Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

ChatGPT Pro: unexplained seven-hour account-wide message restriction across computer and mobile; root cause unresolved after support review

Open
  • Is something wrong?
#52,361 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
5/100
Issue type
Bug
Clarity
Needs clarification
Activity status
Active
Domain
backend

Research direction

The payload names no file, test, or entry point, so there is no code location to start from. It is a customer complaint about a past ChatGPT account-wide message restriction and a support refusal, and it asks maintainers to route it to a team that owns account access. A newcomer should not pick it up as a code change; the useful step is to confirm whether this tracker owns the issue or whether it should be redirected to a support channel.

Written by the indexing model from the issue text.

Description

bug rate-limits

Summary

On 1 August 2026, my ChatGPT Pro subscription (USD 200/month) was effectively unusable for approximately seven hours: I could not send messages through any available model or mode I tried, on either my computer or my phone. Existing conversations, new conversations, and switching models did not provide a working fallback.

The displayed error was "Message limit reached", with a reset time of 23:31, Europe/Amsterdam (CEST, UTC+2). Access returned after the displayed restriction period ended.

This report concerns that historical ChatGPT account-wide inability to send messages. It is distinct from my ongoing conversation-opening/loading report, #48683. I have no evidence that the two incidents share a technical cause. This was not a demonstrated Codex CLI failure.

Observed sequence and environment

  • Product: ChatGPT; subscription: Pro, USD 200/month.
  • Affected surfaces: computer and mobile.
  • Incident date: 1 August 2026.
  • Models I specifically remember trying: GPT-5.6, GPT-5.5, and o3; different modes/settings were also tried.
  • Screenshots captured during the incident show the restriction at multiple points over the afternoon/evening and the 23:31 reset time.
  • Moving between existing and new conversations, models/modes, and devices did not restore message sending.
  • I contacted OpenAI Support while the restriction was still active. The initial escalation response said a specialist would respond in the coming days, which could not restore access during the incident.
  • Historical browser/mobile app versions and a HAR recording were not captured. I cannot reconstruct them retrospectively and will not invent them.
  • This is a documented historical observation, not a deterministic reproduction in a current build.

What Support has actually established

The following is a paraphrased chronology of my private support correspondence. Personal details, case identifiers, screenshots, and raw emails are intentionally omitted.

Date, 2026 Support outcome
1 August I reported the active inability to send messages and requested urgent access restoration and investigation. Support escalated the report.
4 August AI-assisted Support discussed compensation as a refund and asked me to identify the requested charge/billing period, alongside further diagnostic questions. This was not a promise of compensation.
19 August Support explicitly recognized that the reported incident affected ChatGPT across computer and mobile and was not a Codex App issue. It acknowledged that switching chats or modes had not enabled continued work.
8 September Support said the evidence established the temporary restriction and did not establish an ordinary allowance affecting one individual model as the explanation for all the alternatives I had tried. It could not identify the exact backend trigger, a verified corrective action tied to that trigger, or a basis to guarantee nonrecurrence. It acknowledged that reasonable customer-side investigation had been completed and that additional historical diagnostics were not required.
10 September Support confirmed the approximately seven-hour restriction, the unresolved backend trigger, and a final refusal of complimentary Pro months, service credits, waived charges, account credit, or equivalent compensation. It stated that an authorized Support decision-maker had reviewed the request and that no further review route was available through Support.
26 September After my follow-up questioning the scope of that review, AI-assisted Support said it could not identify what the decision-maker had reviewed or the person's role/team, and asked again what remedy I wanted while keeping Pro active. This did not approve a remedy or answer the outstanding technical question.

There are 28 messages in the correspondence chain I have available through 26 September, including my replies and Support's responses. That count describes this correspondence only; it is not a count of affected users.

Earlier replies offered a refund eligibility review on terms that would cancel the subscription if a refund were processed. That did not meet my request to retain the same active Pro subscription. I have not authorized cancellation or downgrade.

Unresolved technical problem

Restoration after the displayed restriction expired is not a root-cause finding. The support review still has not established:

  1. Which allowance, safeguard, account state, counter, or backend condition prevented message sending.
  2. Why every alternative model/mode I tried was simultaneously unusable across devices.
  3. Whether this was intended behavior or a product/account error.
  4. Whether an applicable preventive action or a supported recovery/fallback path exists.

I am not asserting that a specific quota, abuse safeguard, security restriction, or software defect has been proven. The absence of an established explanation is precisely the unresolved issue.

Expected behavior and impact

A paid account should receive an actionable explanation when message sending is restricted, including the scope of the restriction, the applicable reset time, and any available alternative. If an account-wide restriction is unintended, there should be an engineering investigation and a supported recovery path.

The incident blocked planned work for hours. The subsequent correspondence required substantial additional time, including repeated diagnostic questions, without identifying the cause or providing compensation while retaining the subscription. Acknowledging the restriction and later restored access does not address that remaining technical uncertainty or the service complaint.

Requested action and routing

Please identify the engineering owner and provide a substantive assessment of the account-wide restriction, or link the appropriate tracking item and route this report to the team responsible for ChatGPT account access. If this tracker cannot own that part of the product, please identify the appropriate reporting/escalation route rather than treating the historical restoration as a verified fix.

The support/compensation handling is included as relevant customer-impact context and a separate unresolved complaint; I am not asking a code maintainer to process a billing transaction. My requested remedy remains service credit or complimentary Pro time with my existing subscription active and unchanged. Please route that question to someone authorized to address it without restarting the same diagnostic questionnaire.

I ask that the technical report remain tracked as unresolved until there is a substantive technical outcome or a clearly identified owner/tracking destination. I understand that maintainers control issue triage.

Private incident evidence and support correspondence can be correlated through an appropriate private support channel. No private conversation contents or account identifiers are attached publicly.

Dominant language
Rust
Stars
129k
Forks
20.3k
Avg merge
1m
Merged PRs (30d)
998

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from openai/codex

All issues in openai/codex

Similar issues

More Rust issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.