Linux

If you're referencing an isolatedProccess implementation, the benefit is that each site is isolated in its own process, and any exploit would only have access to its own process (the data that the site sees anyways) without further escape (kernel exploit or meltdown, for instance). Without this isolation flag, sites are not sandboxed from each other or from the browser's process itself, meaning an exploit could access any data from any other active site or from the browser's process (such as accessing browser settings, bookmarks, history, or the built-in browser password manager). This has a massive implication on security. I'm unaware of the sandboxing you mentioned before Chrome, so I can't comment on that, but you gain a lot of security from proper per-site process isolation. Yes, the app lives inside its own sandbox, but there's plenty of data within that sandbox that you may not want a site to access, hence the importance of the isolatedProcess flag.