Yes I've tested both bluetooth headsets as well as bluetooth speakers and they work fine. You can even record from a bluetooth headset and use the transport controls on these headsets to control SONAR.
The latency for bluetooth is high so its not useful for input monitoring - you can playback and mix with them however. With WASAPI exclusive you can run at a low buffer size so you can reduce the latency to just the bluetooth latency.

Thanks for the info, Noel. Any thoughts on why record latency compensation would be insufficent and variable? Would this be due to the interface driver mis-reporting input latency? As noted above, I thought it was odd that the real-time RTL was consistent within a given session, but the record latency compensation was not.

There is no mechanism for WASAPI or WDM for that matter to report input or output hardware latency.

The very original reason I joined this forum was because I found my 2003 Home Studio audio tracks were randomly in the wrong place. I had been using a MD 8 digital recorder and an Atari which gave me rock solid timing between the sequences and the audio. But I knew that equipment was becoming dated and looking to use a DAW, so I dove in head first.  I had Cakewalk HS which came with my Roland Sound Canvas.
 
The forum soon pointed out the fact that my Creative Audigy II interface ASIO driver was the issue.  Someone here told me about the loop back test and that was exactly what I found. Not only were my tracks out of sync,, it kept changing.  They also drifted over time.  This my friends is 100% audio driver related, that was 13 years ago and XP 32 bit.
 
I bought an M Audio Fast Track pro interface and the problem was solved.  
 
So even though WASPI might "work" and reported latency will go down, don't trust it until you run this test, I didn't do the second test which is to loop back 3 minutes of audio and check the drift over time.. that's yet another issue that develops when using so, so audio drivers. 
 
It will be good news if WASPI under Windows 10 is as "good" as using ASIO, in theory why not? 
But myself I will be sceptical until I get a solid report from this test.  
Brundlefly's results under W10 are what I'm talking about.
Most newbies to audio might not even notice the bad timing and certainly if your not recording audio it won't matter at all. We will be able to at least do some editing and midi work on our laptops while out and about, that's good stuff.
 
The results may depend on which on board sound card is involved,, a shoot out between Real Tech an others.

Noel Borthwick [Cakewalk]
There is no mechanism for WASAPI or WDM for that matter to report input or output hardware latency.

Hmmm... so WDM and WASAPI modes do no automatic record latency compensation, and you have to do it all with Manual Offset? I guess that makes sense given that the average offset I saw in testing was approximately half the RTL. But that doesn't explain the variability in the recorded click position given that the latency was not varying from one take to the next.
 
I played around with this some more, and after a while every take I recorded was being properly compensated by a Manual Offset of 249 samples, but initially that definitely wasn't the case.
 
That said, although I prefer ASIO's sample-accurate consistency, I would not really be too concerned about a ±1ms variability in record latency being audible in most real-world situations.

yea, I know but... So with these drivers? I can stop using an audio interface, just use my headphones and Sonarworks to mix?
 
peace!

Cactus Music
I didn't do the second test which is to loop back 3 minutes of audio and check the drift over time.. that's yet another issue that develops when using so, so audio drivers.

That shouldn't be possible so long as Playback and Record Timing Masters are referencing channels on the same device.

You just brought up something that then made me think,    Possibly the whole issue with so so audio drivers are they are responsible for the word clock, right!    that might explain why things go out of sync. Iffy word clock.
I do know serious studios use devices like a Apogee Big Ben for word clock. I think they sync up the would studio to the one clock.   
But I believe this comes back to for us with modest little home studios that good audio drivers have fairly solid word clock.  
 
That said, although I prefer ASIO's sample-accurate consistency, I would not really be too concerned about a ±1ms variability in record latency being audible in most real-world situations.   
 
Agreed but as you see in my results it was more than a 1/16 note ( I forgot to change my timeline before taking the screenshot)  I think that's like 1,000 samples. 

Tested... Mixing, Full load of plugins and tracks... Flawless! and sounds no different then my Lexicon Alpha (ASIO)  
I love the WASAPI exclusive!!!
 
Now that's what I'm talking about!
THANK YOU!
 

Cactus Music
Agreed but as you see in my results it was more than a 1/16 note ( I forgot to change my timeline before taking the screenshot)  I think that's like 1,000 samples.

A fixed offset can be addressed by setting a Manual Offset in Preferences > Audio > Sync and Caching. The variability should only be a fraction of the total input latency.