results were comparable with values proposed by Eurocode for the design of timber structures. Conservative solutions were obtained by using simplified design methods and comparing their results to test results and results of advanced modelling. It was shown that the load-bearing performance of CLT may show abrupt changes due to its layered structure. It is strongly recommended that a minimum residual depth depending on the CLT structure should be required in order to ensure robust building products.