Tibo Sottiaux
@thsottiaux
policy change
Tibo explained how an immediate quota reset differs from a reset saved in a user account for later redemption.
Track the next global Codex quota reset with live X signals, verified evidence, and a sourced reset history.
Latest confirmed Codex reset: Jul 16, 04:14 AM UTC
Tibo announced another reset for Codex and ChatGPT Work users, with weekly usage returning to 100% within minutes. No newer confirmed global reset appears in the durable ledger.
View original source →7%
Reset within 24h
Evidence adjustment0 pts
19%
Reset within 48h
Historical interval+6 pts
Public signals and confirmed global quota resets, linked to their evidence.
Durable D1 evidence ledger · 12 assessed posts with linked original sources.
@thsottiaux · Jul 16, 2026
Tibo asked users how they wanted to manage the existing weekly allowance while the five-hour limit remained paused.
Classifier reasoning
This changes quota-policy context but does not promise another reset.
@thsottiaux · Jul 16, 2026
Tibo confirmed a new Codex and ChatGPT Work reset after reaching 9M active users and said weekly usage should return to 100% within minutes.
@thsottiaux · Jul 15, 2026
The public question connected the approaching 9M milestone with another possible Codex and ChatGPT Work reset.
@thsottiaux · Jul 15, 2026
Tibo shared a credits promotion. It adds usage capacity but is not a global quota reset.
@thsottiaux · Jul 15, 2026
A follow-up clarified that the referenced action should not be treated as a new global reset.
@thsottiaux · Jul 14, 2026
Tibo announced that Codex and ChatGPT Work had reached 8M active users and that usage limits were being reset for all.
TikHub polls a small, approved set of high-signal X accounts every hour.
A structured LLM assessment reads each newly stored post and separates signals, banked credits, negative evidence, and explicit global resets.
An empirical Bayesian model recalculates reproducible 24h and 48h scores from historical outcomes and active evidence.
D1 stores every source post, model assessment, confirmed event, forecast input, and ingestion run.
A Codex reset is a global or plan-scoped refresh of usage limits. It is different from a user’s normal rolling quota window.
This monitor focuses on resets announced or observed across a broad group of Codex users, not an individual account’s routine recovery.
There is no guaranteed public schedule. The 24h and 48h scores are experimental estimates derived from stored reset frequency and the historical hit rate of active signal categories.
The model starts from the observed confirmed-reset rate, adjusts it using each active signal category’s historical 24h and 48h outcomes, shrinks small samples, and decays old signals. It is called a probability only after at least 30 resolved forecasts pass the published Brier-score gate.
No. Codex Reset Monitor is an independent project that organizes public evidence and links back to every source.
Independent monitor. Not affiliated with OpenAI. Experimental forecast scores are estimates, not guarantees.
Know when a global Codex reset is confirmed—or when the experimental forecast score rises sharply.
Your address is stored securely in our waitlist. Automated reset emails are not live yet.