This model's maximum context length is 4097 tokens. However, you requested 4238 tokens (1238 in the messages, 3000 in the completion). Please reduce the length of the messages or completion.