That model is currently overloaded with other requests with gpt-3.5-turbo

I experienced the same yesterday, and it’s continuing today.