ChatGPT and Codex may share one usage limit for paid subscribers
OpenAI plans a single usage gauge for ChatGPT and Codex on Plus, Pro and Business plans, with complex tasks drawing more of the allowance, TechRadar reports.
OpenAI is preparing to merge the usage allowances for ChatGPT and Codex, its coding assistant, according to TechRadar, which cites the company's updated documentation. Rather than tracking the two products separately, subscribers on the Plus, Pro and Business plans would see one gauge that covers both.
The documentation indicates that more demanding requests will use up more of that shared allowance. Ordinary Instant conversations, the lighter-weight chat mode, would not count against it, TechRadar reports.
The change arrives as OpenAI readies ChatGPT Work, a set of more autonomous features, for the main ChatGPT experience. Users would be able to ask ChatGPT to carry out complicated projects straight from a conversation, which means those projects would draw on the same pool of resources as coding work done in Codex.
The report does not spell out exactly how heavy a task must be to count as complicated, how the gauge will look to users, or when the change will take effect for each plan.
Background
Subscription AI products typically meter usage because the computing needed to run large models is expensive. Companies usually set limits on messages, files or processing time for each plan so that heavy users do not overwhelm the service and so that prices stay sustainable. Higher-priced tiers generally allow more usage than cheaper ones.
ChatGPT has long offered several paid levels, and OpenAI has expanded what those levels include over time. Codex is aimed at software development, where a single request can involve reading a codebase, writing code, running tests and revising the output. Tasks like these consume far more computing than a short question and answer, which is why coding tools often have their own limits.
The phrase "agentic" describes AI systems that pursue multi-step goals with less moment-to-moment input from the user. Such systems can take actions, check their own progress and keep working on a task over time. That capability is valuable, but it also makes usage harder to predict, because the amount of work a single request triggers can vary widely.
Many companies in this market have moved toward flexible or task-based metering in recent years. Combining several products under one allowance is one way to simplify billing and reflect the fact that users increasingly move between chat, coding and automation in a single workflow.
Why it matters
For people who already pay for ChatGPT, the change could alter how much they can get done in a given period. A user who relies heavily on Codex may find that routine chat work and coding now compete for the same allowance. Someone who mostly uses quick conversational features may notice little difference, since those interactions reportedly do not count against the limit.
The shift also signals where OpenAI sees its product heading. If ChatGPT is meant to run longer and more involved projects, then usage will increasingly depend on how much autonomous work people ask it to do. A single gauge makes sense for that model, but it also means that the cost of each task becomes less visible to the user unless the interface explains it clearly.
For businesses on the Business plan, predictability matters. Teams planning around AI tools need to know whether a shared limit will hold up during busy periods, especially when developers and other staff draw on the same account resources.
The broader market is watching how these trade-offs are handled. Customers tend to accept limits when they understand them, and they push back when limits appear suddenly or feel arbitrary.
What to watch
Watch for the exact definition of a complicated task and whether OpenAI publishes examples or a conversion guide. Clear rules would make it easier for subscribers to plan their work.
Look at how the single gauge appears in the product. A visible, real-time display would help users avoid surprises, while a vague indicator could generate frustration.
Track whether the change reaches all three named plans at the same time and whether the allowances themselves change alongside the new structure. Any shift in the amount of usage included at each price point would matter as much as the merger.
Finally, watch for the arrival of ChatGPT Work in the main experience. Its rollout will show how much autonomous activity users actually request and how the shared pool holds up under that demand.
Viewpoints
Commentary from our columns: three angles on this story.
The User what changes for people who use it
If you pay for ChatGPT, the most useful question right now is simple: what will I actually be able to do in a month? A single gauge covering chat and coding sounds tidy, but it asks people to make a trade they may not have made before. Every big project you hand to ChatGPT could now quietly shrink the room left for your coding sessions, and the reverse is true too.
The good news is that everyday conversations reportedly stay outside the limit. For many people, that means nothing changes at all. For developers and heavy users, though, the experience could feel different, especially once longer, autonomous projects start running straight from a chat window.
What would make this change feel fair is transparency. Users need to see how much of the allowance a task is consuming before they start it, not after they run out in the middle of a job. A plain, readable gauge with clear examples would turn a potential source of frustration into an easy habit to manage. Until those details arrive, the best advice is to pay attention to your usage and plan big projects with some room to spare.
The Ledger the business angle
Shared allowances are a familiar business move. Bundling products under one meter lets a company simplify its pricing story, avoid separate caps that confuse customers, and nudge people toward a single subscription. It also makes the subscription harder to compare with rivals, which is worth remembering when a company describes a change as mere convenience.
The economics are straightforward. Heavier tasks cost more to run, so tying their weight to the allowance keeps the plan's price connected to actual demand. That is sensible, but it also means the real value of a Plus, Pro or Business subscription may now depend on how the company weighs each kind of task. Customers will judge that by how much they get done before hitting a wall.
For business buyers, predictability is the currency that matters. A team choosing between AI vendors will ask whether a shared pool holds up during a busy quarter and whether costs can be forecast. If OpenAI answers those questions clearly, the bundle could strengthen its position. If not, rivals offering more separate or more transparent limits will have an easy pitch.
The Skeptic open questions and weak points
Before we celebrate simplicity, we should ask what the simplicity hides. A single gauge sounds neat, but it depends entirely on how "complicated" gets defined, and the reporting does not say. Until there is a clear definition with examples, subscribers will be guessing how fast their allowance disappears, which is a weak position for anyone paying a monthly fee.
There is also a question of direction. Moving ChatGPT toward more autonomous, project-length work means the tasks users ask for will grow. Combining that with coding under one pool could make the allowance feel smaller over time, even if the plan prices stay the same. Companies often say limits are about fairness, yet fairness depends on whether people can see and predict what they are using.
None of this proves the change is a bad idea. It does mean the burden of proof sits with OpenAI. The company should show how the gauge works, publish conversion guidance, and explain what happens when a user runs out mid-project. Until it does, a shared limit reads less like a convenience and more like an open question.