OpenAI merges ChatGPT and Codex quotas in checkout page update

The hidden quota change

OpenAI recently updated its checkout page to include a new clause regarding how paid plans are metered. While the company publicly announced that it was doubling limits for some Pro tiers, the specific details of a new shared quota were placed at the bottom of the terms section. Many users only discovered the change after screenshots of the checkout page began circulating in online communities.

The new rule states that starting in December, usage for ChatGPT and Codex will be combined into a single pool. Previously, these two types of interactions functioned as separate buckets of usage. This meant that if a user exhausted their coding limits, they could still perform general chat tasks, and vice versa.

By merging these pools, OpenAI is effectively ending the 'freewheeling' nature of chat within paid plans. Users will now have to make strategic choices about whether to use their remaining quota on a few messages or a single complex task run. As Vocal reported, the placement of this information led many to feel the change was intentionally obscured from the main announcement.

How the new metering works

To understand the impact, it is helpful to look at the difference between the two types of tasks. ChatGPT handles general interactions like writing, planning, and answering questions, which are typically contained within single sessions. Codex, on the other hand, handles heavy-duty coding, agent tasks, and workflow runs that require significant processing power.

Because Codex tasks are so resource-intensive, a single five-hour window can consume a significant portion of a weekly quota, ranging from 10% to 47%. Under the old system, these heavy tasks did not drain the pool used for everyday questions. Now, because they share a meter, a single long coding session can leave a user with very few messages left for the rest of the period.

OpenAI has already been sharing some backend quotas between Codex and ChatGPT Work, but this new update pulls general chat into that same pool. This transition aligns with OpenAI's broader product goal of merging the entrances for Chat and Work. Once the user interfaces for these tools are fully integrated, the shared billing system becomes the logical next step for the company's infrastructure.

Who is affected by the shift

The impact of this change will vary depending on how a user interacts with the platform. Light users who primarily use ChatGPT for occasional questions or quick writing tasks may not notice a significant difference in their daily experience. For these users, the quota may still feel large enough to cover their routine needs without hitting a hard ceiling.

However, heavy users and those who rely on Codex for professional workflows will feel the squeeze much sooner. Many Plus users previously treated near-unlimited chat as a safety buffer for when their more intensive tasks were finished. With the shared limit, that buffer is pulled into the same meter as the high-cost tasks, making the remaining balance much more volatile.

The reaction from the community has been mixed, with some users expressing frustration over the lack of a clear, proactive announcement. Some users have already begun calculating exactly how many chat messages they will lose after running specific Codex tasks. Others have expressed their dissatisfaction by moving their cursor toward the cancel button on the renewal page, signaling a potential loss of trust in the communication style.

The business logic of limits

From a business perspective, OpenAI has a clear motive for moving toward shared limits. Because Codex burns a significant amount of compute, it is difficult for the company to manage an overall budget when different tasks are siloed into separate pools. A single pool allows OpenAI to allocate scarce computing resources more directly and bring their billing closer to the actual cost of the operations.

This move reflects a broader trend in the AI industry where the cost of running advanced models is pushing companies away from free or unlimited access. As compute costs remain high, companies are forced to adopt tiered metering to remain profitable while still scaling their user bases. OpenAI is essentially hitting the brakes on unrestricted usage to ensure they can sustain the high costs of their most powerful tools.

While the business logic is understandable, the method of communication has drawn criticism. Rather than providing a clear explanation of why the merge is happening or how the metering will work, the company placed the key clause in the terms section. This has led to a perception that the change was an ambush rather than a transparent update, which can damage the relationship between the company and its power users.

What to watch next

The transition to shared limits is likely just the first step in a larger integration of OpenAI's tools. As the company continues to merge the interfaces for Chat and Work, we can expect more changes to how usage is tracked and billed. Users should keep a close eye on how the

try_itZ_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_it_