Python Asyncio Response Mixing From GPT-4

Different Runs With Different Data In A Short Period Of Time Results In Responses Mixing. Why would this be the case if when triggering a run seperate data is being sent over to gpt-4.