To OpenAI's ChatGPT, Codex and subscription-policy teams:

On November 1, 2026, OpenAI emailed subscribers about a combined Chat + Work experience and announced that ChatGPT and Codex usage would soon draw from a single shared allowance. Official guidance points to a December 2026 transition for Plus, Pro and Business users. Instant chat will be exempt, but advanced models and complex tasks will consume the shared allowance.

We welcome a better interface. We oppose making two essential workflows compete for the same limited allocation without a clear guarantee of equivalent practical access.

People use higher-reasoning ChatGPT for sustained learning, research, cybersecurity investigations, architectural design, accessibility and detailed problem-solving. They use Codex and Work to implement, test and automate tasks. A costly agent run must not unexpectedly exhaust the ability to continue a High-reasoning conversation, or vice versa. Unlimited Instant is useful, but it does not replace advanced reasoning for these tasks.

OpenAI has said most subscribers should not need to wait more often. However, exact new allowances and transparent, model-specific before-and-after comparisons have not been provided. Therefore we are not claiming a reduction has already been proven: we are asking OpenAI to prevent one and demonstrate it with evidence.

We ask OpenAI to:

1. Keep advanced ChatGPT reasoning and Codex/Work usage allowances separate, or offer a protected, non-transferable advanced-chat allocation so one activity cannot consume the other's capacity.
2. Publish exact per-plan usage allocations and costs (tokens, models, reasoning efforts, agent tasks, reset windows and any weekly limits), including realistic comparisons to today's subscriptions before the change goes live.
3. Protect existing subscribers with an optional transition or an equivalent-capacity guarantee at the same subscription price.
4. Offer transparent per-task usage accounting and clear warnings before agent runs can deplete significant amounts of shared capacity.
5. Invite meaningful public feedback from both occasional and heavy subscribers and respond to it before finalizing the policy.

Subscription value should be measured by dependable completed work and meaningful access to advanced reasoning, not merely by a promise of unlimited Instant messages. Unifying the product experience should not make existing paid workflows less predictable.

Sign if you want OpenAI to preserve user choice, transparency and the practical value of advanced ChatGPT reasoning. Share this with other subscribers who rely on both conversational analysis and Codex implementation.

Official announcement details: Understanding ChatGPT and Codex usage limits.