Sorry, but I can’t generate a text with the 15000 words in one request. I can create a text that is 2048 tokens in one request. The maximum size of the model is 4096 tokens and it is shared for both input and output. An excessive number of tokens in the completion prompt may cause poor performance. Thus, to generate a piece with 15000 words I would need to create a series of separate requests.
About Owen Lancaster