It's about 0.1 atom per cc, which is about 100,000 atoms per cubic meter.
the article refers to the fact that space is not a “perfect” vacuum, but it's pretty darn close to one.
Ultra High Vacuum is considered to be 10-12 to 10-15 atmospheres. [One trillionth, to one quadrillionth of an atmosphere.] There are some labs doing Extreme Ultra High Vacuum, which is anything less than 10-15.
In contrast, interstellar space, over which this measurement is taken is on the order of 10-20 atmosphere, which is around 100,000 times better [weaker] than what is considered an extreme terrestrial vacuum.
So, yeah. Pretty darn good.
the old analogy being an atomic nucleus being the size of a baseball, the electron (back in the days they were considered physical objects instead of probability clouds) would be the size of a grain of sand several hundred feet away from the nucleus and the next closest atom would be found a few miles away.
The modern picture is that a proton has a spatial extent of about 1.5 x 10-15 meters, and the Bohr radius of hydrogen, where you expect to find an electron "most of the time" is about 5.3 x 10-11 m. So, if the proton were a baseball, [regulation: 37 +/- 2 mm radius (though it looks bigger when it's a curve ball)] the probability cloud of the electron extends around the proton-baseball in a sphere with a radius of about 1.3 kilometers.
The "next closest atom" depends on a lot of things, but in condensed matter it would actually be "pretty close." For example, molecular hydrogen has an average size about twice the Bohr radius. So, the electron clouds actually get very close, and in favorable bonding orbitals the intra-molecular distances can actually be "smaller" than the constituent atoms. [But, order of magnitude, the overlap is usually about like what you see in atomic hydrogen vs. molecular hydrogen. Some, but "not much" overlap of the electron clouds of the respective single atoms. This is not to say that the molecular orbitals are qualitatively the same as atomic orbitals. They aren't, or there would be no such thing as chemistry.]
This being the case, the space inside a glass lens at the scale of a photon would be as empty as the interstellar medium, to the photon it would be a vacuum, but it travels much slower through the glass than through “space”. Why do you suppose this is?
The correct answer is that condensed matter [and even gases at atmospheric pressure] are not really very much like empty space, because the old, oversimplified picture is wrong. Even if you just consider the nucleons, it doesn't hold up all that well, because proton and electron are not inert baseballs or grains of sand. The electromagnetic force is tremendously powerful, so the likelihood that an electron or an atom or molecule will interact with a photon is quite high.
This is also the reason why condensed matter is largely impenetrable: the electromagnetic forces holding the structure together, and binding the electrons to the lattice, repel [at the atomic level] other condensed matter that tries to "push them too hard".
On the other hand, it also explains friction, because, as long as you don't "push too hard" the protons in the surface are trying to bond to electrons in the materials pushed up against them.
These are, of course huge oversimplifications. The real bottom line is that the electromagnetic interaction is so powerful, that photons, which are excitations of the electromagnetic field, couple so strongly to charge carriers that you don't have to get atoms too close together before it's impossible for photons to avoid them. In "Space" the distance between atoms isn't close enough, but in the atmosphere it is. Go back to the vacuum comparison: Vacuum is a direct function of how many particles there are per unit volume. At atmospheric pressure, there are 1019 molecules per cc, ~20 orders of magnitude more electromagnetic charge carriers than in empty space. That's a lot more chances for the electromagnetic interaction -- which is 1036 times stronger than gravity -- to grab a photon flying by...