Using gpt-4 API to Semantically Chunk Documents

This is the source document: A Court Trial - 64-0412 - Sermon preached by William Branham which is 25,847 Tokens /
105057 Characters

Here is the json hierarchy array:
https://s3.us-west-2.amazonaws.com/docs.scbbs.com/docs/test/node-5979-json.txt

Note that this output is over 7K tokens. I still think GPt-4o is limited to 4K output: What is the token-limit of the new version GPT 4o? - #3 by _j