The results of the sixth phase of the Coupled Model Intercomparison Project (CMIP) are currently being analysed and will form the basis of the IPCC Sixth Assessment Report. Since its creation in the mid-1990s, CMIP has had an increasing influence on climate research. While the principle behind it has always remained the same-comparing different climate models under similar conditions-its design and motivations have evolved significantly over the phases of the project. This evolution is closely linked to the one of the IPCC, since, historically as well as today, the results of CMIP have played a major role in the panel reports. This role increased the visibility of CMIP-over time, more and more people started to be interested in CMIP and to analyze its results. Despite this success, the way CMIP is used today raises methodological issues. In fact, CMIP has promoted a particular way of doing climate research, centred on a single tool, Global Coupled Models (GCMs), and creating a gap between model developers and model users. Due to the debates regarding the interpretation of multi-model ensembles and the validation of GCMs, whether the emphasis on this particular way of studying climate is serving the progress of climate science is questionable.