At short wavelengths, our light detectors tend to be photon counters (e.g. CCD light sensor). At longer wavelengths, we use antennas and bolometers that tend to add up energy instead of counting photons. What is the longest wavelength at which we have "single isolated photon" type detection? Or, at least low number of discrete photons counted directly?
One cannot detect single energy events when there is a lot of thermal noise. So the detector needs to be cold to detect small quanta of energy.
Fortunately, there is a nice mechanism at low temperatures: some materials go superconducting, and this opens a superconducting gap in the density of states around the Fermi level. The size of the gap $2\Delta = 3.53 k_BT_c$ according to BCS theory, which is of the order of 1 meV. Excitations across this gap can be measured electronically. Not just counted, one can even do spectroscopy.
I do not know what the current world record is. But here is a recent article measuring wavelengths up to $\lambda = 90\ \mu$m: http://ieeexplore.ieee.org/document/7151327/