What are the assumptions of probit model?

Can you please explain the fundamental assumptions of the probit model, and how they differ from other regression models like the linear regression? Specifically, how does the probit model handle the binary dependent variable and what are the statistical implications of these assumptions on the estimation process and the interpretation of results? Additionally, could you discuss any potential limitations or challenges associated with these assumptions in real-world applications, especially in the context of cryptocurrency and finance?