Thus far I have found the performance of the BladeRF to be superior to the HackRF, but as we've discussed elsewhere this may well be due to the 12 bit v. 8 bit architecture.
I've also had the BladeRF a lot longer than the HackRF, having received it in January and then the XB-200 HF/VHF transverter coming thereafter.
For me, both are plagued by overload, imaging, and aliasing. Some of this can be overcome with filtering, attenuation, etc. Some of it is my fault, particularly understanding the right balance in software to minimize imaging and aliasing problems.
The BladeRF seems to do quite well with seeing and tuning weak NFM signals in the VHF/UHF/SHF ranges. It has its problem areas; I can't see much at all in the 250-260 MHz area which is unfortunate as I would love to use it to see milsat transponders. I have some considerable interference issues here that I believe may be internal as they continue to appear with no antenna attached. I am plagued by images and aliases of strong signals in the 450 MHz area.
Owning both the BladeRF and the aforementioned transverter, one has to roll their own if they want it in a case. Nuand is behind the curve in providing a case in which both will fit. Therefore my setup sits in a safe stationary position and is not easily transportable, which I think it should be. Again perhaps my shortcoming for not having the proper tools to mount it in a customized case correctly.
The HackRF is a work in progress for me. I love its size, the fact that it's mounted in a case out of the box, and the fact that it's USB 2.0, which may hamstring it a bit bandwidth wise but makes it so easy to use on a number of different PC's. It's easy to install but if you are not a Linux tinkerer doing any command line updating such as firmware becomes more difficult. It seems to suffer greatly from the aliasing/imaging problems I described for the BladeRF, likely due to the same issues. It's ability to see and decode weak signals is not as good as the BladeRF's for obvious reasons. Signals overall seem to be weaker with it although perhaps this is an issue with achieving the proper balance of software settings.
I have tried the BladeRF on HF and was not impressed. It is totally overloaded unless the gain settings are pretty much turned off. I have not tried the HackRF in the HF bands yet. I do have the converter from NooElec that I purchased with it.
The dongles are pretty nifty devices for their low price and I have had a great deal of fun with them.
It also should be noted that at one point I had an AOR AR5000A+3 with an RFSpace SDR-14 as a panadapter. That was the best setup I had for spectrum monitoring and I have regrets that I sold it. It was a leap of faith to the current SDR technology, which is not at a par with it. The BladeRF can approach what the 5000/SDR-14 combination offered when it is properly balanced, but as of yet I've found no equal.