This has been an interesting topic in the past, although mainly around the 2OSC (the predecessor to the 2OSC/D). Robert designed the 2OSC/D to redress some of the noted shortcomings with the original 2OSC - especially the tracking. As Robert himself has said, the use of 1V/octave oscillators was not in the original brief for the Kickstarter, but was added later as a stretch as many customers were requesting it.
Personally, I found that I could only get about 1.5 octaves of good tracking after I calibrated my 2OSC. I have subsequently bought a 2OSC/D, but I haven't used that in the same way yet to compare it. You also have to be careful when you connect modulators to the CV input of the oscillators as the connections in the AE Modular are not buffered and so they will drag the voltage down a bit. I found that I had to re-tune my oscillators once I set up the various modulation paths. Once you realise this, it's quick to do. If you want to make a buffering module, there is something in the DIY section.
As I said, there was quite some discussion and my post here should link you to some of those discussions (although they were for the 2OSC, they are still relevant). You can also read down the thread a bit for more detail on how I set up my oscillators and on my findings on calibration. I am certainly not the best source, but I played a bit to get my AE Modular working well for me.
The 2OSC/d oscillator should track well for 3 1/2 octaves, the VCO around 3 octaves. Is the mentioned issue about the tracking (1V more CV is not +1 octave) or the general stability over time (meaning the frequency drifts after some time)?