What about plugging the numbers into the SRES scenarios, the report that Stern used. These are rough figures in 1990 US$
A1FI - GDP/capita developed countries = 107,000, developing countries = 66,500.
B1 - GDP/capita developed countries = 72,800, developing = 40,200
Median warming for A1FI is 4 degrees. Allowing for 5% damages in 2100 would reduce the welfare here to 100,000 and 63,000 approx. So a warmer but richer world, characterised by modest emissions abatement, is more optimal.
Under A1FI, global inequality is also substantially reduced from today.
You might think that all of this is a bit glib, but these are official figures as used by the IPCC and from the IPCC.
Reference.
http://www.cato.org/pubs/pas/pa-609.pdf