The speed of sound in the glass doesn't play any direct role in this case.
The glass is producing sound by means of the standing wave that forms in its rim, and these waves' properties are a function not only of the material (with its elasticity and sound speed), but also of its geometry: thickness, shape, size, perhaps amount of liquid in the container, etc.
So, from the recorded frequency $f$, you obtain neither the speed of sound in the glass, nor in air, but the speed of the wave in the glass rim. And even for that, first you need to find out which mode has been excited by the sliding finger. Oversimplifying a bit, that means finding what's the number $n$ of wavelengths in that rim: then $\lambda=\pi r/n$, and $v= \pi r f/n$, but probably exciting the fundamental mode ($n=1$) is the easiest.
For more details, there are of couple of easily found papers with measurements and more realistic modeling, such as Jundt et al. Vibrational modes of partly filled wine glasses (e-print).