ChatGPT Is Merging Its Chat and Work Tabs Into One

OpenAI is merging ChatGPT's Chat and Work tabs into a single workspace in the desktop app, according to an email the company sent to users and a report from TechCrunch on November 2, 2026. The combined view will let people move from a quick question straight into a longer, agentic task without switching tabs, and it's rolling out to a small group of users for feedback before OpenAI opens it up more widely. Codex, OpenAI's coding tool, stays as its own separate experience inside the desktop app — but starting in December 2026, ChatGPT and Codex are expected to draw from the same shared usage pool, a change OpenAI says will leave most people unaffected while some heavier users may hit their reset limits more often.

What's actually changing in the ChatGPT desktop app

Since July 2026, OpenAI's ChatGPT desktop app for Mac and Windows has organized itself around three separate experiences: Chat, Work, and Codex. Chat is the familiar conversational mode for quick questions, brainstorming, and search. Work is the agentic mode built for longer, multi-step jobs — research, spreadsheets, presentations, and full documents — using its own file system and browser session. Codex sits alongside both as OpenAI's dedicated coding assistant.

According to the email OpenAI sent to users and reported by TechCrunch, that three-way split is about to get simpler, at least for Chat and Work. The two will fold into one workspace, so a conversation that starts as a casual question can turn into a full research or document task without the person having to open a different tab or re-paste context. OpenAI's email reportedly framed it this way: the change is meant to let people "move from exploring an idea to having ChatGPT carry it through, all in one place."

In practice, that means the toggle at the top of the ChatGPT desktop window — the one that currently lets you flip between "Chat" and "Work" — is going away for the people who get the new experience. Files, plugins, and the ability to ask ChatGPT to build a website, put together a presentation, or work across the open web will all live inside the same conversation thread, rather than behind a separate mode switch.

Why OpenAI is combining Chat and Work

This isn't a sudden pivot. OpenAI's own leadership has been signaling for months that the Chat/Work split was meant to be temporary. ChatGPT and Codex product lead Tibo Sottiaux said on a podcast appearance in early October that OpenAI planned to combine Chat and Work into one experience, and OpenAI president Greg Brockman has separately indicated the two modes would merge by the end of 2026. The original July split itself was a response to an even earlier problem: OpenAI had just folded the standalone Codex app and the ChatGPT app into one unified desktop client, and the initial redesign buried chat history and made switching between modes confusing enough that Brockman reportedly called it "kind of a mess" before OpenAI patched the navigation back in mid-July.

Seen in that light, the Chat/Work merger looks like the next step in a yearlong effort to make one ChatGPT app feel coherent instead of like three bolted-together products. OpenAI has already unified Recents so Chat and Work conversations show up together with sort, filter, and pin options; collapsing the tabs themselves is the logical next move.

Who gets the merged experience first

Per the reporting, OpenAI isn't flipping this on for everyone at once. The unified Chat-and-Work workspace is going out first to a select group of users so OpenAI can collect feedback, with a broader rollout to follow. Neither the email nor TechCrunch's report specified a firm date for when the wider release lands, which fits OpenAI's pattern with other recent desktop-app changes — the original Work launch itself rolled out unevenly, reaching Pro, Enterprise, and Edu accounts first before Plus and Business plans followed days later.

If that staggered pattern holds again, it's reasonable to expect Plus, Pro, and Business subscribers to see the merged tabs before Free-tier users, though OpenAI hasn't said so explicitly for this specific change. Anyone who wants to get a head start can watch for an in-app notice or a prompt the next time they open the desktop app, since early feedback groups for OpenAI features have typically been selected automatically rather than through an opt-in waitlist.

Codex remains separate — but the usage pool is merging

The one piece that isn't changing, at least according to OpenAI's email, is Codex's place in the app. The message reportedly states plainly that "Codex will remain a separate experience in the ChatGPT desktop app." People who use Codex for coding tasks will keep their own dedicated view, separate from the newly merged Chat-and-Work workspace, with its own history and workflows untouched.

What is changing is how usage gets counted. TechCrunch's report says that with this transition, ChatGPT and Codex will draw tokens from the same allowance pool. That lines up with separate reporting from late October that OpenAI plans to unify ChatGPT and Codex token allowances for Plus, Pro, and Business subscribers starting in December 2026 — consistent with OpenAI's broader move earlier in the year to token-based usage limits rather than flat per-message caps.

Today, OpenAI already treats Work and Codex usage as linked in some contexts — both draw from the same "agentic" allowance in several plan tiers, while ordinary Chat usage is metered on its own, separate basis. Folding plain Chat into that same shared pool alongside Work and Codex would be a meaningfully bigger change than anything OpenAI has done with usage limits so far this year, since it means every exchange — not just the agentic, longer-running tasks — could draw on the same finite budget as a user's coding sessions.

Where your usage numbers will live: Settings > Usage

OpenAI already surfaces remaining allowance and reset timing for Work and Codex inside the app — currently split across a few different panels depending on which mode you're in. As part of this transition, that information is expected to consolidate into a single Settings > Usage view, which would match how OpenAI already describes checking weekly rate-limit resets for paid accounts: open Settings, find Usage, and see which allowance is running low along with the next reset time.

That consolidation matters more once the pools combine, since a single number will now reflect activity across Chat, Work, and Codex rather than three separate meters. For anyone on a metered plan, Settings > Usage is likely to become the one place worth checking before starting a big task, rather than an occasional glance at a banner inside Work or Codex specifically.

What OpenAI says this means for most users

OpenAI's email reportedly tries to get ahead of the obvious worry — that merging usage pools will make limits tighter. The company's position, per the reporting, is that the change won't affect most users. The caveat is that, depending on usage and which model someone is running, some people may need to wait for resets more often than they do today, since a single shared pool means heavy Codex use could eat into the budget someone also wants for everyday chat, and vice versa.

That framing tracks with how OpenAI has talked about usage changes earlier in 2026: broad limit increases and plan adjustments paired with occasional notes that specific workflows — long coding sessions, heavy Work-agent tasks — are the ones most likely to run into caps. Casual chat users, who make up the bulk of ChatGPT's roughly one-billion-plus monthly active base, are the group OpenAI is most clearly trying to reassure here.

How this fits OpenAI's bigger ChatGPT Work push

ChatGPT Work and Codex together reportedly passed 40 million active users by early October, and companies including Uber have put large employee bases on ChatGPT Work for internal tasks. That growth is likely part of why OpenAI wants the entry point simplified — a new or casual user who only ever opens "Chat" may never discover Work's document- and spreadsheet-building abilities if it stays hidden behind a separate tab they have no reason to click.

It's a similar instinct to what's driven other recent ChatGPT changes aimed at keeping more tasks inside a single thread, from ChatGPT's slide-making tools to its expanding plugin connections for Slack, email, and project trackers. Pulling Chat and Work together removes one more reason for someone to leave the main conversation window to get something finished. It also puts ChatGPT's interface a little closer to how rival tools like Anthropic's Claude have approached the same problem with its own Cowork feature, which lets Claude work through longer-running tasks alongside ordinary chat.

What to watch next

A handful of details remain genuinely unconfirmed. OpenAI hasn't published a timeline for when the merged experience reaches everyone beyond the initial feedback group, and the company hasn't said whether the model picker — the control that lets paying subscribers choose which underlying model handles a request — survives the merger in its current form. It's also not yet clear whether Free-tier accounts, which already have access to Chat, Work, and Codex in the desktop app, get the merged workspace on the same timeline as paid plans, or later.

The clearest signal to watch for is December 2026, when the shared ChatGPT/Codex token allowance is expected to take effect for Plus, Pro, and Business accounts. That's the point where the usage-limit questions raised by this rollout — including for people comparing Plus and Pro usage limits — should get real answers, since a shared pool changes the math on how far a subscription actually stretches. Anyone doing serious coding work through Codex alongside everyday ChatGPT use, including people who also lean on IDE-based tools like VS Code's Copilot integration, will want to keep an eye on their Settings > Usage panel once the pools combine.

Written byTheo Park

Theo Park runs the AI desk at Pandromeda. He follows model launches from the frontier labs and the open-weight community, tracks the assistants and developer tools built on them, and explains what each release changes on pricing, capability and safety. His reporting leans on primary sources: model cards, technical reports, API documentation and the companies' own announcements.

More from Theo ParkAll AI coverage