OpenAI is changing how usage is calculated for ChatGPT Plus, Pro, and Business subscribers starting in December 2026. Instead of separate limits, ChatGPT and Codex will share a single usage allowance, while Work features become more deeply integrated into the main ChatGPT experience.
Under the new rules, usage will be measured in tokens, the basic units of information that AI models process and generate. The amount of available usage consumed will depend on the selected model, the complexity of the task, and the volume of data being processed.
More advanced models and demanding tasks will naturally consume more of the available allowance than short, simple requests.
Heavy users of advanced models and Codex will feel the biggest impact
The most important detail for most users is that regular Instant conversations will not count toward the shared usage limit. This means everyday questions, short conversations, translations, and similar basic tasks will not consume the same allowance used by Codex, Work, and more advanced features.
The change will therefore be most noticeable for users who rely on ChatGPT for lengthy analyses, working with large documents, programming, research, and other complex tasks.
Work and Codex already share the same usage allowance, and starting in December, eligible advanced ChatGPT features will become part of this system as well.
OpenAI says it expects most users will not reach their usage limits more frequently than they do today, although actual consumption will depend on how the services are used.
Discussion(0)