I work on Mycobacterium genes and have seen a final concentration of 6% DMSO in my PCR reaction mixes give me optimum results. I have never used betaine though.
A really good discussion about this (I think) has always been in Susan Frackman's short publication -- (see attached).
In other reports, about 1.2M betaine in the presence of 4 mM Mg+2 seemed to be an optimal condition for dealing with GC-rich templates (and even stabilizing primer-target interactions in AT-rich regions.
Betaine between 0.25M and 1.25M in steps of 0.25M.
The optimal concentration will be specific to your area of interest, as it depends on the GC content in the surrounding area, not only within your amplicon, as the entire area needs to melt in order for the PCR reaction to proceed.
You could also try adjusting the denaturation temperature for the initial hotstart up to 96°C or 97°C, or digesting/shearing the template DNA.
DMSO and betaine have not the same effect. The stretch of GC induce a particular structure of DNA like a plateau. This structure is modified or disrupt by betaine. This help to open the DNA. DMSO modify the thermodynamic of the linkage between the base with its complementary. It help too to open DNA but it modify too the interaction between primers and their matrices.
I successfully amplified up to 63 percent GC-rich amplicons with the addition of 3% DMSO to GC-rich Hi-FI phusion buffer. Initial denaturing step up to 3 minutes at 95C would yield a better result.