$ less news/make-money-with-chatgpt-when-chat-and-work-merge-one-allowance.md

Make money with ChatGPT when Chat and Work merge: one allowance now covers your agent and your Codex work

2026-11-04

If you make money with ChatGPT by letting an agent run paid tasks, the place you paste the prompt is about to move. In an email to subscribers this week, OpenAI said it will start testing a combined Chat and Work experience with a small group of users in the coming weeks, so you can go from exploring an idea to having ChatGPT carry it through, with files, plugins and tasks on the web in one place. Codex stays a separate experience in the desktop app (BleepingComputer, TechCrunch).

The changelog line I care more about is the second one: your use of ChatGPT and Codex will draw from the same allowance. OpenAI's email says most users won't wait for resets more often than today, usage depends on model and task, more advanced models and complex tasks use more, and Unlimited Instant is included in the plan and doesn't draw from it. Combined remaining usage and reset times show under Settings > Usage, and if you run out you can wait, upgrade or buy credits for eligible features. Separately, Work and Codex allowances on Pro 100, Pro 200 and Pro 500 were doubled from October 29. And the ChatGPT Classic desktop app retires on January 30, 2027 (9to5Mac).

What that means for money-making prompts: until now, an earning agent in Work and a coding session in Codex were two separate meters. Once the pool is shared, a long Codex refactor and a run of paid research tasks eat the same budget, and whichever runs first decides when the other one hits a reset. Nothing in my prompt changes. It's still one line, and /setup still holds the rules: a bitcoin address or an email (verified only at cash-out), tasks one after another until you say stop, and cancel_earning_task for anything stuck.

My test plan for the week the merge reaches your account: 1) screenshot Settings > Usage before you start, so you have a baseline for the combined pool and the reset time; 2) run the earning prompt in one conversation and nothing else for one paid task; 3) check usage again and write down what that task burned next to the sats it was credited; 4) don't buy credits to keep an earning agent going, because then you're paying per task for income you haven't measured yet; 5) if you also code in Codex, decide which work gets the pool first and stop the agent before a reset, not after. I haven't been moved into the merged test yet, so I'm not claiming anything about how it behaves there.

Marcus wrote the money-side half of this check on ubitokens. His note on why a finished agent job isn't income until the buyer signs off on it is the right column to put next to the usage you just wrote down: count what's credited, not what ran.

Why ubiprompt is the best place for money-making prompts when the app underneath gets rearranged: I treat the prompt like code. It's one versioned line that doesn't depend on which tab you're in, the behaviour rules are written down on a public setup page instead of buried in a chat, and when the platform changes I publish a test plan and say plainly what I haven't tested. I don't publish earnings I can't show, mine or anyone's.

— priya · get the prompt

And before you budget your weekend around a reset timer, read who collects when the agents do the work on neetbros.