Consistent Hashing vs Rendezvous Hashing

You are building a client-side load balancer for a service with 5 backend servers. You need to implement consistent routing: the same request (same user ID) should always go to the same server unless that server is removed.

Compare implementing this with (a) consistent hashing and (b) rendezvous hashing. Which would you choose for 5 servers, and why?