I think, quite honestly, we should publish a new version (or release candidate)
We can not make releases candidates with known bugs, that would be a shortsighted strategy IMO. I believe a good beta is better than a bad release candidate or a bad stable release. We can not obviate the fact that there is significative bugs pending to be fixed. Unstable releases would doom this project much faster than lack of developers IMO.
There is room for improvements about how the betas are released though. Maybe we should shift to installers again instead of using special Blender builds.
What are the future plans for YafaRay itself?
I don't think there is a plan, only day-to-day work. There are many people commited to work on YafaRay right now, but the fact is nobody is skilled enough to work on YafaRay core structures at the moment, except from small fixes.
I can't really say much more. I am very proud of the work done so far. I have come to realise that it is worthless to rush things out, YafaRay will progress as long as there are people willing to work on it. Of course it will not progress as much as people would expect or wish, but we can do little about it. I can not make people work on YafaRay against their disposition. Besides, Blender is an special case in the field of FOSS desktop projects, comparing every other FOSS project with Blender development pace is a common mistake and a terrible idea.
I am optimistic about YafaRay, I believe I will do well against Cycles, as long as we do things right. I think the code sprint could be a very good thing for YafaRay right now. The idea is to motivate this new generation of developers (Olaf Arnold, Povmaniaco, Gabriele Ferreri, Rubdos, etc) interested in YafaRay. As soon as I find hosting premises for the code sprint, it will be oficially announced.