The models participating in CMIP6 include the world's most complex, comprehensive representations of the climate system.
However, this complexity and comprehensiveness comes at a computational cost.
Here we examine the degree to which key characteristics of these models' response to climate forcings, for example global-mean temperature increase due to anthropogenic emissions of greenhouse gases, can be emulated with simplified models.
Evaluating this performance is of key importance as the simplified models are often used as emulators of the more complex models, allowing researchers to perform experiments which would be too expensive to run with full complexity models.
For example, simplied models will be used to perform climate assessment of the hundreds of socioeconomic scenarios submitted to Working Group 3 of the IPCC's Sixth Assessment Report.
By evaluating their emulation ability, we can better understand which parts of the Working Group 3 climate assessment capture the behaviour of the CMIP6 ensemble and where the limitations may lie.