The lightest photon ever detected

Okay, so I am taking this question to mean what is the lowest-energy photon that can be individually detected. This is certainly an interesting technological question.

I can't give an authoritative answer, but the lowest energy detectors I am familiar with is at the CMB microwave background energy of ~ $3Kk_B$, which corresponds to a wavelength of about 5 mm, or a frequency of around 60 GHz. These are the transition edge sensing bolometers that are used in the Bicep 2 experiment and similar experiments.

The way that these work is that they use a superconducting material just below the superconducting transition temperature, which is heated up enough by absorption of one photon to transition to the normal temperature. This changes the resistance, which is ultimately read out as a slight increase in the amount of heat dissipated.

The limit to the lowest energy photon one can detect with these sensors is given most directly by the size of the superconducting bandgap. 3 K already corresponds to a ~0.25 meV gap, which is on the low end of materials as far as I know (compare for example with this chart). So I don't think one could use this to get a whole lot farther, certainly not down to radio wave scales. But I welcome any thoughts on this matter.

Edit 03/2019: A new result shows detection of photons at around 200 MHz (corresponding to a wavelength of 10 m) that are stored in a resonator, using coupling to a superconducting qubit similar to the systems mentioned by Daniel Sank in the comments. The natural frequency of the qubit is in the GHz, so they had to do some clever designing to make it sensitive to such a lower frequency.

The article says "Consequently, there can only ever be an experimental lower bound on the mass of a supposedly massless particle; in the case of the photon, this confirmed lower bound is of the order of $3×10^{−27} eV = 10^{−62} kg$." It is saying that the rest mass of the photon, if it exists, is less than $10^{−62} kg$ which is different to the frequency of the photon. Effectively it is making a comment on the range that the inverse square is accurate, which turns out to be $\frac{\hbar}{mc} \approx 10^{20}m$.