The expectation is that GPT-5 would be the next GPT-N but 100x the training compute of GPT-4, but that would probably cost tens of $billions, so GPT-N scaling is over for now.
The expectation is that GPT-5 would be the next GPT-N but 100x the training compute of GPT-4, but that would probably cost tens of $billions, so GPT-N scaling is over for now.