I've only used CATT-Acoustic. Its pretty good for all three of the above and does some good stuff with Ambisonics if you set it up and have the right package/add-ons.
Be aware though that they all have the same limitation in that they do ray-tracing. I.e. they don't model sound as a wave which is a pretty basic error.
For most, simple, models this is ok but the user needs to be aware of what implications this will have when interpreting the results.
