Request to https://api.openai.com/v1/chat/completions failed, reason: socket hang up

I am trying the new mini model today.
I’m using the latest version of the npm library, with js.

For any small prompt ( a sentence ) it works fine.
For any large prompt, I keep getting an error

request to https://api.openai.com/v1/chat/completions failed, reason: socket hang up

more specifically:

APIConnectionError: Connection error.
    at OpenAI.makeRequest (file:///.../node_modules/openai/core.mjs:297:19)
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)

Why are all the API calls failing with long prompts?

and btw, I’m charged for all the tokens somehow!

( trying the same exact prompt in the mini model in the page, works fine and gets a pretty good answer )