In working with the IBM quantum computer, one can use for example the composer in order to build simple quantum algorithms. Essentially, one usually feeds a phase in order to affect one of the qubits. While for instance phases like $\pi/2$ can be applied "faithfully" into the real backends, one can add in the composer phases like
I would like to ask what is the actual "precision", i.e. up to which significant figure does the real backends can achieve. I suppose it also depends on the backend that one chooses, so it should be a list of how much precision we can achieve in each one of the real machines.