The H716 was introduced in 1972. The references I've found for it mention this CPU being part of "System 700". It isn't very clear (to me) what exactly System 700 was, but the H716 CPU seems to have been fairly central to it.

As far as I can tell not very many H716 machines were built. I think that the machine just came too late. Microprocessors were coming, and the market for mini-computers, especially in the embedded control market, was about to disappear.

I've come across a rather prophetic reprint from IEEE Computer magazines "32 and 16 years ago" feature. This shows that in the same issue from July/August 1972 (cached) on page 58 there is an announcement of "System 700" and the H716 CPU costing from a mere $30000, and on page 60 an article about Intel introducing its second "computer on a chip" the 8008.

I don't know a great deal about the H716 and most of what I do know is gleaned from comments in verification and test program source code. It seems that an attempt was made to paper over the Series 16's most obvious architecural "quirk". This is the fact that its code is inherently self modifying because of the operation of the JST instruction in which the return address is stored in the target address of the instruction, this being the first location of the subroutine. The H716 introduces a hardware stac pointer.

Location 3 of core memory is the stack pointer. Stores to location 4 cause data to be pushed onto the stack. Loads from location 5 pop data from the stack. Similar "adjustments" to JMP and JST instructions allow the program counter to be pushed and popped from the stack.

The A register shadows location 1 of core memory in the same way as the X register shadows location zero on all Series 16 machines.

The B register shadows location 2 of core memory.