Not OP but, I finally got access to GPT-4 32k and I thought that for a thing as simple as a summary it would be better than the smaller context-window model. Bu

Not OP but, I finally got access to GPT-4 32k and I thought that for a thing as simple as a summary it would be better than the smaller context-window model. But then I realized that now I have (practically) infinite context in input but what the LLM can output is always between 1-1.5k tokens. That is, I think, because of the samples in its training, not because it can't output a more lengthy output. In short, I think that multiple steps are better, for example in summarization.