dmesg(8) output wraps on systems which do not clear RAM on reboot.... note that the kernel messages start mid-buffer on both of these.
I can see that APM is used on both 4.3 and 4.5, so this is not an ACPI-related problem.
ACPI trouble is quite common, see the song for 4.5.
Other than that, I do not see anything obvious or helpful in the dmesgs. Someone else might.
A hung machine can sometimes be in panic and have dropped into the ddb(4) debugger -- yet the operator may not be aware of it, if X is active at the time of the hang. And the machine will really be hung, as there's no way to communicate with X to stop it, and no way to switch to the console, which is stuck in a ddb> prompt.
Was X in use at the time of the hang?
A machine which is hung, but looping in the kernel, can be forced into the ddb> debugger if the OS is still able to accept console interrupts, and not in X, and if the sysctl ddb.console=1 is set. This sysctl must be set while in single-user mode, or from /etc/sysctl.conf. The crash(8) man page has additional guidance.