Cluster creation best practices

Hello everyone. I need to deploy an Opensearch cluster. To do this, I have a server in the data center with the following specifications that I need to maximize.
CPU: 100 GHz
RAM: 1.7 TB
DISK: 15 TB SSD
I can use these resources to create the virtual machines that will form the cluster. Note that the cluster must have data nodes, master nodes, and machine learning nodes. For now, we need this to test the machine learning capabilities and how we can use them. What is the best way to allocate these resources and how many nodes of these types should you have?