Purely in performance terms, how do CONTAINS([MyCol], 'foo') and [MyCol] LIKE '%foo%' compare?

Consider CONTAINS([MyCol], 'foo') and [MyCol] LIKE '%foo%'. Assume the best possible indexes on [MyCol], whatever those may be (at minimum, we need a unique key and a full-text index). I know that the two expressions are not equivalent, but how do their performance characteristics compare? Assume SQL Server 2019. I am asking because I have inherited a codebase where the first expression was frequently preferred over the second, without any documentation explaining why.