The models are fine ... just need better hardware to run the simulations ... we've reached exoscale and Moore's Law is still in effect ... but as I understand the system, it is the dynamic models they're using here ...
I think the radiative climate models are fine as well ... but we MUST keep in mind the several assumptions being made ... and that if any one assumption fails, then the whole of the modeling fails ... and I also think the IPCC's assumption of no increase in cloud cover is foolishly wrong ... but required or they'd have no reason to exist ...
Higher temperature --> higher humidity --> more clouds --> more rain ... as they say in the trade: || ..