Many have already asked, what to do, if they finished the 5000 words lingvist provides.
As you've claryfied, the vocubalary for each language is collected from real world content and ordered according to its relevance in general texts/content. When one knows 5000 words and is able to understand approximately 90 % of general texts, then each common word learned above the first 5000 is getting less and less valuable, because the chance that one has to read/understand this word is getting lower and lower.
I think, that many users may want to understand texts of some specific type or domain (like medicine, laws, economy, science fiction, cooking, gardening, politics, middle Ages). And even if they learn the 10.000 most common words of a language, the chance, that the needed words to understand the texts they want are among the 10.000 most common is very low.
My suggestion for Lingvist is the following: If you've already collected the 5000 most common words through a big data analysis, then you may also be able to collect an unspecified amount of words, which are common to different topics (e. g. like the ones I mentioned above).
Maybe this could also be a good idea for the monetization of Lingvist. I'm sure a lawyer or an MBA are willingly to pay, if they could learn the 1,000, 2,000, ..., X,000 most common words of their domain.