ChatGPT and Codex Usage Limits Explained (Nov 2026)

On October 29, 2026, OpenAI doubled the Work and Codex allowance on Pro 100 and Pro 500. It also cancelled a planned cut to Pro 200.

A bigger change lands in December, when ChatGPT and Codex start sharing one usage pool. This guide explains every plan’s limits, how to check what you have left, and how to make it last.

TLDR

Plans: Free ($0), Go ($8), Plus ($20), Pro 100, Pro 200 and Pro 500, Business (from $20 per seat), Business Premium, and Enterprise/Edu.
October 29 change: Pro 100 and Pro 500 got twice their previous Work and Codex allowance. Business Premium was also raised.
Pro 200: the cut announced in a September 29 email is cancelled. Accounts on the lower allowance were doubled to match the higher one.
Plus: unchanged. It still has a 5-hour window plus a weekly limit for Work and Codex.
From December: Plus, Pro and Business share one token-based allowance across ChatGPT and Codex. Instant chats do not count.
Check usage: Settings → Usage in ChatGPT, or /status in the Codex CLI.

ChatGPT and Codex limits by plan

OpenAI sells seven consumer and team plans right now. Everyday text chat is listed as unlimited on every plan, subject to abuse guardrails. The real differences are in models, files, Work and Codex.

OpenAI does not publish exact numbers for most caps. Where it only says “limited” or “expanded”, the table says so too.

Plan	Price	Main models	Work and Codex	File uploads
Free	$0	GPT-6 Luna, GPT-5 Thinking Mini	Limited, desktop app only	3 per day
Go	$8/month	GPT-6 Luna, GPT-5 Thinking Mini	Limited, desktop app only	80 per 3 hours
Plus	$20/month	GPT-6.1 Sol, GPT-6 Astra, GPT-6 Sol, GPT-6 Luna	5-hour window plus weekly limit	80 per 3 hours
Pro 100	$100/month	All Plus models plus Pro models	Weekly limit, no 5-hour window, doubled Oct 29	80 per 3 hours
Pro 200	$200/month	All Plus models plus Pro models	Weekly limit, no 5-hour window, cut cancelled	80 per 3 hours
Pro 500	$500/month	All Pro models plus Ultrafast mode	Weekly limit, no 5-hour window, doubled Oct 29	80 per 3 hours
Business	$20/seat/month yearly ($25 monthly), 2+ users	Same family as Plus	Same per-seat Codex ranges as Plus; can add credits	Not published
Business Premium	Per seat, higher tier	Same family as Plus	Raised on Oct 29	Not published
Enterprise / Edu	Contact sales	Workspace settings decide	Flexible pricing plans have no fixed rate limits	Not published

A few more caps matter in daily use. These come from OpenAI’s pricing page and help articles:

Context window: Go and Plus get 256K tokens with a reasoning model. Pro gets 400K tokens.
Files per project: 5 on Free, 25 on Go and Plus, and 40 on Pro.
Voice: Go and Plus get 3 hours per rolling 24 hours. Pro 100 gets 15 hours.
Ultrafast speed: only Pro 500 and eligible Enterprise and Edu plans.

For a feature by feature view of each plan, see my ChatGPT Free vs Go vs Plus vs Pro comparison.

The October 29, 2026 change explained

OpenAI posted the update in its ChatGPT release notes under the title “More Work and Codex usage on Pro and Business Premium Plans.” Thibault Sottiaux, who leads Codex and core ChatGPT work, shared it on X as “Day 25” of a 28-day run of improvements.

What doubled

Pro 100 and Pro 500: the Help Center says these plans “now include twice their previous Work and Codex allowance.”
Business Premium: the release notes say the extra Work and Codex usage applies here too.
Prices: unchanged at $100, $200 and $500 per month.
Timing: effective the same day. Existing subscribers did not need to renew or resubscribe.

What happened to Pro 200

Pro 200 had two allowance levels in October. Since late September, new subscribers were placed on a lower allowance. A September 29 email told existing subscribers they would move to that lower level after October 29.

That cut is now off. The Help Center says “The reduction announced in our September 29 email is canceled.” It also says that on Pro 200 “only the lower allowance doubled”, so anyone on the lower level was lifted to match the higher one.

Quick answer: If you pay $200, your Work and Codex allowance is now the higher Pro 200 level, whenever you joined. Pro 100 and Pro 500 users got twice what they had before.

What did not change

Plus: its Work and Codex limits stay the same.
Standard Business seats: not named in the update. Their Codex ranges still match Plus.
Absolute numbers: OpenAI still does not publish a fixed message count for any Pro tier. It only describes the change as a multiple of the old allowance.

Note: Dots, OpenAI’s always-on personal agents, use a separate allowance. Sottiaux said their extended limits continue “for now”, which means they are not promised forever.

Codex messages per 5 hours by model

For Plus and standard Business seats, OpenAI’s Codex pricing page gives ranges of local messages per 5-hour window. The range is wide because a short question costs far less than a long task that reads a big repo.

Model	Local messages per 5 hours (Plus)	Credits per 1M tokens (input / cached / output)
GPT-6 Astra	5 to 45	250 / 25 / 1,250
GPT-6.1 Sol	15 to 160	50 / 2.5 / 250
GPT-6.1 Terra	35 to 285	25 / 1.25 / 125
GPT-6 Sol	15 to 150	50 / 5 / 250
GPT-6 Luna	350 to 3,000	2.5 / 0.25 / 12.5

OpenAI calls these estimates, not fixed limits. A weekly limit may also apply, and cloud tasks can use more than local messages.

Speed modes cost more

Faster output burns your allowance faster. Here are the multipliers on included plan usage:

Standard: 1x.
Fast: 2.5x on plan usage, or 2x when paying with credits.
Ultrafast (GPT-6 Astra or GPT-6.1 Sol): 8x on plan usage, or 6x with credits. Pro 500 only.

So a doubled Pro allowance used only in Fast mode gives you less Standard-equivalent work than you might expect. Use Fast when time matters, not by default.

Shared ChatGPT and Codex limits from December

Today, Chat, Work and Codex mostly have their own limits. Plus users have had a 5-hour window shared across Work and Codex since August 25, 2026, on top of the weekly cap.

On November 2, 2026, OpenAI updated its Help Center to describe the next step. Starting in December, the “usage allowance will be shared across ChatGPT and Codex” on Plus, Pro and Business.

How the shared pool works

Measured in tokens: how much you use depends on the model you choose and the task you give it.
Heavier models cost more: a long Codex session leaves less room for advanced chats, and the reverse is also true.
Instant chat is free: OpenAI says “Instant chat will not draw from your usage.”
One meter: Settings → Usage will show your combined remaining usage and reset times.

OpenAI says it expects most users won’t wait for resets more often than they do now. It has not given the exact December start day, the pool size per plan, or how Free and Go fit in.

Warning: If you lean on Codex heavily, plan for December. A big refactor could leave less allowance for GPT-6 Astra chats later that week.

How to check your remaining usage

There are a few places to look. They all read from the same account data.

Open ChatGPT settings. Click your profile picture, then Settings, then Usage. You will see your remaining Work and Codex allowance and when it resets.
Open the Codex usage page. Go to chatgpt.com/codex/settings/usage in your browser for the Codex dashboard.
Ask the CLI. In an active Codex CLI session, type /status to see your limits and token totals.
Check deep research. ChatGPT shows a counter of remaining deep research tasks inside the tool.
Watch for notices. Uploads, images and voice have no meter. ChatGPT tells you when you hit them, often with a reset time.

Tip: OpenAI suggests checking the Codex dashboard every week or two. That way a busy week does not surprise you on a Friday.

What happens when you hit the limit

What you see depends on which limit you reach. Here is how each one behaves:

Mid-task in Codex or Work: the agent can finish the turn it is on, subject to fair use limits.
Plus 5-hour window: wait for it to reset. The window starts with your first message after the last one ended.
Weekly limit: wait for the weekly reset, or buy ChatGPT credits on Plus or Pro to keep going.
Thinking models: ChatGPT may switch you to another available model so the chat can continue.
File uploads: wait for the rolling 3-hour window to clear.
Images on Free: wait for the reset time in the notice, or upgrade.

Your options to keep working

Buy credits: Plus and Pro users can buy credits without upgrading. Business, Edu and Enterprise plans with flexible pricing buy workspace credits.
Upgrade: moving from Plus to Pro removes the 5-hour window entirely.
Use an API key: any user can run Codex with an API key, billed per token at normal API rates. See my OpenAI API pricing guide for the rates.

Switching models does not give back what you already used. It only changes how fast the rest goes. OpenAI support will not reset limits on request.

How to stretch your limits

Most limit pain comes from sending too much context. These habits come straight from OpenAI’s own Codex advice and the way the meters work:

Pick the smallest model that works: GPT-6 Luna gets up to 3,000 Plus messages per 5 hours, against 5 to 45 for GPT-6 Astra.
Use Instant for simple chats: from December, Instant chat will not draw from the shared pool.
Stay on Standard speed: Fast costs 2.5x, and Ultrafast costs 8x.
Write tight prompts: give only the files that matter and say the format and length you want.
Trim AGENTS.md: keep it short, and nest smaller files in sub-folders. My AGENTS.md guide shows how.
Turn off unused MCP servers: each one adds context to every request.
Prefer local over cloud tasks: cloud tasks can use more allowance than local messages.
Start fresh chats: a long history is re-read every turn, which costs more tokens.

For more Codex-specific tips and the full app, CLI and IDE picture, read my OpenAI Codex guide.

Troubleshooting usage limit problems

My Pro limit did not double

The change applies automatically, so first refresh the app and sign out and back in. If you are on Pro 200, note that only the lower allowance doubled. If you were already on the higher level, your number stays the same.

I hit a limit after only a few messages

Check the model and speed. A few GPT-6 Astra messages on Fast mode, with large files attached, can use a whole 5-hour window on Plus. Switch to GPT-6.1 Sol or Luna on Standard speed.

Settings → Usage looks empty or wrong

Check the workspace: if you belong to a Business workspace and a personal plan, switch to the right one.
Update the app: older desktop and mobile builds may not show the newest meter.
Use the Codex page: open chatgpt.com/codex/settings/usage as a cross-check.

Codex says “model at capacity” or will not respond

That is often a server-side issue, not your allowance. Check OpenAI’s status page and try again later. My ChatGPT not working fix guide covers the common causes.

I bought credits but Chat is still limited

Credits extend eligible features, such as Work and Codex. They do not lift every cap in Chat, such as file uploads.

Frequently Asked Questions

Did OpenAI double ChatGPT Pro limits on October 29, 2026?

Yes, for Work and Codex. Pro 100 and Pro 500 got twice their previous allowance, and Business Premium was also raised. Prices did not change.

Is the Pro 200 usage cut still happening?

No. OpenAI cancelled the reduction announced in its September 29 email. Pro 200 accounts on the lower allowance were doubled to match the higher one.

Does ChatGPT Plus have a 5-hour limit?

Yes, for Work and Codex. OpenAI brought the 5-hour window back for Plus on August 25, 2026, and a weekly limit also applies. Pro plans have no 5-hour limit.

Do ChatGPT and Codex share the same usage limit?

Starting in December 2026, yes, on Plus, Pro and Business. Usage is counted in tokens across both, but Instant chat does not count.

How do I check my remaining ChatGPT or Codex usage?

Go to Settings → Usage in ChatGPT, or open chatgpt.com/codex/settings/usage. In the Codex CLI, type /status.

Can I pay to keep using Codex after the limit?

Yes. Plus and Pro users can buy ChatGPT credits, and anyone can use an API key billed per token. You can also upgrade to a bigger plan.

Is ChatGPT free text chat really unlimited?

OpenAI lists everyday text chats as unlimited on all plans, subject to abuse guardrails. Advanced models, uploads, images and agents still have caps.

ChatGPT and Codex Usage Limits Explained (November 2026)

ChatGPT and Codex limits by plan