I didn't think I had to, since 72 is the optimal temp, but I was told that since my organism has a high A-T content that I should decrease the extension temp
It's complicated. There isn't an easy way to predict this. You should test multiple things. But all things being equal, if you can test only one thing, I would, on average, lean towards increasing the time.
If you're away from your enzyme's optimal temp (72C) in either direction, it should take more time. Maybe. If it wasn't PCR. Since it is PCR. Good luck :)
how long is your predicted amplicon? For best results with SYBR green qPCR, I usually choose shorter amplicons (100-150 bp) and use a two-step cycling protocol with no extension step at all.
If the longest is 160 bp you can likely just do a two-step cycling protocol, especially if your primer Tm is on the higher side (60C or more). I'd be sure to run a standard curve and check your efficiency, as well as inspect your melting curves.
Speaking of Tm, I am coming across different temperatures for my primer, depending on which source I use, but I am not sure which one I should rely on. I've used http://www.basic.northwestern.edu/biotools/oligocalc.html and http://www.idtdna.com/calc/analyzer, but they're both giving me different Tm. The first one doesn't take Mg under account, but even when I don't account for Mg in the second tool the temperatures are off (56 vs 59 (or ~65 when I account for Mg conc.)). Any recommendation on which is more accurate?
Use whichever calculator or formula is recommended by the manufacturer of your SYBR-Green assay, but be prepared to do some additional optimization depending on your PCR efficiency and melt curve.