Has the speed of light ever been measured in vacuum?

Science is full of ideals in its wordings. This is one of them.

SI has fixed the speed of light in a vacuum to be 299,792,458 m/s. If there was indeed light propagating through a perfect vacuum, that would be its speed ... because we define it to be.

For practical purposes, however, we need to be able to design experiments with which to measure distances using this definition. We have done these sorts of experiment regularly in high vacuum, on par with or more extreme than the vacuum of interstellar space. When we look at the effect matter has in slowing the speed of light, we find that the difference between its speed in a perfect vacuum and an achievable vacuum is smaller than the measurement error on our experimental devices. Before we had fixed the speed of light to be a constant, we had measured it to within 1 m/s.

How much of an effect does it have? I'm having trouble finding sources to give a definitive answer, but based on the refractive index of hydrogen as a function of pressure, I would expect interstellar levels of hydrogen to slow light by a factor on the order of µm/s. It's very difficult to measure physical things to 8 or 9 digits, and µm/s is 15 digits away from the speed of light, so our measurements in a high vacuum are as usable as if they were in a perfect vacuum.

If, at some point in the future, we discover that this approach is flawed, we will amend it, as has been done several times before—the most recent amendment being fixing the kilogram as a function of several fundamental constants.

The “few atoms of hydrogen per cubic meter” is a vacuum for visible light. The adjective “perfect” is both unnecessary and irrelevant in describing light propagating through a vacuum.

The wavelength of visible light is around 500 nm. So roughly $8 \ 10^{18}$ wavelengths fit in a cubic meter. A few hydrogen atoms are utterly irrelevant. Their presence cannot explain the wave behavior of light and from the perspective of such a phenomena the region is vacuum.

Since neutrinos don’t interact with light, their presence is also unimportant, regardless of their quantity. And photons are the same thing as light so it hardly makes sense to complain about the presence of photons when discussing the presence of light.

Trying to address your confusion about whether the definitions are circular... I think the root of your confusion may be in understanding what this means from the Wikipedia article you quoted:

Its exact value is 299,792,458 metres per second

You bolded the above sentence, but it is not really the important one. The important one was later in your quote:

a metre is defined as the length of the path travelled by light in vacuum during a time interval of 1/299792458 second

I think you also may be confused about the difference between values and the physical quantities they represent.

So to determine what a meter is based on the above definition, you first need to measure the speed of light. Imagine you are able to make an experiment that measures how far light goes in 1 second. Perhaps it draws pencil marks on two walls very far apart. Say you run the experiment 10 times and look at the different results. The outputs from the different runs of the are not values like [299792458.1 m, 299792457.7 m, 299792458.3 m, ...]. Instead, the outputs are all called "299792458 m". Of course they do have measurement error, but the error is in the actual physical measurement (the locations of the pencil marks).

If someone wants to use this experiment to, say, manufacture a very accurate meter stick (maybe a "light-second stick" in this case), they would need to actually use their meter stick to measure the distance between the pencil marks on the wall, and use that to know whether their meter stick is too large or two small and adjust their manufacturing process based on that.

Does this give you any insight, or am I off base here about your confusion?