All low-end VPS offerings use shared cores, so # of cores is a fairly useless comparison unless they also tell you the oversubscription rate or all of your neighbors are idle.
I don't know about the others, but on AWS with the T2.* instances it always seems like I've got the server to myself. They are either well provisioned, or well managed. The burstable CPUs are great.
You do have them to yourself, as long as you don't use them.
T2 instances work on a credit system, where you are allocated X credits per minute. If you use less than you're allocated, you will rack up hours of CPU time (there is some cap). You can see the credits in the AWS dashboard.
Problem is, that buffer you have is also the delay between when you start doing too much work and when you find out you're doing too much work because the instance isn't responding. And it's the delay between when you fix the problem and your balance goes back to "normal".
I recently hosted my personal tree of Fast-Growing-Programming-Language on a T2, and "accidentally" became the #1 google search result for the entire long-tail of every stack trace and every error message. Took me awhile to piece together what was wrong.