Independent public-interest evaluations of how systems respect dignity and support agency.
HXR Evaluation
ChatGPT Capability Reduction Should Explain the Limit and Recognize Existing Pro Subscribers
How does this compare with your experience?
Good, difficult, mixed, and different experiences all matter. Share a non-sensitive summary through the manual contact route; evidence and permissions are reviewed separately.
Nothing is published automatically or sent to OpenAI. The private contribution steps and limits are explained below.
One direct screenshot and anonymous contributor testimony support a bounded finding: ChatGPT disclosed a temporary capability reduction, reset time, and possible lower quality, but did not identify the affected capability or safe limit category and paired the notice with a generic Get Pro CTA that was incongruent with the contributor’s reported current Pro subscription. The artifact does not establish the trigger, the system’s entitlement state, the CTA destination, or prevalence.
- Evaluation ID
- HXR-EVAL-0026
- Status
- Open — gathering evidence
- HXR Incident connections
- 1
- Last updated
- August 25, 2026
Current HXR status
Current work and participation
- Evaluation work
- Open — gathering evidence
- Contribution intake
- Pilot
- Reform and verification
- Reform Proposed
Evidence, independent corroboration, outreach, organization response, implementation, and verification are tracked separately. An open contribution doorway does not imply any later state.
What would help now: Have you seen a ChatGPT capability-reduction notice? Record your current plan, exact notice text, affected model or tool if shown, reset time, platform, and where any upgrade button leads—without sharing private conversation content.
Who and what this Evaluation covers
Naming an organization, product, service, or workflow identifies the subject of this Evaluation. It does not by itself prove which internal team caused or controls the issue.
- Organization in scope
- OpenAI
- Product, service, or workflow
- ChatGPT capability-reduction notice and tier-aware recovery or upgrade interface
Evaluation scope
User-facing temporary capability-reduction notices, entitlement-aware labeling, affected-versus-unchanged capability disclosure, reset timing, safe limit-category explanation, actionable recovery, factual prevention guidance, plan-management or next-tier CTA logic, diagnostics, support escalation, accessibility, and cross-surface consistency. Excludes undisclosed anti-abuse rules, quota algorithms, backend root cause, internal team ownership, legal conclusions, and population prevalence.
Question being evaluated
When ChatGPT temporarily reduces a subscriber’s capabilities, does the interface accurately identify the affected capability or limit category, explain the practical effect and reset conditions, recognize the user’s current paid plan, and offer a relevant recovery or plan-management action without implying that an existing subscriber lacks the plan already purchased?
Current findings
What works well or deserves recognition
The interface states a specific reset time and warns that responses may have lower quality. The message indicates a temporary rather than permanent reduction.
Difficulties and opportunities to improve
The notice does not say which capability changed, why the user is seeing the restriction at a safe actionable category level, what remains available, or whether any action can help. Get Pro is plan-incongruent for a reported existing Pro subscriber and can create doubt about account recognition and paid-plan value. The result shifts diagnosis and decision burden to the user.
Mixed findings and limits
A product need not disclose anti-abuse secrets or sensitive internal thresholds to provide an actionable category, practical effect, reset condition, and relevant recovery step. The screenshot supports a plan-incongruent appearance but does not prove that the underlying entitlement was unrecognized or that the CTA would permit a duplicate purchase.
Organization information and response status
No OpenAI response or verified receipt recorded.
How HXR would test performance
For a Pro account, the interface never displays generic Get Pro. It states the current plan, names the limit category and affected capability, identifies what remains available, shows reset time/countdown, offers a relevant action, and supplies a diagnostic identifier and feedback route. Any next-tier offer names the tier and explains why it resolves this exact gate; otherwise no upsell appears. The experience is consistent across iOS, web, and desktop and understandable to screen-reader users.
Evidence coverage and limits
Current evidence is one contributor-supplied screenshot plus first-person testimony from the same contributor. Screenshot facts and testimony are separate atomic records, and their shared-source dependency is disclosed. Additional occurrences, counterexamples, official explanation, plan-state verification, and cross-platform tests are needed.
Connected HXR Incident reports
1 public HXR Incident report is connected to this Evaluation. Canonical roles: 1 supports; 0 partially supports; 0 narrows; 0 context only; 0 challenges. Connections are not votes and do not establish prevalence.
Next observation or verification
Obtain OpenAI’s safe classification of the gate, confirm expected Pro entitlements and CTA destination, capture recurrence with plan state and app version, test iOS/web/desktop consistency, and verify any revised notice against the acceptance test.
Revision history — Public Revision 1
Initial public projection from canonical HXR-EVAL-0026. The public payload separates direct screenshot facts, anonymous contributor testimony, unresolved internal questions, requested reform, and acceptance criteria.
Technical record details
- Evaluation type
- Interface
- Benchmark posture
- Needs More Evidence
- Blue Score readiness
- Developmental Data Only
- Reform / positive practice
- Reform Proposed
- Public revision
- 1