ChatGPT and Codex to share one usage allowance in December

OpenAI says paying ChatGPT subscribers will draw on a single pool of usage across ChatGPT and Codex starting next month. In a Help Center article, the company wrote: “Starting in December, Plus, Pro, and Business plans’ usage allowance will be shared across ChatGPT and Codex as we bring the agentic capabilities of Work into the main ChatGPT experience.”

Measured in tokens: OpenAI says usage for ChatGPT and Codex “will be measured in tokens, the units of information a model processes and generates.” How much of the allowance a user spends “will depend on the model you choose and the task you ask it to do,” the article says, and “more advanced models and complex tasks generally use more of your allowance.” One carve-out: “Instant chat will not draw from your usage,” according to OpenAI.

What OpenAI promises: The company says it is “preparing carefully for this transition to help prevent disruption to your workflows,” and that it expects “most users won’t need to wait for their usage to reset more often than they do now.” Subscribers will be able to see their “combined remaining usage and reset times in Settings → Usage,” OpenAI says. If they reach the limit, the article says, they “can wait for it to reset, upgrade your plan, or buy credits to continue using eligible features.”

What’s shared today: The December change extends a setup OpenAI already uses for its agent products. “Work and Codex share the usage allowance included with your plan,” a separate Help Center page says, and that allowance “does not correspond to a fixed number of messages or tasks.” Depending on the plan, OpenAI says, limits “may apply over a five-hour window and a weekly window,” and where both apply, users need allowance remaining in both to continue. OpenAI says Pro 100, Pro 200 and Pro 500 “currently have no five-hour usage limit in Work and Codex.” The company also notes that “switching models does not restore allowance in a shared usage pool.”

Coders vs. chatters: OpenAI’s own estimates show how wide the spread is. For Plus and Standard Business, it estimates 5–45 local messages within a five-hour window on GPT-6 Astra, versus 350–3,000 on GPT-6 Luna, and calls those ranges “estimates, not guaranteed message counts.” Cloud tasks “may use more allowance than local messages,” OpenAI says, and larger inputs and outputs, higher reasoning settings, Fast mode and multi-step tasks “can also increase usage.” A third page says Work “follows the same usage structure as Codex.” For someone who mostly uses Instant chat, OpenAI says that use will not count against the allowance.

Our read: OpenAI is making tokens the unit of account for a consumer subscription, not just for its API. That is a more direct reflection of what a task actually costs to run, and OpenAI’s own limits table lists many times more messages per window for Luna than for Astra. The trade-off is clarity: the company already says the allowance “does not correspond to a fixed number of messages or tasks,” so the answer to “how much do I get” becomes a range that moves with model, reasoning level and task. For people who live in Codex, the shared allowance means heavy coding can now eat into agent work in ChatGPT, and the reverse. Users who mainly chat should see little change if, as OpenAI says, Instant chat stays outside the meter. OpenAI’s expectation that most users won’t hit resets more often is a forecast, and the fallbacks it lists, waiting, upgrading or buying credits, are what a user who runs out will face. How often people reach for credits will show what the shared allowance really costs.

What’s next: OpenAI says the change starts in December; the article does not give a specific day. Users can check their combined remaining usage and reset times in Settings → Usage once it does. On the coding side, we recently reported on GitHub Copilot Fusion reaching general availability.

Sources: OpenAI: Understanding ChatGPT and Codex usage limits (Nov. 1, 2026); OpenAI: Managing usage with GPT-6 Astra in Work and Codex (Oct. 31, 2026); OpenAI: ChatGPT Work and Codex (Oct. 29, 2026)