Gpt 3.5 turbo worse after upgrade to paid account

I was experimenting with ChatCompletion using the free version of gpt-3.5-turbo and got good results.
But, when I upgraded to a paid account to improve the latency, the results were poor. This was done using the same script, the same data, and the same API key as the free version.
Any clue on what’s happening?
Thank you,

Closing this topic.

Please consider using this topic instead.

GPT-4 has been severely downgraded (topic curation)