An open letter on the $200 ChatGPT Pro subscription
Trust needs clear terms.
I value ChatGPT and Codex. I stopped renewing because a lower recurring allowance at the same price, together with inconsistent explanations, undermined my confidence in the subscription.
The original three requests
Publish a verifiable comparison.
Specify the affected products, previous and revised recurring allowances, effective date, shared pools and reset rules. Explain what customers can actually verify.
Offer a meaningful continuing arrangement.
Preserve the previous recurring allowance for continuously subscribed customers, or offer a meaningful ongoing alternative such as recurring credits or a price adjustment.
Address the unresolved policy concern.
Acknowledge the difference between explaining a change and resolving the objection to it. Provide a response from the responsible team, including treatment of paid billing periods that cross the transition.
What the correspondence says
These are redacted summaries of correspondence I received. They describe the explanations given in my case; they are not independent measurements of usage or a guarantee of policy for every account. Original messages are retained privately.
30 Sep 2026
Transition and temporary credits
Support confirmed that my account was eligible to retain its previous allowance through 29 October while the subscription stayed active. It also confirmed a one-time grant of 62,500 credits expiring on 31 December. I acknowledge this temporary benefit.
4 Oct · 08:15 UTC
No verified numerical comparison
Support said it could not verify an exact 20×-to-10× comparison from the published guidance and recorded my objection to the lower recurring allowance at the unchanged price.
4 Oct · 10:15 UTC
A conflicting explanation
A later reply described the change as 20× to 10× and referred to five-hour and weekly windows. This was an explanation in the correspondence, not a figure I independently verified.
4 Oct · 14:10 UTC
The precise claim was withdrawn
Support then said the earlier exact multiplier explanation went beyond published guidance and could not be verified. It clarified that Work and Codex share an allowance, ordinary Chat limits are separate, and Pro $200 did not have a separate five-hour Work/Codex cap. No policy change or response timeline was promised.
21 Oct 2026
Cancellation statement and a further reply
I publicly stated that I had cancelled renewal. A mailbox check also found a reply dated 21 October, labelled AI-assisted, repeating the lower-allowance transition at the same $200 monthly price. It asked about confirmation of the App Store cancellation. This page does not treat that email as proof of my billing status.
The original letter (21 October 2026)
Dear OpenAI Support and the team responsible for subscription pricing and product policy,
I have cancelled renewal of my $200-per-month ChatGPT Pro subscription. This is my statement of that decision, not a request to delete my account or data.
I relied heavily on ChatGPT and Codex for software development, studying and research. As a student and developer, $200 per month was a substantial financial commitment. I was willing to make it because these tools were valuable to my work.
My concern is the announced reduction in the recurring Work/Codex allowance at the same monthly price. In the correspondence I received, Support confirmed a transition through 29 October 2026 for eligible existing subscribers who remained subscribed, followed by a lower included allowance. I am not claiming that I independently measured an early reduction on my own account.
Support also confirmed a one-time grant of 62,500 credits expiring on 31 December. I acknowledge that benefit. My objection concerns the ongoing subscription after temporary benefits end: I need to know what recurring value I can reasonably plan around.
I also need to correct my original public post. Support has replied to me. The record includes replies on 30 September and 4 October and an AI-assisted reply on 21 October. It would be inaccurate to describe the entire correspondence as unanswered.
Those replies have not resolved my policy objection. On 4 October, an explanation describing an exact 20×-to-10× change was subsequently withdrawn as going beyond the published guidance. I therefore do not present an exact 50% reduction, or a fixed loss of completed tasks, as an independently established fact. The inconsistency itself is part of why a clear and verifiable comparison matters.
Please publish a product-specific comparison of the previous and revised recurring allowances, the effective date, shared pools and reset rules, and explain how subscribers can verify their allocation and consumption.
Please preserve the previous recurring allowance for continuously subscribed customers, or provide a meaningful continuing alternative such as recurring credits or a price adjustment. Additional features, model-efficiency improvements and expiring credits do not by themselves establish that a lower allowance preserves the value of my particular workload.
Please address the treatment of already-paid billing periods that cross the transition, and provide a substantive response from the team responsible for the policy. I recognise that frontline Support cannot promise a company-wide policy change. Recording feedback and resolving the underlying objection are different outcomes.
I would have preferred to continue renewing. I still value the tools. My decision reflects a loss of confidence in predictable subscription value and in the consistency of the explanations I received.
If OpenAI provides further clarification or a meaningful continuing arrangement, I will update this page and distinguish that development from the earlier record.
For clarity, I am not requesting deletion of my account or any of my data.
Sincerely,
Omica Chow
Share a documented experience
If you have a related notice or experience, you can email me a redacted account with the date and affected plan. Remove account identifiers and payment details, and say whether I may quote it publicly. This page does not collect signatures or claim to represent other subscribers.
Updates
30 October 2026 — Marked the letter closed following Tibo’s 29 October allowance announcement.
21 October 2026 — Published this bilingual record after checking the correspondence. Corrected the broad “no response” statement and distinguished the withdrawn multiplier claim from the unresolved policy concern.