What happened
The Decoder reported on November 1, 2026 that OpenAI is applying a device limit to paid ChatGPT accounts. When a user reaches the limit, ChatGPT displays “Device limit reached,” and the affected account can no longer be used on that device. This is more consequential than a failed attempt to buy another plan: under the report’s description, an existing paid account can lose access on that particular device after the limit is reached.
The report does not state the number of accounts allowed. It describes an OpenAI support page but the supplied material does not reproduce that page in full, so it does not establish the exact language of the rule, how long it has been in force, or whether the threshold is identical for every user. The visible operational detail is that logging out of an account does not automatically free a place for another account. In other words, repeatedly signing in and out is not described as a reliable way to reset the device’s capacity.
The policy covers ChatGPT Go, Plus, Pro 100, Pro 200, and Pro 500. Free accounts do not count toward the limit. Business, Enterprise, and Edu plans are also excluded, according to the report. Those distinctions matter because the reported restriction is not framed as a general maximum for every ChatGPT identity on a computer or phone; it applies to a defined group of consumer paid subscriptions.
The same report describes related card and device restrictions. If too many subscriptions are associated with one payment card or device, a user may be unable to start a new subscription. The device cap adds a separate reported consequence: the account can no longer be used on that device once the limit is reached. The supplied source gives no numerical threshold for either the card rule or the device rule, and it does not say whether a card restriction and a device restriction are calculated in the same way.
The Decoder says the policy is part of OpenAI’s security measures. It may particularly affect users who have switched among multiple paid accounts on one device to obtain more usage from flat-rate subscriptions. The report says OpenAI points those users toward additional credits or usage resets instead. It characterizes those alternatives as generally more expensive than another subscription, but supplies neither prices nor worked comparisons. The material therefore supports a clear access change, not a quantified assessment of its cost.
What the archive adds
This is a new fact on a known account-access story. OpenAI help page sets device limits on paid ChatGPT accounts, affecting stacked Codex accounts covered card and device restrictions; the current report adds that reaching a device cap can prevent use of an affected account and identifies the plans covered. A related account-switching note described keeping two accounts signed in, a separate feature rather than this cap.
Trend signal
The report does not materially support a listed SITE_PATTERNS thesis: it concerns subscription and device access rather than agent architecture, capabilities, or runtime controls. No pattern link is warranted.
Why now
The report appeared November 1, 2026, after the earlier coverage of OpenAI’s help-page limits on paid accounts. It presents the device restriction as part of security measures, but the source does not explain the rollout schedule, identify a triggering abuse pattern, or say why OpenAI is applying the threshold now. The timing is therefore clear only at the level of the report and its relationship to the earlier documented restrictions.
Connections to other events
The account-switching feature described in the earlier Field Note allowed two accounts to remain signed in and be switched between on web, iOS, and Android. The reported device cap concerns a different question: whether paid accounts may be used on the device at all. Together, the two changes make a distinction between convenient switching and permitted multi-account access.
Why it matters
Teams and individual builders sometimes maintain more than one account for separate usage needs, testing, or higher available capacity. A device-level rule means access is not governed only by the subscription attached to each login: the device itself can become a constraint, and logging out may not restore capacity. That makes it risky to assume that adding another paid plan will reliably increase usage on a shared workstation.
The plan distinctions matter for procurement and account administration. The report says consumer paid tiers are in scope while Free and organizational Business, Enterprise, and Edu accounts are excluded. A team should not treat that as a recommendation to shift accounts between plans; the source does not state eligibility rules, migration paths, or whether organizational plans solve an individual’s access case. But it does indicate that a workflow built around rotating personal paid accounts has a newly reported operational dependency.
For builders using ChatGPT or related paid access as part of daily work, the immediate task is to identify whether critical work depends on several consumer subscriptions sharing one device. If it does, the fallback should be an authorized access arrangement and a tested usage plan, rather than an assumption that account rotation will continue to work. Earlier coverage: OpenAI help page sets device limits on paid ChatGPT accounts, affecting stacked Codex accounts.
What happens next
- OpenAI will clarify the per-device threshold or the recovery path after a device block. (3 months, 55%). Watch for: an updated support-page explanation or an in-product account-management notice.
- Users who relied on stacked consumer subscriptions will shift toward credits, resets, or organizational access where available. (6 months, 65%). Watch for: new purchase flows or plan guidance that explicitly addresses additional usage.
Limits and uncertainty
The report does not establish the numerical device cap, whether the threshold differs by plan, how OpenAI counts devices or accounts, or how users can appeal an incorrect block. It also does not provide the full support-page text in the supplied excerpt, independent user testing, or a rollout timetable. The description of credits and usage resets lacks prices, while the claim that these are generally more expensive is not quantified.
The source attributes the policy to security measures but does not provide OpenAI’s detailed rationale. It also does not establish whether the restriction applies uniformly across operating systems, regions, devices that are shared by several people, or accounts that have previously been used on the same hardware. Those missing details limit what teams can infer from a single reported enforcement message.
Skill to build now
Audit any workflow that relies on multiple consumer ChatGPT subscriptions from a single workstation. List the accounts, their plan types, the device they use, and the work that would stop if one account showed “Device limit reached.” Then test and document an approved fallback for the affected work, such as a separately authorized organizational account or a defined additional-usage purchase path. The goal is not to evade the limit; it is to remove an untested access dependency from a production routine.
Reader takeaway
Treat a paid ChatGPT subscription as subject to a device-access policy, not as a login that can necessarily be rotated freely. Several consumer plans are reportedly covered, and signing out may not restore a slot. The exact cap remains unknown, so multiple personal subscriptions on one device are no longer a predictable way to expand usage.
Open question
What is the per-device threshold, and can users see or contest which accounts count toward it?