OpenAI to merge ChatGPT's Chat and Work tabs and share one allowance with Codex

OpenAI plans to merge the Chat and Work tabs in ChatGPT, according to an email it sent to users and reported by TechCrunch and BleepingComputer. A small group of users will test the combined experience first. OpenAI will roll it out more widely after that. Codex, its coding agent, stays a separate section of the ChatGPT desktop app.

The interface is only one part of the change. OpenAI also said ChatGPT and Codex will draw from the same usage allowance. A conversation, a delegated task and a coding session will all count against one budget, measured in tokens. Some heavy users may have to wait for resets more often.

One place for chat and delegated work

The ChatGPT desktop app has two sections, ChatGPT and Codex, and the ChatGPT section is split again into Chat and Work tabs. TechCrunch calls the result confusing: it is easy to lose track of which tab you are in and where a project lives. The Decoder reports that users currently pick between two operating modes and 11 model settings.

OpenAI's email explains what the merged version is meant to do. "This will enable you to move from exploring an idea to having ChatGPT carry it through, all in one place," it said. "You can bring in your files and plugins and ask ChatGPT to build a website, create a presentation, or do things for you on the web, while reviewing and guiding its work. Codex will remain a separate experience in the ChatGPT desktop app."

BleepingComputer, which also saw the email, reports that testing starts "in the coming weeks". The Decoder, citing an updated OpenAI support page on ChatGPT and Codex usage limits, reports that the merger begins in December.

OpenAI is not the first to do this. TechCrunch notes that Anthropic merged Claude's chat and Cowork functions into one screen in September.

A single token budget

"Your use of ChatGPT and Codex will draw from the same allowance," OpenAI's email says, according to BleepingComputer. The details reported so far:

• The Decoder, citing the support page, says Plus, Pro and Business plans will share one token budget across ChatGPT and Codex.
• Chats with Instant models do not count toward the shared limit. BleepingComputer describes unlimited Instant use as included in the subscription.
• More advanced models and more complex tasks use up the allowance faster.
• Users can see their combined remaining usage and reset times in Settings, under Usage.
• When the allowance runs out, users can wait for a reset, upgrade their plan, or buy credits for eligible features.

OpenAI told users the change will not affect most of them. It added that, depending on usage and model, some may need to wait for resets more often. No numeric limits or credit prices appear in the reporting.

The pooling follows an earlier change. BleepingComputer reports that on October 29 OpenAI doubled the Work and Codex allowances for its Pro 100, Pro 200 and Pro 500 subscriptions. The Decoder presents that increase as preparation for the merger.

How the product is changing

Until now, ChatGPT kept talking and acting in separate places. Chat was for questions and drafts. Work and Codex were for tasks the model carries out over several steps, often with files, plugins or a browser. Putting chat and Work in one view removes that split for most users. Developers keep a dedicated space in Codex.

Metering is changing too. The more expensive activities, such as advanced models, agentic tasks and coding, now draw on one shared budget, and the cheapest model stays outside it. The Decoder headlines its report as the end of "all-you-can-prompt" pricing. It argues that ChatGPT stays flat-rate "on paper", but that heavy use on cheaper plans will likely mean falling back to less capable models.

Two questions remain open. OpenAI has not published how many tokens each plan includes. The reports also differ slightly on scope: BleepingComputer frames the shared allowance around Pro subscriptions, while The Decoder lists Plus, Pro and Business.

What it means for work

In the merged view, the same window handles a question and the job that follows from it. Building a slide deck, drafting a website or completing a web form can start inside the conversation. ChatGPT does the steps and the user reviews and steers. The person's role moves from writing the output to briefing and checking it. Shared budgets also give users a new decision to make: someone who uses both ChatGPT and Codex must choose which model and which task deserves the allowance, since a long agentic run can leave less for coding later. For teams, the Settings view of remaining usage and reset times becomes something to watch. Still unknown: how large the allowances are, how fast typical agentic tasks consume them, and when the merged experience reaches users outside the test group.