OpenAI is merging ChatGPT and Codex quotas

OpenAI is consolidating usage limits across subscription plans into a single pool. According to an announcement published in the company’s Help Center, Plus, Pro, and Business subscribers will move to a single shared quota for ChatGPT and Codex starting in December 2026. Instant chat will not draw from this shared pool.

Under the new structure, usage will be measured in tokens, the unit of information a model processes and produces. How much of a subscriber’s quota is consumed will depend on the model chosen and the complexity of the request; more advanced models and harder tasks will generally use more quota.

The reasoning behind the change

OpenAI describes this step as part of the process of bringing the agentic features of the Work experience into the main ChatGPT interface. Work is a version of ChatGPT designed for long research and tasks that produce deliverable outputs. The company says it has carefully prepared for the December transition so as not to disrupt user flows, and that most users should not need to wait for their quota to reset more often than they do now.

Once the change takes effect, subscribers will be able to see their remaining combined usage and reset times in the Usage section of the Settings menu. Users who run out of quota can wait for the reset, upgrade their plan, or purchase credits to keep using eligible features.

Features already drawing from shared pools

OpenAI already measures some items from shared pools. On plans that use shared quotas or credits, Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents draw from the same quota and credit pool. ChatGPT for Word also draws from the shared usage quota for Codex and other premium features.

Codex is available on all ChatGPT plans, including Free and Go. Codex Cloud, which runs coding tasks on computers managed by OpenAI, is available on eligible Plus plans, all Pro tiers, Business, Enterprise, Healthcare, and Education accounts. According to the company’s documentation, consumption depends on the model, where the task runs, complexity, context, reasoning, speed, and the tools used. A long-running task can use significantly more quota than a short request.

Five-hour and weekly windows

Codex limits work with five-hour and weekly windows. Using one of the accumulated resets refreshes both windows and changes the weekly reset date. OpenAI support does not reset ChatGPT or Codex usage limits.

On the Business side, ChatGPT, Work, and Codex are sold together; there are Standard and Premium seats. Premium seats include more Work and Codex usage and do not have a five-hour limit. On Standard seats, estimated message ranges by model are applied within the five-hour window: 15-150 messages for GPT-6.1 Sol or GPT-6 Sol, and 250-2,000 messages for GPT-5.6 Luna. The Pro model chat quota is kept separate from Work and Codex usage; Business Standard includes 15 Pro messages per month, and Business Premium includes 50 Pro messages per week.

Recent adjustments to the same pool

The December plan is the latest in a series of changes to the same limits. According to release notes dated October 29, 2026, OpenAI doubled the Work and Codex quota on the Pro 100 and Pro 500 plans and canceled the Pro 200 discount announced on September 29, 2026. As a result, affected Pro 200 subscribers moved to twice their previous quota and reached the same level as the other plans. The September notes also introduced the $500-per-month Pro 500 plan and reopened Pro 200 subscriptions at $200 per month.

Source: Unite.AI