Experimental historical estimate, not a promise or your account reset time. Waiting longer does not raise this memoryless model's odds.
How we calculate this
We use up to eight completed intervals between source-marked confirmed reset records, deduplicate by UTC date, and exclude banked grants, previews and future records. Rate = intervals / total days; P(t) = 1 − exp(−rate × t). Displayed estimates are capped at 46% and 71%. This is our own implementation, not the original provider's exact model or a validated prediction.
OFFICIAL SERVICE REPORT
Service status
Waiting for OpenAI Status.
Only Codex-named components actually reported by the official API are shown. Other surfaces and your personal quota are unknown.
A conservative wording rule, separate from the historical percentage. A watch score is a rule level, not a measured probability. Banked grants do not reset this model's clock.
Data: codex-reset.com. Service status: OpenAI Status. Our backend fetches and briefly caches these public sources on request. It does not operate its own X API collector or account-monitoring fleet. No request is sent with your saved personal time or usage.
The visible page checks about every two minutes; source timestamps may be older than our fetch time. If an upstream fails, we show failure or stale data, not the old archived snapshot as a live result.
Day 11/
We've landed a series of changes and more compute to very significantly decrease the occurrence of the disruptive "Selected model is at capacity. Please try a different model". When it does happen, it should also eventually just continue by itself and not just block on it.
Let us know if you still see it out there in the wild and we'll take a look.
Day 9/
We've rolled out optimizations to how we store state on disk for the chatgpt & codex desktop app that should reduce the amount stored on disk by 60%. Small ship but we've heard enough feedback about power users running out of disk space.
Perhaps I shall try to explain. There are significant differences between the 6.1 generation and prior models and therefore the optimal defaults are different between GPT-6 Astra and GPT-6.1 Sol.
And even for 6.1 Sol we wanted to be careful with the change and make sure we felt good about the results. We do now and therefore we changed it.
The results of this poll has gotten me confused. I definitely thought it would sway much more favorably in the 2X direction across Pro 100/200/500 at the cost of bringing back 5h limits in order to smoothen the demand curve. t.co/zwn7kboMhT
Day 7/
We have increased default context length for GPT-6.1 Sol to 500k when used through Codex & Work, up from 272k or an ~ 80% increase.
We will go higher in the future, but it helps reduce occurrence of compactions and with tasks that require keeping a lot of context and nuance over periods of time. It was carefully tested and is the new sweet spot for most of our users. You can still override it if you want to be higher our lower, with maximum context length of GPT-6.1 Sol of 1M.
Last 30 days. 374 different accounts since Jul 31, 2026; replies in his own threads are not counted.
About this archive
Posts and replies by @thsottiaux as codex-reset.com's monitor collected them through the X API: every post it saw from July 31, 2026; before that, only his reset posts.
Text only: images, videos, quotes and reposts are not shown, and each post links back to X. Days and the default clock are UTC; “Tibo time” is US Pacific.
The archive is refreshed with site updates; newer posts load from the live feed while this page is open. “Reset” shows the posts about a Codex reset. A post is removed here on request.
Questions
What is Tibo's X account?
@thsottiaux. Tibo is Thibault Sottiaux; his public profile on X says he works on Codex and ChatGPT at OpenAI.
Does Tibo announce Codex resets on X?
Yes. A global reset of Codex usage limits is usually announced in a post of his first. The posts the tracker counts as resets carry a “Codex reset” mark here, and the full record is on the timeline.
How often does Tibo post?
11.9 posts and replies a day over the last 30 days. The busiest day in the archive was October 6, 2026, with 43 posts.
About Tibo
Tibo is Thibault Sottiaux. His public profile on X says he works on Codex and ChatGPT at OpenAI, and he is the one who announces the global resets of Codex usage limits.
