After having tried this on my production hardware also I have unfortunately concluded that is isn't limited to virtualized instances of pfSense. My now live physical firewall does the same thing (ntp service crashes) when having it's ntp server peers on two different interfaces (WAN and LAN). :'(
Only using external ntp servers on the WAN interface have been stable for a few days now so that is the workaround I 'll have to settle with until the bug is resolved. Only connecting to the internal ntp server on the LAN interface is not an option for me for redundancy reasons, so I haven't tried that but my guess is that it would also work.
Maybe it isn't very common to have ntp server peers on multiple interfaces and maybe that is the reason this haven't been discovered in beta testing?