ChatGPT Usage Limits: What Changes for Australian Businesses

ChatGPT usage limits are changing in December, and Australian businesses on ChatGPT Plus, Pro or Business plans should plan for it now. OpenAI has updated its help centre to say that from December the usage allowance on those plans will be shared across ChatGPT and Codex, its coding agent, and measured in tokens. For a team that uses ChatGPT for documents and Codex for software work, that means one pool of usage instead of two. Here is what is changing, what OpenAI has not said yet and five practical steps to take before the switch.

What OpenAI is changing with ChatGPT usage limits

According to OpenAI’s help article Understanding ChatGPT and Codex usage limits, the change comes as OpenAI brings the agentic capabilities of ChatGPT Work, its agent for longer tasks, into the main ChatGPT experience. The key points are:

• One shared allowance. On Plus, Pro and Business plans, ChatGPT and Codex will draw from the same usage allowance.
• Measured in tokens. How much you use depends on the model you choose and the task. More advanced models and complex tasks generally use more.
• Instant chat is exempt. Quick conversations using Instant will not draw from the allowance.
• One place to check. Combined remaining usage and reset times will appear in Settings → Usage. At the limit you can wait for a reset, upgrade your plan or buy credits for eligible features.

OpenAI says it expects most users won’t need to wait for their usage to reset more often than they do now. BleepingComputer reports that OpenAI has emailed subscribers about the change and will first test a combined Chat and Work experience with a small group of users, while Codex stays a separate experience in the ChatGPT desktop app.

Who the new ChatGPT usage limits affect

The teams most affected are those that mix everyday ChatGPT use with heavier agent work. OpenAI’s Codex help article already says Codex, ChatGPT Work, ChatGPT for Excel and workspace agents use a shared allowance and credit pool on plans that have one, and that a long-running task can use substantially more than a short request. From December, a long coding session or a big document task could leave less room for advanced chat, and the reverse.

For ChatGPT Business, OpenAI’s plan overview lists two seat types. Standard seats are US$25 per user per month billed monthly or US$20 billed annually, and Premium seats are US$125 monthly or US$100 annually, with five times the usage of Standard, no five-hour limit and weekly resets. Teams can mix seat types, and OpenAI notes pricing may vary by country and currency, so check what your Australian account is actually charged. Per OpenAI’s release notes, it increased the Work and Codex usage included in Pro and Business Premium plans on 29 October.

What we don’t know yet

The help article does not give an exact December start date, how many tokens each plan includes or whether the current five-hour and weekly windows will change. Until OpenAI publishes those numbers, treat any estimate of “how much you’ll get” with caution, including ours. The practical response is to understand your own usage now, so you can compare it with the new figures when they arrive.

Five steps to prepare before December

1. Map who uses what. List which staff use ChatGPT chat, Work, Codex, the Excel or Word add-ins and workspace agents. Business workspace owners can download usage reports from Workspace settings → Billing.
2. Match seats to workloads. Give Premium seats to the few people running heavy agent or coding work, and keep Standard seats for everyone else, rather than upgrading the whole team.
3. Set alerts before you buy credits. OpenAI’s flexible pricing FAQ says Business owners can set credit usage alerts and automatic recharge. Credits are valid for 12 months and are generally non-refundable, so start small.
4. Teach model choice. Because advanced models use more of the allowance, staff should use Instant for quick questions and save higher reasoning for work that needs it. Our free AI prompts course and AI in the workplace course cover this kind of everyday judgement.
5. Write down your rules. Agree which tasks belong in ChatGPT, which in Microsoft Copilot or other tools, and who approves extra spend. If you are weighing options, start with our choose AI tools for business course or our Microsoft Copilot training.

When a subscription stops being the right tool

Seat-based plans suit people working interactively. They are less suited to high-volume, repeatable jobs such as processing every incoming invoice, answering the same customer questions or running nightly reports. Those workloads usually belong in software that calls a model through an API, where each task has a known cost and you choose the model. OpenAI’s release notes also say eligible ChatGPT Business workspaces can now activate promotional API credits, which are separate from ChatGPT usage credits.

That is where ChatGPT and LLM integration and purpose-built AI agents come in. A model-agnostic design also lets you switch to cheaper models such as GPT-6.1 Terra, or to a private LLM option for sensitive data, without a rebuild. If you are not sure whether a task needs an agent at all, our guide to automation vs agents explains the difference, and our custom AI development team can cost both approaches.

This is general information based on OpenAI’s help centre articles and release notes and published media reporting as at 2 November 2026. Plan prices are as listed by OpenAI in US dollars and may differ for Australian accounts. Limits and dates may change; check OpenAI’s current pages before changing plans or budgets.

Related reading: Once you know where the spend is going, our guide to the best AI tools for small business maps free and paid options by the job you need done (phones, admin, marketing, bookkeeping and more), and links the free Start Smarter beginner course.

Featured image: Photo by Burak The Weekender on Pexels.