Back to blog
3 min read

Codex Reset for 25 Million Users: What Is Available Again

The global Codex reset of August 31, 2026 explained: who it covered, which limits matter, and how to use the renewed capacity without guessing.

  • #Codex
  • #OpenAI
  • #Usage Limits
  • #AI Agents
Miniature globe with a reset ring and connected network nodes.
Share

Short answer

On August 31, 2026, OpenAI reset usage for all paid ChatGPT Work and Codex subscriptions. The public announcement tied the reset to 25 million active users. This was a one-off global hard reset, not a new fixed schedule or API credit. Check your own Codex usage view before planning around the renewed capacity.

From reset signal to a safe work block

Four steps separate the public announcement from the capacity you can actually use.

  1. Confirm the signal

    Check the primary post, date, audience, and at least one independent counter-source.

  2. Check your account

    Read the five-hour and weekly state in the real Codex usage view.

  3. Budget the block

    Choose one outcome, limit context, and reserve capacity for tests.

  4. Preserve the result

    Verify the diff, tests, and live behavior, then record the next step.

On August 31, 2026, another public Codex reset arrived: usage was reset for every paid ChatGPT Work and Codex subscription, according to the announcement. This time the reason was not an outage but a milestone of 25 million active users. People who had nearly exhausted a weekly allowance potentially regained room for coding tasks, reviews, and cloud runs.

The headline is simple; the operational meaning is not. A global post cannot tell you the percentage visible in your account at this moment, and one act of goodwill does not create a new subscription promise. This guide separates the confirmed event from what you still need to verify.

What was confirmed on August 31

The current primary post came from Tibo Sottiaux. It connects two statements: ChatGPT Work and Codex reached 25 million active users, and usage had been reset for all paid subscriptions to celebrate. The wording describes the reset as completed rather than merely forecast.

Creaiter added the event only after two independent trackers linked the same current X post. That check matters because a reset had already been promised for 6 p.m. Pacific the previous day, while an earlier near-duplicate post was later corrected. A single screenshot could easily mix those records.

What a global hard reset means

A hard reset refreshes affected usage automatically. It is different from a banked reset, which eligible users redeem later. It is also not credit for the OpenAI API. API billing, rate limits, and spending controls remain a separate contract.

The global event does not replace your regular account counters. Codex may show a short usage window and an additional weekly budget. Only your usage dashboard or Codex status view can reveal which values changed for your plan. Our evergreen guide to Claude and Codex limit resets explains those categories in more detail.

How to check whether your account received it

Open the usage view for the exact account and organization where you work. Record the weekly state, short window, and next personal reset time. Refresh the page if it was already open during the announcement. Then verify that you are looking at ChatGPT Work or Codex—not API billing or another ChatGPT limit.

If a counter does not jump to 100 percent, that is not immediate proof of failure. Rollout can become visible at different times, and surfaces may update differently. The public statement describes the intended audience; your dashboard describes your account. If it stays unchanged, record the plan, organization, time, and exact warning before contacting support.

How to spend the renewed capacity well

A reset tempts you to start the largest task on your backlog. A bounded work block is safer: one feature, one bug fix, or one release check. Define the outcome, affected files, and acceptance criteria before the run. A good first prompt saves usage by reducing detours, not merely by being short.

Reserve capacity for verification. An agent needs room to read, test, inspect, and correct—not only to write. Small commits or clear intermediate checkpoints make long sessions recoverable. When another limit appears, you know what is finished and where to continue.

Checklist for your next Codex block

  • Account, organization, and usage view are verified.
  • The desired outcome is bounded to one testable block.
  • Relevant files and acceptance criteria are in the request.
  • Capacity remains for tests, browser review, and corrections.
  • The diff, test result, and next step are saved before the session ends.

Why you should not calculate the next date

Recent global resets sometimes landed close together. That still does not create a contract. Milestones, incidents, launches, and goodwill decisions have different causes. An average or median only describes previous events.

Use the Codex reset history to classify a signal, not to bet a release on it. Until another explicit source appears, the safe assumption is: no extra global reset is coming, and your personal usage timer remains authoritative. That turns the 25M reset into helpful extra capacity instead of the foundation of a risky delivery plan.

Mini quiz

Can you classify the 25M reset?

Choose the answer that separates public evidence from your account state.

1 / 3

What scope did the public announcement name?
Show solutions
  1. 1. What scope did the public announcement name?

    Correct answer: Paid ChatGPT Work and Codex subscriptions

    The announcement names paid ChatGPT Work and Codex subscriptions. API billing and personal windows are separate.

  2. 2. What should you check before planning around new capacity?

    Correct answer: The usage view in your own account

    A global post proves the announced event, not the exact state of one account.

  3. 3. How do you determine the next global reset?

    Correct answer: Wait for a new explicit source

    Historical gaps provide context, not a promised date.

Sources

  1. 25M active users and usage resetTibo Sottiaux · accessed 2026-08-31
  2. Codex Reset Timeline APICodex Reset · accessed 2026-08-31
  3. Codex reset status todayCodex Limit Watch · accessed 2026-08-31
  4. Using Codex with your ChatGPT planOpenAI Help Center · accessed 2026-08-31

Frequently asked questions

Was there really a Codex reset on August 31, 2026?

Yes. Tibo Sottiaux publicly wrote that usage had been reset for all paid ChatGPT Work and Codex subscriptions after the products reached 25 million active users. Two independent ledgers link to the same current primary post.

Who was supposed to receive the reset?

The announcement names all paid subscriptions for ChatGPT Work and Codex. It does not prove the state of any individual account, so verify rollout, organization, and plan in Codex itself.

Did it reset both weekly and five-hour limits?

The public wording says usage was reset. Only your account can show which counters changed for your plan. Do not confuse this global event with the regular schedule of your personal windows.

Was this a banked reset or API credit?

No. The announcement describes an executed reset, not a voucher saved for later. API billing and API rate limits remain separate.

When will the next global Codex reset happen?

There is no verified fixed cadence. Previous events are history, not a calendar. Wait for a new explicit source and plan against your account usage display until then.

Privacy setting

With your consent, PostHog EU measures which pages are opened and additionally records your session: mouse movement, clicks, scrolling and the rendered page content are stored as a replayable reconstruction. Typed input is masked before sending. The contents of sign-in, sign-up, password recovery, the founder chat, the newsletter field, your email display, your quiz answers and the checklists are not recorded at all; on the sign-in and sign-up pages recording is paused. Addresses are always stored without query parameters. A random device identifier and technical connection data such as the IP address are processed as well. In addition, strictly defined usage events are measured: how far a post was read, which article element was used, which call to action was clicked, how a newsletter sign-up ended, how far you get in a course or in the feed (recording only whether a task was correct or incorrect, never your answer itself), which video you start, whether you switch the language, and how a sign-in or sign-up attempt ended — without the email address, without the password, and without the error message. Only values from a fixed list and whole numbers from fixed ranges are transmitted — no input and no free text. We also label production, internal, and automated test visits separately and derive whether a visit came from a known interface such as ChatGPT, Claude, or Perplexity from a known referring domain or a strictly allowed campaign value. The full referring address and query parameters are not sent, and this cannot identify a specific AI model. If PostHog's IP-based geo enrichment is enabled in the project, the service can derive an approximate country, continent, and region; we do not request browser or GPS location. Every visit is also assigned to a page area from a fixed list (for example home, blog, tools, account) — the area is sent, not the address. Page loading and stability metrics are measured as well (Core Web Vitals: LCP, CLS, INP, FCP), without network payloads. A click on a link leading away from the site is recorded only as the target domain from a fixed list, without path, query parameters, or link text. For the Idea Wheel, AI Labelling and Limit Reset tools only the kind of action is measured — spin, option changed, copy or export, for instance — never your input and never a result. Learn more about privacy

Without your consent no analytics code is loaded and nothing is recorded. You can withdraw at any time: withdrawal ends collection immediately; data already collected may have been transmitted by then and is deleted after the storage period.