Unifying usage allowances between ChatGPT and Codex
OpenAI has confirmed that starting in December 2026, subscribers to the Plus, Pro, and Business plans will share a single usage allowance between ChatGPT and Codex. Until now, these services operated with separate limits: Work and Codex shared a quota, while standard chat had its own restrictions. With this change, token consumption on either platform will draw from the same common pool.
This modification aims to simplify consumption management for users who alternate between conversation, document analysis, and software development. The company states that the goal is to prevent a user from running out of capacity in one tool while still having balance in the other, although this means an intensive coding session will directly reduce availability for chat tasks.
Plans affected by the allowance unification
Summary of the subscription plans that will share the token allowance between ChatGPT and Codex starting in December 2026.
| Plan | Included in unification | Specific notes |
|---|---|---|
| Plus | Yes | Maintains a five-hour limit per session. |
| Pro | Yes | No five-hour limit, but may have weekly limits. |
| Business | Yes | Maintains a five-hour limit per session. |
| Free, Go, Enterprise, Edu | No | Outside the scope of this specific announcement. |
Integration of Work features into Chat
In addition to the allowance unification, OpenAI is preparing a combined experience that will integrate the capabilities of the Work tab directly into the Chat interface. This integration will allow users to move from exploratory conversations to executing complex tasks, such as creating websites or presentations, without switching modes.
The company has indicated that this functionality will be tested initially with a small group of users in the coming weeks before expanding. It is important to note that Codex will continue to exist as an independent experience within the ChatGPT desktop application, maintaining its specific focus on code development.
Token-based consumption mechanism
The new system will measure usage in tokens, the basic unit of information that artificial intelligence models process and generate. The amount of allowance consumed will depend on two main factors: the specific model used and the complexity of the requested task.
Operations using advanced reasoning models or tasks requiring multiple steps and tools will consume a larger portion of the shared quota. In contrast, brief interactions or the use of Instant mode will not deduct from this common allowance, allowing users to maintain fluid access for everyday queries without exhausting the budget intended for more demanding work.
Exemptions and limits of Instant mode
OpenAI has clarified that Instant mode, designed for quick responses and general conversations, will not consume the shared ChatGPT and Codex quota. This means users can continue to use this mode for low-consumption tasks without affecting their capacity to execute agent work or coding.
However, official documentation states that although Instant usage does not deduct from the main quota, the company maintains safeguards against abusive use. No specific numerical threshold has been published that defines when these restrictions apply, so users should be aware that access is not technically unlimited in terms of extreme request volume.
How to check remaining consumption
Process to check the status of the shared quota once the new system is implemented.
- 01Access the ChatGPT application.
- 02Go to the Settings menu.
- 03Select the Usage section.
- 04Review the remaining balance and the next reset date.
Implications for users and teams
For professionals who use these tools intensively, the change requires more careful planning of resources. Teams that rely heavily on Codex for software development may find that their consumption directly impacts the capacity of their colleagues to perform analysis or writing in ChatGPT.
OpenAI states that most users will not experience an increase in the frequency with which they have to wait for their quotas to reset. However, the company has not provided comparative figures that allow this claim to be verified independently. Users should monitor their usage patterns and be prepared for potential adjustments in their workflows if actual consumption differs from expectations.
Open questions
- The exact effective date within December 2026 has not been specified.
- Specific numerical token limits for each plan (Plus, Pro, Business) under the new system have not been published.
- It has not been detailed how the five-hour limits of Plus and Business plans will interact with the shared token quota.
- The claim that most users will not wait longer for resets is not supported by publicly available comparative data.
Keep exploring
Sources consulted
Corrections and transparency
If you spot incorrect or outdated information, send us a correction with the page and source we should review.
Submit a correction