TodoMVC is a popular UI example use-case, which illustrates basic interactivity concepts. Webdevs will often implement TodoMVC when learning a new framework to get the hang of all the core concepts.
And well, there's a lot of frameworks, which may all have different performance in different browsers, so this benchmark tests many different implementations of TodoMVC, all done in different frameworks.
Ultimately, it tries to simulate normal web usage, it's not some speciality benchmark.
I would still use FF for moral reasons but I'd understand if uses it for the things you mention, but saying it's "faster" isn't really a good term in this case, faster in what? I mean, I'm not saying this is done in bad faith or anything, but would be even better if we could know that instead of simply clamoring over "fastness".