The correct definition of a vacuum in terms of general relativity, is a system with an isolated mass. You can have an isolated mass, with an asymptotically flat spacetime, which implies that the stress-energy tensor in the Einstein field equations is zero. Just for your own information, there is only one spherically symmetric, static, and vacuum solution in general relativity, it is the Schwarzschild solution.
You will notice from the Schwarzschild metric that although it is a vacuum solution, there is a "Mass" term. The most unfortunate thing about how people describe vacuums is they say "space is empty", this is not correct, because a vacuum solution has the existence of a mass. What they mean to say is that a vacuum physically constitutes in this case an isolated mass in an asymptotically flat spacetime.
The second possible case of a vacuum that is non-empty is if one considers a non-zero cosmological constant. Looking at the field equations, with a vacuum (Tab = 0), we see that:
Rab - 1/2Rgab + Lambda g_ab = 0
Which implies that upon taking the trace:
R = -4Lambda.
Clearly, this is a vacuum, but the Ricci scalar is non-zero, so the spacetime is clearly not empty. The Lambda has been thought to correspond to dark energy.