OpenAI Merges ChatGPT Chat and Work Tabs, Unifies Limits
OpenAI is merging ChatGPT's Chat and Work tabs. You can now turn ideas into finished projects in one place without switching modes.
Tools · Source: TechCrunch
What happened
OpenAI is finally fixing its confusing desktop software. The company is merging the Chat and Work tabs inside the ChatGPT app. An email sent to users confirmed the upcoming changes. A small group of users will get access to this combined experience first. OpenAI wants to gather feedback during this testing phase before rolling the unified interface out to the wider user base.
The goal is to stop users from losing track of their projects. You can now move straight from exploring an idea to executing it. Everything happens in one place. You can bring your own files and plugins into the conversation. From there, you can ask ChatGPT to build a website or create a presentation. The AI will do things for you on the web while you review and guide its progress.
One major feature is staying exactly where it is. Codex will remain a completely separate experience within the desktop app. However, the billing mechanics are changing behind the scenes. OpenAI is unifying the usage allowances. Both ChatGPT and Codex will now draw tokens from the exact same pool.
Key facts
- October 29 — Date OpenAI doubled Work and Codex allowances for Pro 100, Pro 200, and Pro 500 subscriptions
- September — When rival Anthropic merged Claude's chat and Cowork functions
Why it matters
Building products with AI just got significantly less fragmented. You no longer have to jump between different tabs to brainstorm and execute complex tasks. This interface update mirrors a move by a major competitor. Anthropic already merged Claude's chat and Cowork functions into a single screen back in September. OpenAI is playing catch up to reduce user friction. This matters for founders who rely on AI agents for multi-step workflows. You can maintain your context without breaking your train of thought.
The second-order effect is all about your token budget. The unified allowance pool changes the math for heavy users and engineers. Since ChatGPT and Codex now share a single limit, running complex web tasks in Chat will drain your coding capacity. Advanced models and heavy workloads will consume this shared allowance much faster. Builders will need to monitor their usage closely under the settings menu. If you are not careful, you will hit reset walls more frequently during intense development sprints.
For builders
Streamlined agentic workflows in one window
You can now guide ChatGPT through complex web tasks and file manipulations without switching contexts. This saves massive amounts of time for founders who use AI to build websites. You pay nothing extra for the convenience, but you gain a smoother workflow.
Shared allowance drains your limits faster
ChatGPT and Codex now pull from the exact same token pool. Heavy coders lose out in this scenario. Complex tasks in your regular chat will cannibalize your Codex capacity and force you to wait for allowance resets or buy more credits.
Unlimited Instant provides a usage loophole
OpenAI explicitly stated that Unlimited Instant queries will not count toward your shared usage limit. Builders can exploit this to handle basic questions for free. You win by saving your shared allowance strictly for heavy lifting and advanced models.
My take
OpenAI is finally fixing a bloated desktop app. Forcing users to switch tabs to get real work done was always a terrible design choice. But bundling the Codex and ChatGPT usage limits is a sneaky way to throttle power users. Even though they doubled the Pro allowances late last month, heavy builders will absolutely feel the squeeze.
Original reporting: TechCrunch. This is my rewrite and opinion.