Managing OpenAI Rate Limits with Request Prioritization