This model's maximum context length is 4097 tokens. However, you requested 4098 tokens (1002 in the messages, 3096 in the completion). Please reduce the length of the messages or completion.