The results of IOPS performance tests depend for a large part on the settings used in the benchmarking tool. For any meaningful comparison with other drives (from the same manufacturer or not) you need to know what parameters were used in the test.
In addition to shodanshok's anwser :
4K = 4 kilobyte = the size of the read/write IO requests used in the test.
Aligning the IO request size with the native sector size used on a drive results in optimal (test) performance.
QD16 = the Queue Depth used in the test.
By using the native command queuing a disk can optimize the order in which read/write commands are executed on the actual permanent storage. A high queue number is mainly of interest with SSD's as with their parallel channels maximizing the utilization of each channel requires sufficiently high queue depths, but even with spinning rust you see performance improvements by allowing the disk controller to optimize the order of requests.