Chrome is 63% (+ edge & Samsung as forks, makes it over 70%) of users, Safari is 20%, Firefox is 3%

Most organisations don’t want you ignoring 1 in 5 people, so unfortunately Safari testing (especially with all its shitty bugs) should be second only to Chrome for any professional work.

Also unfortunately Firefox is only 0.5% of mobile traffic, so it’s not representative of real users to only do your mobile viewport testing in that browser.