langgraph/Q/Is the graph a big runnable and do I need to prefix model configuration

• TODO In LangGraph, is the entire graph a single LangChain/Runnable, and if so, do I need to prefix the configurable if I want to use more than one model in the graph, for example, in different nodes?

  • It seems like each node assembles its own runnable independently, in which case I wouldn't necessarily have to prefix models in LangChain/chat_models/init_chat_model if I wanted to use just one model per node