Llama Index Router Query Engine from Building Agentic RAG With LlamaIndex course fails with gpt-4o-mini Open AI model

Hi! If you use gpt-4o-mini the Router Query Engine will fail. It can’t parse a json. Do you now how to fix that?

ValueError: Failed to convert output to JSON: ‘json\n[]\n

It don’t will throw if use models like gpt-4o, is a specific mini model error.

It will work if you use PydanticSingleSelector