Developers are secretly hoping OpenAI fails to ship this month
The New Stack Amanda Caswell
OpenAI kicked off a 28-day Codex sprint with a speed boost. Developers want the promised resets, but that means hoping OpenAI misses a ship.
Based on reporting by The New Stack, Amanda Caswell — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI’s Codex sprint has already produced its first visible change, and it’s not the free usage reset some subscribers were hoping for. On day one of Thibault Sottiaux’s 28-day push, the company said GPT-6 Astra and GPT-6.1 Sol were made about 50% faster by default. Sottiaux later put that at roughly 50 tokens per second, up from about 30.
That upgrade matters beyond OpenAI’s own tools. People using Sign in with ChatGPT can pick it up through OpenCode, Pi, Amp and Devin too. And while the exact numbers sound a little squishy — the source itself treats them as rough estimates rather than formal benchmarks — the move still sets the tone for the month: speed first, reset later, if at all.
The timing has a bit of a sting in it. Two days before the sprint began, Sottiaux had already apologized for a slow launch of GPT-6.1 Sol after a major load spike in its first two days, then announced a global reset for paid ChatGPT accounts to make up for that. So the first “ship” in the challenge ended up being a fix to the same model that had just stumbled.
This is also where OpenAI’s new ritual gets awkward. Sottiaux has been training users to treat usage like a kind of currency, with resets showing up after April rate-limit changes, July regressions, and even on stage at DevDay on September 29. Now people are gaming the calendar, hoping OpenAI misses a day so they get their allowance back. That’s easier to do when limits are already annoying: one Plus subscriber said a five-hour cap shapes when they can work.
The company also picked a bad moment for restraint. Last week it said that starting October 30, it will cut the Codex and ChatGPT Work allowance on its $200 Pro plan from 20 times the Plus allowance to 10 times. At the same time, it introduced a $500 plan with 25 times the Plus allowance and an Ultrafast tier for GPT-6 Astra that promises up to eight times standard Codex speed. So yes, OpenAI is shipping faster Codex. It’s just doing it while asking developers to applaud with one hand and count resets with the other.
My take — AI-written commentary, not fact-checked reporting
OpenAI has turned usage into a suspense mechanic, which is a very Silicon Valley way to make billing feel like a sport. Developers aren’t really cheering for failure; they’re reacting to a company that keeps making them track limits like a scoreboard. If your product strategy depends on people hoping you miss deadlines, the product may not be the only thing that needs a reset.
Read more about this at: The New Stack