They opened the whole thing up for unlimited access for a brief period during the pandemic, but decided to walk that back after just three months. If you know where to look, there's a 500gb torrent floating round with the 480k+ papers that were accessible as of June 2020.
It's sad that in this day and age, particularly with the widespread acceptance of open source, most academic publications are still behind a paywall. We shouldn't even be having discussions about "open access"; the "open" part should just be implicit.