OpenAI Merges ChatGPT and Codex Limits: What Changes

OpenAI’s December Usage Change: ChatGPT and Codex Will Share One Token Pool

OpenAI has confirmed that starting in December, Plus, Pro, and Business plans will use one shared token-based allowance across ChatGPT and Codex, a change that has already triggered concern from early users. The update means activity in either product will draw from the same pool, while Instant chat will remain excluded from the shared usage limit.

What is changing in December

OpenAI’s Help Center says the usage allowance for Plus, Pro, and Business plans will be shared across ChatGPT and Codex as the company brings agentic Work capabilities into the main ChatGPT experience. The shared usage will be measured in tokens, which OpenAI describes as the units of information a model processes and generates.

• Plus, Pro, and Business plans will share one usage pool.
• ChatGPT and Codex usage will both count against that pool.
• Instant chat will not draw from the shared allowance.
• More advanced models and complex tasks will generally use more of the allowance.

Why users are pushing back

The biggest concern is that heavy use of one product will reduce capacity for the other. Users who split their workflow between planning in ChatGPT and building in Codex worry that one long reasoning session or coding task could consume the same budget they need for everyday chat.

Early feedback has been especially critical among subscribers who feel the change makes the paid plans less predictable. Some users argue that a shared pool may blur the value of separate tools, especially for people who rely on both for different parts of the same workflow.

How token-based usage affects subscribers

Token-based accounting means the exact amount of usage consumed depends on the model selected and the complexity of the task. In practical terms, larger or more advanced requests are more expensive in usage terms, so mixed workflows may exhaust the shared allowance faster than users expect.

• Simple tasks should consume less usage than complex tasks.
• Advanced model use will likely reduce available capacity more quickly.
• Users will need to monitor combined usage across both products.

What remains unclear

OpenAI has not fully detailed the exact December rollout date or the new allowance sizes for each plan. That leaves some uncertainty around how much usage subscribers will actually get once the shared system goes live.

It is also not yet fully clear how all advanced features will be counted under the new system, beyond the stated token-based approach and the exclusion of Instant chat.

Why this matters for ChatGPT and Codex users

This change matters because it effectively ties two distinct workflows together under one consumption model. For casual users, the impact may be minimal, but for power users who alternate between chat-heavy and code-heavy sessions, the shared pool could make paid plans feel tighter.

As OpenAI folds more agentic Work features into ChatGPT, the company appears to be moving toward a unified product experience. The backlash shows that many subscribers still prefer clearer separation between tools, quotas, and use cases.

Conclusion

OpenAI’s December update will unify ChatGPT and Codex usage for Plus, Pro, and Business subscribers under one token-based pool, with Instant chat left out of the limit. The move may simplify OpenAI’s product structure, but early user reaction suggests that many subscribers see it as a step backward for flexibility and value.