When I wrote the above comment, I was ballparking it to be around 100nm, which I think is closer to µm than nm.
Looking a bit closer at it, I would say that the bands you see on the feet and body seems like the "precision" of the system. These are around 4 pixels wide. The 20µm ruler is 84 pixels wide, giving us 20µm / 84pixels ≈ 0.25µm/pixels.
So the band is 4 pixels * 0.25µm/pixels = 1µm. (Assuming my premises are right)
That's pretty amazing. I had the privilege of demoing repraps to a Vienna UT researcher a while ago, and he reciprocated by giving me and my friends a tour of the 3D printers at their institute. The 2-photon printing facility was already in place (and oh my is it impressive), but it hadn't made a lot of interesting stuff at the time - I think they were focusing on fabbing organic molecules at the time. I'm pretty sure it wasn't as fast as this - 4 minutes for a sculpture of that size, at that level of detail is pretty great.
That was about 2 years ago now. Very glad to see they made such great progress in the meantime (-: