This OpenAI poll is pretty clever.

Tibo, who runs Codex, asked Pro users to pick one:
A: bring back the 5-hour limit, double the weekly quota
B: no 5-hour limit, quota unchanged

Quoted post: https://x.com/i/status/2109152716954222849

Some context first: B is exactly Pro’s current state. In July Codex temporarily removed the 5h limit; Plus already got it back at the end of August, and for Pro they only promised not to add it back for the next few months.

So what this poll is really asking is: are you willing to trade the 5h limit for a 2x?

  1. 2x isn’t as sweet as it looks
    Tibo himself has said the 5h limit is there to flatten compute peaks. Put that gate back, and how much you can use in a week gets capped by the 5h window. For people who never hit the weekly cap anyway, doubling it adds nothing.

  2. They can’t lose either way
    If A wins, bringing the limit back was the users’ own choice; if B wins, they keep the status quo and earn some goodwill on top. No 5h limit was one of the reasons some people picked Codex over Claude Code. Taking it away outright would get them flamed; letting users vote makes it a lot softer.

  3. Not much value as research
    Who hits the 5h limit and who hits the weekly cap is crystal clear from backend data. An X poll can’t filter for real Pro users either; of the 40,000-plus votes, nobody can say how many come from actual users.

  4. Interesting timing
    For new Pro 200 subscriptions, the quota was already cut starting September 29, and existing users keep the old quota until October 29. The poll doesn’t say which baseline the 2x is based on.

How to choose:
Running agents continuously for long stretches, lots of parallel sessions → B
Spreading usage evenly each day, often hitting the weekly cap → A

Asking publicly beats changing it quietly. Which side are you on?