OpenAI has revealed plans to combine usage allowances across ChatGPT and Codex, as part of a broader effort to integrate its conversational AI and agentic capabilities into a single experience.
In a help article quietly published over the weekend, the company confirmed the changes will take effect “starting in December,” affecting Plus, Pro and Business subscribers. Under the new system, ChatGPT conversations and Codex tasks will draw from a common pool of tokens, with more demanding models and complex tasks consuming a greater share — though OpenAI has yet to disclose how generous the combined allowance will be.
Instant chat, ChatGPT’s faster-response mode for everyday questions and tasks, will be exempt.
OpenAI’s products come together
The changes are the latest step in OpenAI’s efforts to bring its growing collection of AI products under one roof.
In July, the company launched ChatGPT Work, an agentic tool built on its Codex coding technology that allows users to delegate complex tasks, such as researching topics, analysing data and producing documents, presentations and spreadsheets. The idea was to extend the capabilities of Codex beyond software development, allowing knowledge workers to use AI agents for everyday business tasks.
As The New Stack reported at the time, the launch also saw OpenAI consolidate ChatGPT and Codex into a single desktop application, with dedicated interfaces for conversations, agentic tasks and coding.
ChatGPT Work joined an expanding collection of specialised tools, including ChatGPT for Excel, which debuted in March, which later extended to Google Sheets.
It’s worth noting that these products already have overlapping usage arrangements. Currently, Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents draw from a shared agentic allowance, while regular ChatGPT conversations have separate usage limits.
And it’s that very separation that’s set to change from December, putting developers in a position where intensive Codex sessions eat into the allowance available for ChatGPT conversations, and vice versa.
‘ChatGPT and Codex serve different purposes’
OpenAI isn’t alone in bringing its conversational and agentic products closer together. In September, Anthropic merged Claude Chat and Cowork into a single interface, allowing users to move from asking questions to delegating more involved tasks within the same conversation.
This shows how two of the biggest frontier AI labs are pursuing a similar goal: making it easier for users to switch between asking questions, and carrying out more substantial tasks, without having to navigate separate products. However, OpenAI’s decision to combine usage allowances has prompted objections from subscribers concerned that the change could leave them with less access to the tools they already pay for.
“ChatGPT and Codex serve different purposes, and their usage limits should remain separate.”
In a discussion kickstarted on OpenAI’s developer forum on Saturday, users called on the company to abandon the proposed consolidation, or guarantee that subscribers won’t end up with less usable capacity under the new system. The forum member who started the discussion, going under the moniker Smalikou, argues that the two products should retain independent allowances and called for greater transparency over how usage will be calculated.
“ChatGPT and Codex serve different purposes, and their usage limits should remain separate,” they write.
Another user, going by the name of PZA, calls for more detailed usage reporting, including a breakdown of consumption by product, model and task.
“If OpenAI decides to combine them anyway, a real-time usage dashboard must be available before the change takes effect,” they write. “It should show the remaining allowance, reset time, and consumption by product, model, and task. Without that transparency, I will consider the change a downgrade to my paid plan and reconsider renewing.”
“Without that transparency, I will consider the change a downgrade to my paid plan and reconsider renewing.”
As it happens, OpenAI has confirmed that subscribers will be able to see their combined remaining allowance and reset times in ChatGPT’s settings, but hasn’t specified what level of detail it will include.
The concerns come just days after OpenAI reversed a controversial decision to halve usage limits for its $200 Pro subscription. On October 29, the company confirmed it was cancelling the Pro 200 reduction, restoring the higher Work and Codex allowances for subscribers already moved to the lower limits, while doubling allowances for Pro 100, Pro 500 and Business Premium subscribers.
Tibo Sottiaux, OpenAI’s head of ChatGPT and Codex, acknowledged the criticism surrounding the proposed reduction, writing on X that the company had heard subscribers’ feedback and would preserve existing Pro 200 allowances.
“We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same,” he wrote.
However, one forum participant questioned how much those increases would benefit subscribers once ChatGPT conversations begin drawing from the same allowance as Codex. The user, posting as Zigg88, notes that they already regularly approach their weekly Codex limit.
“Doubling a resource allocation while simultaneously expanding the number of things consuming it doesn’t inherently translate into more usable capacity,” they write.
“Doubling a resource allocation while simultaneously expanding the number of things consuming it doesn’t inherently translate into more usable capacity.”
Others raised concerns about the impact on their ability to use ChatGPT for research and technical planning while relying on Codex for software development, with one commenter arguing that OpenAI should preserve independent allowances or guarantee equivalent access under the new system.
The New Stack reached out to OpenAI to ask when, exactly, the changes would take effect in December, and whether the new combined allowance would be equivalent to the existing separate allowances, or leave some subscribers with less usable capacity. The company declined to answer the questions, instead reiterating its recent increases to Work and Codex usage limits, and directing us to its existing documentation.
For now, OpenAI says it expects most users to encounter usage resets no more frequently than they do today. However, without publishing the new token allowances or explaining how they compare with existing limits, subscribers have little way of knowing whether that reassurance will hold true for them.
The post OpenAI’s new shared usage limit pits ChatGPT against Codex appeared first on The New Stack.
