Unfortunately, OpenAI's api offering has slowed down a lot and is extremely inconsistent in response times. (I'm assuming they are struggling to scale and/or sometimes you get session-ed up with a congested data centre)
This is a pretty common theme for the past few months in their community forums -> https://community.openai.com/t/api-gpt-3-5-turbo-sucks-slow/147230/8
If anyone knows a less congested alternative I would greatly appreciate it (don't mind if it's of lesser quality responses)