1 month after GPT-5 comes out, I will create a poll that will run for a month with three options:
GPT-5 exceeded the general expectations for it.
GPT-5 exactly met the general expectations for it.
GPT-5 fell below the general expectations for it
This market will resolve "Exceeded" if more people vote that GPT-5 exceeded the general expectations around it then vote that it fell below general expectations. It will resolve "fell below" if the opposite happens. Yes, this effectively means the votes of the people who say it exactly met general expectations won't do anything. The price of neutrality I guess.
Respondents will be asked about general expectations around GPT-5, not whether it exceeded or fell below their own expectations. Exactly how they interpret general expectations I will leave up to them.