#Which ChatGPT plans include Codex
OpenAI's pricing page opens with the list: Codex is "included in your ChatGPT Free, Go, Plus, Pro, Business, Edu, or Enterprise plan". Free is $0 and Go $8 a month. Plus is $20 and covers Codex "on the web, in the CLI, in the IDE extension, and on iOS". Pro starts at $100, with plans at $100, $200 or $500 a month. Business is $20 a user a month billed annually, or $25 monthly, for two or more users, and Enterprise and Edu are priced through sales.
Two details shape everything else. "ChatGPT Work and Codex share usage", so heavy use of one leaves less of the other. And Codex can also run on an API key, paid at API rates, with "No cloud-based features".
#The five-hour window on Plus and Business
| Model | Local messages per 5 hours (Plus, standard Business) |
|---|---|
| GPT-6 Astra | 5-45 |
| GPT-6.1 Sol | 15-160 |
| GPT-6 Sol | 15-150 |
| GPT-6 Luna | 350-3,000 |
OpenAI's estimates as published on its Codex pricing page, read 21 October 2026. The page adds: "These estimates are not fixed message limits".
The ranges are wide because OpenAI treats them as estimates rather than quotas: how far a window goes depends on what you ask Codex to do. OpenAI publishes these numbers for Plus and standard Business only. The page calls it a five-hour period and does not say whether the window rolls or resets on a schedule.
#Pro, Enterprise and Edu
"Pro plans currently have no five-hour limit." The page gives no Pro message numbers, and the general note that "Weekly limits may also apply" still stands. The $500 Pro plan adds an Ultrafast mode.
Enterprise and Edu workspaces on flexible pricing "have no fixed rate limits. Usage scales with credits." Without flexible pricing they "have the same per-seat usage limits as Plus for most features."
#How to check Codex usage
The usage dashboard. OpenAI's pricing page sends you to chatgpt.com/codex/settings/usage: "You can find your current limits in the usage dashboard", with reset times. OpenAI's own advice is to "Check the dashboard every week or two".
/status in the CLI. The same page says that "during an active Codex CLI session, you can use /status". In the ChatGPT desktop app, /status shows "the chat ID, context usage, and rate limits", and /usage opens usage and billing settings when you are signed in with ChatGPT.
Command output has changed between Codex versions, so if /status in your version shows only session settings, the dashboard is the reliable view. The API dashboard at platform.openai.com tracks API spend only, not plan limits.
#What drains the allowance faster
Four things move the meter faster than a plain local message. Cloud tasks "may use more of your allowance than local messages", and local messages and cloud chats share one allowance. Images use included limits "~3-5x faster on average". The Fast speed mode uses 2.5 times the included usage and Ultrafast 8 times. And the model matters: the same window holds far fewer GPT-6 Astra messages than GPT-6 Luna ones.
#When you run out
Hitting a limit mid-task does not cut the task off: "the agent will be able to continue working on that turn, subject to fair use limits." After that, Plus and Pro "can purchase additional credits to continue working without needing to upgrade", and Business, Edu and Enterprise workspaces on flexible pricing can buy workspace credits. Any plan can also "run extra local chats using an API key" at standard API rates. Otherwise the dashboard shows when the window resets.
Credits are priced per million tokens and vary by model; OpenAI's own example is that "A typical GPT-5.6 Sol task may use 2-15 credits." How Codex's metering compares with Claude Code and Cursor is in Claude Code vs Codex vs Cursor.
#Keeping track on a Mac
Codex has no always-visible meter by default: the numbers live on the dashboard and behind a slash command. Claude Code, by contrast, can print its five-hour and weekly percentages in its status line, as shown in Claude Code usage: how to check it. NotchBay's notch meter reads the local Claude Code logs and does not cover Codex, so for Codex the dashboard tab remains the place to look.
#Frequently asked questions
How many Codex messages do you get on ChatGPT Plus?
OpenAI estimates local messages per five hours by model: 5-45 on GPT-6 Astra, 15-160 on GPT-6.1 Sol, 15-150 on GPT-6 Sol and 350-3,000 on GPT-6 Luna. The range depends on task size, and OpenAI says these are not fixed limits.
Does Codex on ChatGPT Pro have a usage limit?
Pro plans currently have no five-hour limit, according to OpenAI's pricing page. Weekly limits may still apply, and the usage dashboard shows the current state.
Is there a way to keep using Codex after hitting the limit?
Yes: buy ChatGPT credits on Plus or Pro, or run local chats with an API key at standard API rates. Otherwise wait for the reset time shown on the usage dashboard.
#What this means for the notch
Codex's limits live on a web page: a five-hour estimate per model and weekly caps with no published number. The habit OpenAI suggests is checking the dashboard every week or two, and a habit is a weaker tool than a surface that is always in view. The band around the MacBook notch is exactly that kind of surface, which is why agent status keeps ending up there. For Claude Code it already does; for Codex, the dashboard is still a tab you have to open.
