Attached a PR to fix this for the Assistants API, as those are the requests, that are failing right now without setting the Content-Type explicitly to application/json. However, having a look at the documentation, it seems like setting the Content-Type for all requests is the correct way:

curl https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
     "model": "gpt-3.5-turbo",
     "messages": [{"role": "user", "content": "Say this is a test!"}],
     "temperature": 0.7
   }'

(example request from their documentation)

Hi @benedictstrube. I was having the same issue but out of nowhere it seems to be working again.

Thanks @propertius for the heads up! I already posted a PR that sets the Content Type explicitly according to specification. I think it still might be a good idea to merge this, as this might happen again anytime in the future, if OpenAI decides to be restrictive regarding that. 😄