Elsevier's ScienceDirect platform publishes over 500,000 articles a year, across 4,300 journals, providing over 18 million documents and 42,000 e-books, and serving over 18 million unique monthly page views (https://www.relx.com/~/media/Files/R/RELX-Group/documents/re...). Everything published goes through a process of peer review, editing, layout, final review, etc, and then access to the publication is maintained in perpetuity for thousands of institutions around in about a hundred countries.

Here are their pricing policies: https://www.elsevier.com/about/policies/pricing and their other policies, which gives you an idea what else is involved with managing journals: https://www.elsevier.com/about/policies

So, yeah, journals do have costs. I am not defending those costs. But it is worth pointing out that there's more to publishing than a free peer review and uploading it to GitHub.

Peer-review is free. Editing and Lay-out etc is done with the journal (not much though as they supply latex templates for your work). So there is some cost associated with this.

I haven’t heard of any reviewer that was compensated for a review, though I once won a reviewer award that had a small honorarium.

Editing is generally a paid service — the journal will not correct your typos or your prose. They will find incomplete reference data and ask you to fix them, and that’s about it