ChatGPT Work Limits: Usage Metering, Plan Gates, and Connector Caveats (2026)
ChatGPT Work has no fixed task quota and no per-task price — every task consumes a variable slice of your plan’s allowance depending on its complexity, and OpenAI hasn’t published a rate card. That’s the single most important limit to understand about OpenAI’s new agent, launched July 9, 2026. A quick summarization job and a four-hour research-and-build job draw from the same pool at wildly different rates, and the only way to learn your real capacity is to spend it.
Here’s every limit that matters — metering, plan gates, approval friction, connector caveats, and the structural one nobody’s headline mentioned.
The usage meter: variable draw, no published rates
ChatGPT Work uses the same metering structure as Codex: tasks are billed against your plan allowance, weighted by complexity. What that means in practice:
- No per-task price. OpenAI has not published what a task costs, and the same request can cost more or less depending on how much gathering, reasoning, and building the agent decides it needs.
- Long tasks are the expensive ones. The marquee feature — an agent that works independently for hours — is precisely the usage profile that drains an allowance fastest. You’re budgeting for open-ended compute, not a count of runs.
- You discover your ceiling by hitting it. There’s no calculator that tells you “this plan buys N reports per month.” Enterprise admins get spend controls; everyone else gets a meter and a surprise.
- Allowances reset on plan cycles, so a heavy project week can leave you rationing the agent for the rest of the month.
If you’ve hit ChatGPT’s regular message caps before, this will feel familiar — we covered that pattern in ChatGPT Plus limits. Work applies the same logic to a far hungrier workload.
Plan gates: who can even use it
Rollout is staged. Pro, Enterprise, and Edu got ChatGPT Work first (web and mobile) on July 9, with Plus and Business “over the next few days.” The new macOS desktop app that houses Chat, Work, and Codex side by side is available on all plans including Free — but having the app is not the same as having Work allowance to spend in it. Free users get the surface, not the agent.
Approval and check-in friction
Autonomy in ChatGPT Work is deliberately gated:
- Plan mode shows the steps before anything runs — you review and approve the approach.
- Check-ins pause long tasks for your input at decision points.
- Action approvals gate consequential moves individually.
These are sensible safety choices for an agent with access to 1,400 connectors. But they mean “works independently for hours” comes with an asterisk: it works independently between the moments it needs you. If you kick off a task before a flight, you may land to find it waiting on an approval it hit twenty minutes in.
Connector limits: the fine print on Gmail and Calendar
The 1,400-app directory is real, but depth varies a lot by connector, and the two most-used office apps show the pattern:
- Gmail can now send email — with four caveats. Sending works on web only (not mobile), without attachments, is not available in the EU or UK, and requires per-message approval — the agent can’t fire off a batch of follow-ups unattended. We dug into the connector’s full behavior in ChatGPT Gmail integration.
- Google Calendar is read-only. ChatGPT Work can see your schedule to inform a task, but it cannot create, move, or cancel events. An “assistant” that can read your calendar but not touch it still leaves the scheduling to you.
For research-and-build tasks, these gaps barely matter. For anything resembling executive-assistant work — sending, scheduling, following up — they’re the whole job.
The structural limit: no standing triggers
ChatGPT Work is task-based. You start each job; it runs; it ends. There is no “when a new lead arrives, triage it,” no “every Friday, send the pipeline report,” no “if an email goes unanswered for three days, nudge them.” The agent doesn’t watch anything, and it doesn’t fire on its own — and even if you manually re-ran a task daily, the metered allowance makes always-on operation the most expensive possible way to use it.
That’s not an oversight; it’s the product’s shape. ChatGPT Work is built for projects, not processes.
The predictable alternative for recurring work: Carly
If what you actually need is the process half — an assistant that acts whenever the trigger fires, with a bill you can predict — that’s Carly.
- Runs 24/7 in the cloud on triggers: schedules, incoming email, CRM changes, calendar events. Nothing to kick off, nothing to keep open.
- AI agents start at $35/month, and steps in a workflow that don’t use AI run free and unlimited. A workflow that moves data between apps a thousand times a day and only invokes AI where judgment is needed stays cheap — the meter isn’t running on every step.
- Actually sends email on Gmail and Outlook — no web-only, no per-message approval queue — and writes to your calendar and CRM instead of reading them.
- Connects to 200+ tools natively, and anything else via your own API key (paste it at dashboard.carlyassistant.com/integrations).
- Set up by conversation. Describe the workflow in plain English; Carly interviews you and builds it. No code.
The clean division of labor: brief ChatGPT Work on the big one-off project, and hand the recurring, trigger-driven work to Carly — where it doesn’t consume a mystery allowance every time it runs. For the full landscape, see the best ChatGPT Work alternatives.
Frequently Asked Questions
Does ChatGPT Work have a task limit?
Not a fixed one. Tasks consume a variable amount of your plan allowance based on complexity — the same structure Codex uses — so your effective limit depends on how heavy your tasks are. There’s no published per-task price or rate card.
Which plans have ChatGPT Work?
Pro, Enterprise, and Edu got it at launch on July 9, 2026, with Plus and Business rolling out over the following days. The macOS desktop app ships to all plans including Free, but the Work agent itself requires a paid plan with allowance. Full rundown in What is ChatGPT Work.
Can ChatGPT Work send emails automatically?
Only partially. The Gmail connector can send, but web-only, without attachments, not in the EU/UK, and each message requires your individual approval — so unattended sending isn’t possible. Google Calendar is read-only. For an assistant that sends on Gmail and Outlook without a per-message approval queue, see Carly.
Can I schedule ChatGPT Work to run every day?
No. ChatGPT Work is session-based — you start each task manually, and it doesn’t run on standing triggers or schedules. Recurring trigger-based work is a different product category: always-on assistants like Carly, where non-AI workflow steps run free and unlimited.
More: What is ChatGPT Work · Best ChatGPT Work alternatives · ChatGPT Plus limits · ChatGPT email assistant · AI news, July 9
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
