This was a routine global reset: the team proactively wiped and refilled Codex usage for all paid subscriptions (Plus / Pro / Business). It applies automatically once announced — nothing to claim.
The announcement: “We are reseting usage for all paid users of Codex and ChatGPT Work. Please conti…”
Reading the announcement: the 390-word notice follows the classic incident structure — status, fix, compensation — a template that hardened after the March 2026 cluster.
Quick factsTHE SNAPSHOT
| Item | Data |
|---|---|
| Announced | 2026-08-29 20:43 UTC (28:43 Beijing, Saturday) |
| Type | Global Usage Reset (reset_type: regular) |
| Sequence | #48 of 53 tracked resets |
| Gap before | 2 days |
| Gap after | 1 days (next reset: 2026-08-31) |
| Running average at that time | 7.4 days (mean of the prior 46 gaps) |
| Where this gap sits in history | longer than 17% of historical gaps (ranked #28 longest among the gaps recorded up to then) |
| Source | public announcement by @thsottiaux ↗, archived by codex-resets.com |
TimelineWHAT HAPPENED
-
2026-08-29 20:43 UTCThe official global-reset announcement went out.
-
Original announcement
“We are reseting usage for all paid users of Codex and ChatGPT Work. Please continue reading for an update on Codex usage limits. The team has been working around the clock, going through thousands of reports and shipping fixes. Depending on how you use Codex, you should see your usage go between 10% and 50% further than before. We really went with a fine comb, with many uncovered small things being longstanding and here is what we found and fixed: - Compaction. We were keeping old images during compaction, sometimes making the context large enough to trigger compaction again. After the fix, usage dropped around 10% for users making heavy use of images. Fixed. - Memory. Background memory workers could inherit Stop hooks and keep running when the hook wouldn’t let them finish. This affected fewer than 1% of users, with the long tail being pretty bad and we saw one example thread check whether it could stop 15,000 times. Fixed. - Goals. In some cases, a set /goal could finish and then keep going past the intended stop condition, or the model would keep retrying broken tools without stopping. We saw examples consume anywhere from 15% to 70% of a weekly allowance. Fixed. - Automations. Some custom schedules could run more frequently than configured. Fixed. - Subagents. Smaller models (e.g. Luna) sometimes picked more capable helpers without being explicitly asked. The same was true where the orchestrating model not running in /fast mode could request sub-agents to run /fast. Fixed. - Computer History. The older implementation could lead to repeatedly summarizing overlapping activity. For some cases we saw it consume up to one fifth of the weekly usage per week. Fixed. - Rolling task summaries. Ordinary turns were triggering extra background requests. These added about 1% to token usage. Small each time, but it adds up. We have disabled this. - MCP. Some tool results could be encoded twice. We also found tool instructions getting cut off and fetched again. Fixed. We’ve also made architectural changes to prevent these from regressing and our teams will get paged if it happens regardless. We are also working on showing you directly in the app where your usage goes so you don’t have to guess. Goes without saying that we’re resetting usage limits and I hope you enjoy a very nice Saturday!”— @thsottiaux on X ↗
-
EffectThe reset propagated across all paid accounts and quotas were refilled.
The era: Milestones & banked creditsCONTEXT
The 20M and 25M milestones landed back to back, and banked resets graduated from occasional perk to the primary format: the August 21 batch, then celebration resets two days running at month's end — a clear warm-up for the Astra rollout.
In the full sequence, this is event #5 of 6 in the "Milestones & banked credits"; 1 day later, the next reset landed on schedule on 2026-08-31. Announced on a Saturday, 2 days after its predecessor — those numbers together are this event's coordinates on the timeline.
What it means for youTAKEAWAY
Routine resets historically average about 6.9 days apart, but the spread is wide — back-to-back days to two-week gaps. Watching the radar beats counting days.With a very short gap, this sits squarely in a streak — burning quota in such periods had the highest odds of being refilled in the whole archive.
FAQCOMMON QUESTIONS
What was the Codex global reset on August 29, 2026 (Saturday)?
On August 29, 2026 (Saturday), the OpenAI Codex team announced a global usage reset — reset #48 of 53 in the archive, 2 days after the previous one, part of the "Milestones & banked credits". Usage across all paid plans (Plus / Pro / Business) was wiped and refilled automatically.
Only 2 days after the previous one — are back-to-back resets normal?
Yes. The archive shows several such streaks (this event's era, the "Milestones & banked credits", alone holds 6 events), typically tied to milestones, launch weeks or incident wrap-ups. Streak periods are also when the radar's SEND-IT advice appears most often.
When was the next global reset after this one?
The archive has the exact answer: the next reset landed on 2026-08-31, 1 day later. That's the unique value of these event pages — the "what happened next" that people at the time couldn't know. For the future, use the homepage radar's projection instead.