When to use probit vs. logit?

In the realm of statistical modeling, the choice between probit and logit models can often be a perplexing one. So, when should one opt for a probit model over a logit model, and vice versa? What factors should be taken into consideration to make an informed decision? Is there a specific type of data or research question that leans more towards one model than the other? As a researcher or analyst, how can one ensure that their choice aligns with the assumptions and characteristics of each model?