Your ChatGPT chats are about to share a meter with Codex

👋 Easy read · Nov 3, 2026 · ~7 min read

Quick terms: Codex is OpenAI’s coding agent, which you can use in the ChatGPT desktop app, a terminal, your code editor or the web. A token is a small chunk of text a model reads or writes, so a long task means many tokens.

Your allowance is the amount of use your plan includes before you hit a limit. A reset is when that allowance refills. Credits are a pay-as-you-go balance you can buy to keep going after the included amount runs out.

What changes in December

Starting in December, OpenAI says the usage allowance on Plus, Pro and Business plans “will be shared across ChatGPT and Codex as we bring the agentic capabilities of Work into the main ChatGPT experience” (OpenAI Help Center: understanding ChatGPT and Codex usage limits).

Put simply, your regular chats and your Codex work will draw from the same allowance.

That change notice is four short paragraphs. It doesn’t give a size for the combined allowance, and it doesn’t name a December date. What it does say, as of Nov 3, 2026, is this:

• Usage for both will be measured in tokens.
• How much you use depends on the model you pick and the task you give it, and more advanced models and complex tasks generally use more.
• “Instant chat will not draw from your usage.”
• You’ll see your combined remaining usage and reset times in Settings → Usage.
• At the limit you can wait for the reset, upgrade your plan or buy credits.

We first flagged the change in a Daily on Nov 2. This piece goes one step further and pulls together the rest of what OpenAI has published about how the meter works.

“We expect most users won’t need to wait for their usage to reset more often than they do now.” OpenAI Help Center

That line is OpenAI’s expectation, not a measurement, so treat it as a promise to watch rather than a fact about your own account.

Why a quick question and a long task cost differently

Counting tokens instead of messages is the part that changes how it feels. A message is a message, but a token count grows with everything the model reads and writes along the way.

OpenAI’s Codex help page lists what moves the number. Usage depends on “the model, where the task runs, task complexity, context, reasoning, speed, and tools” (OpenAI Help Center: using Codex with your ChatGPT plan). It also says it plainly.

“A long-running task can use substantially more than a short request.”

So a one-line question to a light model and a two-hour Codex job on the most capable model are not the same size, even though each starts with one prompt. Picking the model is the lever you control most directly.

What already shares one pool today

Sharing isn’t new for everything. On plans that use shared allowances or credits, OpenAI says Codex, ChatGPT Work, ChatGPT for Excel and Workspace Agents already “use a shared allowance and credit pool” when those features are on your plan (OpenAI Help Center: using Codex with your ChatGPT plan).

The credits page adds that on Plus and Pro, PowerPoint and Word use that same agentic allowance too (OpenAI Help Center: using credits in ChatGPT personal plans).

Voice is split. A task you start by voice in Work or Codex on desktop uses that same allowance, while the connected voice time itself is metered separately where flexible pricing applies.

What December adds is regular ChatGPT use. If you already use Codex or Work, your chats are the new guest at a table that was already shared.

What doesn’t touch the allowance

A few things stay off the meter, according to OpenAI’s pages as of Nov 3, 2026:

• Instant chat. The December notice says it won’t draw from your usage.
• Auto-review safety checks in Codex. When you’re signed in with a ChatGPT account, those checks don’t count toward your plan’s usage limits.
• Regular Chat Voice. It has separate limits and doesn’t consume Codex usage.
• File uploads, image generation and voice limits in ChatGPT. These have their own limits and reset periods, and today’s Codex help page says they don’t apply to Codex.

The last two describe today’s rules. OpenAI may update those pages when the shared allowance goes live, so check them again in December.

Where to see what’s left

You don’t have to guess. OpenAI points to these places:

• Settings → Usage in ChatGPT, which will show the combined remaining usage and reset times from December.
• Settings or the usage dashboard in Codex, where you can see which allowance is used up, your credit balance and any reset time.
• /status in the Codex CLI (the version that runs in your terminal), typed during an active session.
• Per-chat credit usage in ChatGPT Desktop, where supported on credit-based plans. It covers individual Work and Codex chats, not regular Chat.

Your credit balance shows in Settings → Usage in ChatGPT or Usage & Billing in the Codex app.

When you hit the limit

First, the good news if you’re mid-task. On plans with included allowances or credits, OpenAI says Codex can keep working on the turn it was already doing when you hit the limit, subject to fair-use limits. After that turn, you have four options.

1. Wait. Your allowance refills at the reset time shown on the usage page. OpenAI Support won’t reset it for you, and starting a new chat doesn’t reset it either.
2. Buy credits. Your included usage is spent first, and credits cover eligible use after that. They’re valid for 12 months from purchase and are generally non-refundable. If a task finishes after your balance runs out, the balance can go negative, and your next purchase pays that off first.
3. Buy an instant reset (eligible Plus and Pro personal accounts). It restores both the 5-hour and the seven-day allowance right away. OpenAI says it pulls your normal allowance for the week forward rather than adding extra usage on top (OpenAI Help Center: paid Work and Codex rate limit resets). It applies the moment you buy it, can’t be saved for later, and moves your next automatic reset to seven days after your first Work or Codex request.
4. Upgrade your plan. Your plan tier sets how much is included.

One setting to look at before you lean on credits is automatic reload. If it’s on, OpenAI buys credits for you whenever your balance drops below a minimum you choose, up to any monthly cap you set.

Before December

A few small habits make the switch less of a surprise:

1. Open Settings → Usage once now, so you know what “normal” looks like for you before the meters merge.
2. Use Instant chat for quick questions, since OpenAI says it won’t draw from the allowance.
3. Save the most capable model for the tasks that need it. Model choice is the first thing on OpenAI’s list of what drives usage.
4. If you buy credits, decide on automatic reload on purpose, and set a monthly limit if you turn it on.
5. Check the usage limits page again in December for the final details.

Chef’s pick: look at Settings → Usage once now and again after the switch. Two readings of your own meter will tell you more than any prediction about how fast the shared allowance drains.