I wouldn't say reliability or other features of metal case parts is poor (in terms of advantage). Though I don't know by how much that actually matters. So, I don't have the numbers, or even the test methodology to establish that; I would just suspect they have at least some reason besides inertia.
The main direct / advertised difference is TJ going up to 175 or 200C, which is, at best, only rarely available in plastic packaged parts.
But yeah, if you aren't doing high temperature, mil spec, down hole, etc. stuff, you'll likely never see the difference; plastic case parts are amazingly reliable.
You'll see package differences on the highest speed parts (fT > a few gigs), but that's about parasitics. The die, in and of itself, will be perfectly identical across parts of a given type but different package.
Tim