Why it mattersA heavy day on Codex will now eat the same budget as regular ChatGPT queries. Teams that pay for one plan have to pick which work matters more.
A heavy day in Codex will soon eat into the same budget as regular ChatGPT questions. OpenAI announced the change on its Help Center page about usage limits. The page was updated around six hours before this item. The company plans to turn the change on in December. The reason given on the page is a product change: Codex is being folded into the main ChatGPT product.
What is changing
The page states that "Plus, Pro, and Business plans' usage allowance will be shared across ChatGPT and Codex" from December. Today each product has its own limit. Teams can run Codex hard against one bucket and still ask ChatGPT questions against another. In December those become one bucket.
The company says the new allowance is measured in tokens. The page states that cost will depend on which model the user picks and how large the task is. More advanced models and bigger tasks use more of the budget. OpenAI says "Instant chat will not draw from your usage" and remains free of the limit.
The reason OpenAI gives
The reason given on the page is a product change. OpenAI says it is bringing "the agentic capabilities of Work into the main ChatGPT experience". A user may start a long Codex task and then ask ChatGPT a short question. Under the new rule, both pieces of work count against one allowance.
What the help page does not say
The page does not state the number of tokens in a Plus, Pro or Business allowance. It does not say how much a single Codex task costs. It does not list the models that count against the budget, or how much each one uses. OpenAI says it will show "combined remaining usage and reset times in Settings → Usage" after December. The number only becomes clear once the change is live.
The page does name three ways out when a user runs out. OpenAI says a user can wait for the limit to reset, upgrade to a higher plan, or buy credits. Credits are the only way to keep working once the plan allowance is gone and before it resets.
Why some subscribers are worried
The Hacker News thread about the change has 59 points. One user there described what the change could feel like after a heavy day in Codex. They wrote that it kills them to run out of coding usage and then lose their regular ChatGPT work too. The complaint names Claude Code. The same pattern applies to Codex under the new rule.
Many teams pay one subscription per engineer. The engineer uses it for coding and for general questions. Now those two kinds of work share one budget. OpenAI says it does not expect most users to reset more often than they do today. The company has not published numbers behind that claim.
The team should check Settings → Usage after December, before planning any heavy work. The page tells users to look there for the combined remaining allowance. For teams whose Codex use already comes close to the limit, buying credits is now part of the plan.
Source
Understanding ChatGPT and Codex usage limits, OpenAI Help Center. Hacker News discussion.
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.
