📅 October 31, 2026 • OpenAI
OpenAI published a new Help Center page on October 31 describing how ChatGPT and Codex usage limits work together. Starting in December, the usage allowance on Plus, Pro and Business plans becomes shared across ChatGPT and Codex, as the company brings the agentic capabilities of Work into the main ChatGPT experience. The update was independently covered the same day by TokenPost, AICoder and other outlets that track OpenAI billing changes.
Usage will be measured in tokens rather than messages — the units of information a model processes and generates — so how much of the allowance a task consumes depends on the chosen model and the complexity of the task. Advanced models and complex tasks generally use more. Instant chat is excluded and will not draw from the shared pool. Combined remaining usage and reset times are shown in Settings → Usage, and users who reach the limit can wait for a reset, upgrade the plan or buy credits to continue. OpenAI says it expects most users will not need to wait for resets more often than they do now, and that it is preparing carefully for the transition to prevent disruption to workflows.
The change formalizes a convergence that has been building for months. Under the current setup, Work and Codex already draw on one shared allowance — the December change extends that pool to ChatGPT usage itself. The Business pricing page already describes a shared limit across ChatGPT and Codex starting in December, and checkout text stating the December start was spotted on October 29. Thibault Sottiaux, who heads ChatGPT and Codex, previously confirmed the plan to converge chat and work experiences by end of year. The new page does not publish per-plan token amounts, so the practical size of each pool is not yet known.
How it works in practice: automation builders on Plus, Pro or Business should budget one shared token pool across chat and coding agents. A heavy Codex session now reduces what is left for model-heavy chat work in the same window, and vice versa — while instant chat stays unmetered. The practical routine is concrete — check the combined balance in Settings → Usage before launching long agent jobs, align batch runs with the reset time shown there, and keep credits as the overflow buffer for runs that would exhaust the pool. Teams on Business plans should also ask admins whether per-seat usage visibility will exist in the new Usage screen, since one person's agent-heavy week can consume allowance that used to be effectively earmarked for the whole team's chat.
Sources:
- help.openai.com
- openai.com
- tokenpost.com
- aicoder.com
Continue the conversation: https://t.me/AIAutomationArchitecture · Bluesky
Top comments (0)