There is a new Model Selector node. Documentation says it “enables implementing fallback mechanisms for error handling“.
The problem is – it’s completely not clear how to use this node for implementing fallbacks. It only receives data from incoming nodes. Sub-nodes or responses are not working here.
Are there any clarifications on how to create fallbacks in Model Selector?
BTW, I’m not talking about built-in fallback feature inside LLM-root nodes. They only support a single fallback option.