Ok, here comes a long post.
😉
The way to read the graph - the distance between the peaks of the graph corresponds to the frequency of the signal, so the closer they are to each other, the higher the frequency and the higher pitch you're going to hear.
The displacement along the y-axis corresponds to intensity or how loud you're going to hear the signal.
Based on that, you cannot say many nice things about the graph of 2/4/5. There is something that might resemble a beat frequency pattern (the really high peaks) but it's very rough and will sound more like noise, if you could hear such low frequencies.
This is due to really poorly picked frequencies in the problem. To talk about beat frequency, you want two signals with a small difference which is also noticeably different from the their frequencies. So 1000/1002 Hz would be a good pair, 3/5 Hz not really good.
First let's look at this:
This is the combination of 50 Hz and 51 Hz signal. The rapid oscillations up and down are what you're going to hear as frequency of the sound (it will be 50.5 Hz). The big sine wave that is formed around the graph is what you're going to hear as intensity - it will be increasing from zero to loud and back to zero once each second.
Now, if you go to some higher freq, like 1200/1202, you get an even more obvious picture:
The oscillations are so close to each other here that you don't really see them. The big wave is just a sound with a pitch fo 1201 Hz going from quite to loud and quite again twice a second. (also, ignore the white space in the solid regions - it's just the graphing program averaging things, it should be equally 'solid' everywhere)
Now if we introduce a third signal, things get more complicated. Let's see what 1200/1202/1204 looks like:
As you can see, we don't have such a nice patter as we had for two frequencies. We'll still get fluctuations in the intensity, but they'll follow a more complex pattern, something like quiet-loud-quiet-really loud-quiet-loud etc..
And it gets even worse if the difference between each two frequencies is different, like 1200/1202/1205.
As you can see, you're dealing with 3 separate peak intensities here and you're not going to completely quiet state either.
While the differences between the different peaks can be difference between the original frequencies, it's not clear what you would like to call 'beat frequency' in that case. Is it any change of increasing to decreasing intensity? Is it between the loudest peaks? Or something else?
All this goes a bit further than intro physics, so I'm not sure why they are bothering with the 3 signals problem. But the problem is deficient in so many ways that I really doubt the writer was paying attention while he was creating the question.