VM memory usage is a percentage of currently used memory by the VM, as you can see from the screenshot it is calculated by dividing memory active by the machine memory amount. You cannot really compare the numbers taken from Task Manager and VMware, as there are many variables that can force Guest OS think that it is using maximum of allocated memory, while it might not be 100% true.
The point is that there are multiple memory management techniques that allow you to use more memory than you have on the physical host (memory sharing, memory compression, zipped memory etc,), this will all lead to discrepancy between Task Manager and VMware memory views. Moreover, if you configure resource reservations and limits for the VM, this will also lead to different utilization numbers.
See this similar topic for more details: Differences between Task Manager and Veeam
In addition to this, please note that some applications like SQL Server will consume all memory that was granted to the VM, but it does not mean that SQL Server needs more memory and you cannot reclaim it.
That said, if you want to know whether you can reduce configured memory amount or not, looking at real-time stats is not the best idea, as you might have a peak load that happens every time SharePoint users log on to the portal and start working with it, in this case you definitely need to reserve some amount of RAM to satisfy the peak memory demand.
If I were you, I would run an Oversized VMs report, that will analyze and show you the average and peak memory usage for the past week, month etc. Based on that figures you will see your average workload and figure our whether you can reclaim memory resources from this machine or not.
Sorry for the long post, but I've tried to cover as much details as possible
Hope this helps!