OpenAI Codex lead Thibault “Tibo” Sottiaux has committed the team to a 28-day stretch in which it will either deliver a meaningful improvement for most Codex and Work users or issue a full usage reset.
Sottiaux posted the commitment on X, writing: “Over the next 28 days, each day we’ll either ship one thing that is a clear improvement and relevant for most codex/work users or ship a full reset.” He followed with a challenge to others: “I challenge you to do the same.”
The wording does not promise 28 resets. A reset is the alternative when the team does not ship a qualifying improvement that day. Sottiaux did not publish a formal definition of how OpenAI will determine whether an improvement meets that standard, nor did he specify which plans would qualify for a reset.
OpenAI’s developer forum identifies Sottiaux as the person leading Codex. The public pledge therefore comes directly from a senior member of the team responsible for the product rather than from an unaffiliated commentator.
The commitment also follows a separate post in which Sottiaux outlined four areas the Codex team is concentrating on: simplifying the product, improving efficiency and usable capacity, shipping major new features, and introducing new models. He said user feedback was pushing the team toward a simpler Codex experience.
The reset part of the pledge is relevant because usage limits have been a recurring issue among Codex users. OpenAI’s Codex help documentation describes both automatic and banked resets and explains that Codex and Work usage varies according to factors including the model, task size, reasoning settings and tool use.
OpenAI’s developer community has also hosted an extended discussion around Codex rate limits, with users reporting problems involving token consumption, long-running tasks, context handling and available capacity.
The team has already used resets during periods of heavier demand. In late September and early October, public announcements associated with Sottiaux included additional Codex and Work resets, while the team addressed performance and capacity issues surrounding newer models.
Those updates included changes tied to GPT-6.1 Sol, which OpenAI introduced for coding and computer-use workloads. OpenAI also reported faster token generation in Codex and improvements intended to reduce the amount of usage consumed by some workloads.
The new 28-day commitment sets a daily condition for the team: a broadly relevant product improvement must be shipped, or users receive a full reset. Sottiaux has not published a day-by-day roadmap or additional eligibility details for the resets.
Discover more from Aree Blog
Subscribe now to keep reading and get access to the full archive.
