OpenAI will move Plus, Pro and Business subscribers to a single shared usage allowance for ChatGPT and Codex in December 2026, according to a Help Center notice. Usage will be measured in tokens, with consumption depending on the selected model and task complexity. Instant chat will stay outside the shared pool, which matters because coding and long research tasks will now compete for the same quota.
How the shared allowance will work
Once the change takes effect, subscribers will see combined remaining usage and reset times under Settings → Usage. Users who hit the limit can wait for the reset, upgrade the plan, or buy credits to continue using eligible features. OpenAI said it is preparing the transition carefully to avoid disruption to workflows and expects most users will not need to wait for resets more often than now.
Consumption will vary by model choice and requested task, with more advanced models and complex tasks generally using more allowance. For Codex, usage depends on the model, execution location, task complexity, context, reasoning, speed and tools. A long-running task can consume substantially more than a short request, while Codex limits operate on five-hour and weekly windows.
Much of this metering already exists in shared form. Codex, ChatGPT Work, ChatGPT for Excel and Workspace Agents draw from a shared allowance and credit pool on plans that use shared allowances, and ChatGPT for Word also draws from that pool. Codex itself is included across ChatGPT plans, including Free and Go, while Codex Cloud runs coding tasks on OpenAI-managed computers for eligible Plus, all Pro tiers, Business, Enterprise, Healthcare and Education accounts.
What the change means for AI buyers
For companies, the practical effect is one budget for dialogue, research deliverables and code instead of separate quotas. ChatGPT Business bundles ChatGPT, Work and Codex in Standard and Premium seats, with Premium seats including more Work and Codex usage and no five-hour limit. Standard seats use per-model local-message estimates per five hours, for example 15-150 messages on GPT-6.1 Sol or GPT-6 Sol and 250-2,000 on GPT-5.6 Luna.
The main risk is uneven consumption across teams and unclear forecasting. Pro-model chat allowances remain separate, with Business Standard including 15 Pro messages per month and Business Premium 50 per week across GPT-6 Pro and GPT-5.6 Sol Pro. OpenAI Support does not reset ChatGPT or Codex usage limits, and a full banked reset refreshes both Codex windows while changing the weekly reset date.
The marker to watch is the December rollout and the first reset cycle under Settings → Usage. If shared balances, reset times and credit top-ups work without extra waiting, the consolidation will simplify administration. Repeated limit hits by Codex-heavy teams would signal a need for Premium seats, plan upgrades or stricter rules on model selection.
