Introduced by Hewlett-Packard in 1974 at an MSRP of $795[1] (equivalent to $4,912 in 2023)[2], it featured nine storage registers and room for 100 keystroke instructions.

Like all Hewlett-Packard calculators of the era and most since, the HP-65 used Reverse Polish Notation (RPN) and a four-level automatic operand stack.

The HP-65 had a program memory for up to 100 instructions of six bits which included subroutine calls and conditional branching based on comparison of x and y registers.

HP also sold a number of program collections for scientific and engineering applications on sets of prerecorded (and write-protected) cards.

The HP-65 had an issue/design flaw whereby storage register R9 was corrupted whenever the user (or program) executed trigonometric functions or performed comparison tests; this kind of issue was common in many early calculators, caused by a lack of memory due to cost, power, or size considerations.

Another program for the HP-65 allowed the crew to compute pointing angles for the spacecraft antenna for aiming at the ATS-6 communications relay satellite.

In the same year, Mitchell Feigenbaum, using the small HP-65 calculator he had been issued at the Los Alamos National Laboratory, discovered that the ratio of the difference between the values at which successive period-doubling bifurcations occur tends to a constant of around 4.6692...

Steve Wozniak sold his HP scientific calculator to partly fund the first batch of Apple 1 computers.