I've published several articles in various prestigious scientific journals, the last one was in 2020. I never had to pay a cent! The only limit was the size of the article. If it exceeded a certain number of pages (I think the limit was about 10 pages of two-column format), then I would have to pay for the extra pages.
2000 € / review is outrageous!
Ok, it might be their "open access" journals to compete with PLoS. I think it was Science Advances or something like that and Nature was doing one too. If open access then yeah you have to pay but I believe the overall quality suffers (like PLoS)
Arxiv is great, Perelman published his proof of the Poincaré Conjecture there. Free, and serves its purpose just fine.
As we say, the odds are good, but the goods are odd.