ChatGPT and Codex Shared Usage Limits: What's New, How They Work, and Alternatives (2026)
OpenAI is changing how usage limits work on its paid plans. Starting in December 2026, the usage allowance on Plus, Pro, and Business plans will be shared across ChatGPT and Codex. Usage will be counted in tokens, not messages.
The change matters most to people who use ChatGPT for long tasks, such as research reports, slide decks, and spreadsheets. Those tasks will draw from the same pool as coding work in Codex. This guide covers what OpenAI has announced, how the shared limits work today, what to check before December, and the alternatives. Every detail comes from vendor pages read on November 4, 2026.
What OpenAI has announced
OpenAI announced the change in a help center article titled "Understanding ChatGPT and Codex usage limits." News coverage dates the article to the end of October. On November 4, the page showed it was last updated two days earlier.
The article is short. Its main points are these:
- One allowance for two products. OpenAI says the "usage allowance will be shared across ChatGPT and Codex." This applies to Plus, Pro, and Business plans.
- Token-based measurement. "Your usage for ChatGPT and Codex will be measured in tokens," the article says. Tokens are the units a model reads and writes.
- Heavier work costs more. "More advanced models and complex tasks generally use more of your allowance."
- Instant chat is exempt. OpenAI states that "Instant chat will not draw from your usage."
- One place to check. You will see your combined remaining usage and reset times in Settings, then Usage.
OpenAI also gives the reason. The change comes "as we bring the agentic capabilities of Work into the main ChatGPT experience." ChatGPT Work is the mode for longer research and deliverable tasks. Today, Work sits apart from regular chat.
What has not been announced yet
The article leaves several questions open. As of November 4, OpenAI has not published:
- A start date. The article says "Starting in December" without a day.
- The size of each plan's allowance. No token totals are listed for Plus, Pro, or Business.
- Token rates per model. The article says advanced models use more, but gives no ratio.
- Changes to Free and Go. The shared pool is described for Plus, Pro, and Business only.
OpenAI does offer one reassurance. It says, "We expect most users won't need to wait for their usage to reset more often than they do now." Treat that as an expectation, not a guarantee. Watch Settings, then Usage, during your first week in December.
How ChatGPT and Codex usage limits work today
Part of this pool already exists. OpenAI's guide to GPT-6 Astra says "Work and Codex share the usage allowance included with your plan." The pricing docs add that Work "uses the same pricing, credits, and usage limits as Codex."
So the December change mostly adds regular ChatGPT chat to a pool that Work and Codex already share. Instant chat stays outside it.
Here is how the current Work and Codex usage limits behave, according to OpenAI's help pages:
- Two windows. Depending on the plan, limits can apply over five hours and over a week. You need allowance left in both to keep working.
- No fixed message count. "Your allowance does not correspond to a fixed number of messages or tasks."
- Model choice matters. Larger inputs, higher reasoning settings, Fast mode, and multi-step tasks all use more.
- Switching models does not refill the pool. OpenAI says, "Switching models does not restore allowance in a shared usage pool."
- Pro has no five-hour window today. "Pro 100, Pro 200 and Pro 500 currently have no five-hour usage limit in Work and Codex." A weekly allowance still applies.
OpenAI also publishes estimated local messages per five-hour window for Plus and Standard Business seats:
| Model | Plus, per five hours | Standard Business, per five hours |
|---|---|---|
| GPT-6 Astra | 5–45 | 5–45 |
| GPT-6.1 Sol | 15–160 | 15–160 |
| GPT-6 Sol | 15–150 | 15–150 |
| GPT-6 Luna | 350–3,000 | 350–3,000 |
OpenAI says "These ranges are estimates, not guaranteed message counts." Cloud tasks can use more than local messages.
What draws from the pool, and what does not
OpenAI's help pages name several items that sit outside the shared allowance today, and a few that use more of it.
Outside the pool:
- Instant chat. The December article says it "will not draw from your usage."
- Auto-review checks in Codex. OpenAI says these safety checks "are free and don't count toward your plan's usage limits" when you sign in with ChatGPT.
- Composer predictions. This Codex beta suggests your next message. OpenAI's release notes say predictions don't count toward Codex usage limits during the beta.
- API usage. "API usage is billed separately at API rates."
- GPT-6 Pro in Chat, for now. The Astra guide says its message limits "are separate from the shared Work and Codex allowance." OpenAI has not said whether that changes in December.
Uses more of the pool:
- Fast mode. OpenAI says it "uses more of your included allowance."
- Ultrafast on Pro 500. It uses included usage "at 8× the Standard rate for the selected model." Other self-serve plans do not include Ultrafast at launch.
- Cloud tasks. These "may use more allowance than local messages."
Plans and prices
OpenAI's Codex pricing docs list these plans on November 4. The shared pool in December covers Plus, Pro, and Business.
| Plan | Price on OpenAI's page | Shared pool from December |
|---|---|---|
| Free | $0/month | Not named in the article |
| Go | $8/month | Not named in the article |
| Plus | $20/month | Yes |
| Pro | From $100/month (tiers at $100, $200, or $500) | Yes |
| Business | $20 per user per month billed annually, or $25 billed monthly | Yes |
| Enterprise and Edu | Contact sales | Not named in the article |
Business has two seat types. OpenAI says "Premium seats offer more included usage for ChatGPT Work and Codex." The Business help page says individual chat model allowances are separate today. How that changes in December is not yet stated.
Pro limits also moved recently. On October 29, OpenAI's release notes said "Pro 100 and Pro 500 plans now include twice their previous Work and Codex allowance." The same note cancelled a planned cut to Pro 200.
What you can do when you hit the limit
OpenAI lists three options. "If you reach your limit, you can wait for it to reset, upgrade your plan, or buy credits to continue using eligible features."
The Astra guide adds two more for some accounts:
- Banked resets. These are saved resets that restore your allowance when you apply them. You use them from Settings, then Usage, in the desktop app.
- Purchased instant resets. Eligible personal Plus and Pro accounts can buy one. It refreshes the five-hour and weekly Work and Codex allowance right away.
A purchased reset cannot be saved for later. It also moves your weekly period forward, so check the terms before you buy.
Who feels the change most
Most people who only chat will notice little, especially if they use Instant. The people to watch are those who mix heavy work types on one account.
Developers who also write in ChatGPT. A long Codex session and a long research chat now draw from one pool. A busy coding day can leave less room for analysis later.
Analysts who build deliverables in Work. Reports, decks, and spreadsheets are long, multi-step tasks. OpenAI says these "generally use more of your allowance."
Teams on Business Standard seats. Several people may plan their week around separate limits today. In December, each person's chat and Codex use come from one allowance.
Anyone who picks the strongest model by default. GPT-6 Astra gets 5 to 45 local messages per five hours on Plus. A lighter model can stretch the same allowance much further.
How to prepare before December
You can take a few steps now to avoid surprises.
- Check your current usage. Open Settings, then Usage, and note your weekly pattern for two weeks.
- Match the model to the task. OpenAI's Astra guide suggests GPT-5.6 Luna for "extracting information, categorizing items, or making short edits."
- Lower reasoning effort first. OpenAI notes that "Astra at Low effort can outperform Sol at High effort."
- Use Instant for quick questions. Instant chat will not count against the pool.
- Split recurring deliverables out. If a weekly report eats a large share, consider a tool built for that job.
- Plan for credits. If you often hit the limit, compare the cost of credits with an upgrade.
Limits to keep in mind
A few gaps remain on November 4:
- Numbers are missing. Without token totals, you cannot calculate your December allowance in advance.
- Estimates are wide. A range of 5 to 45 messages makes planning hard for heavy users.
- Mixed signals on Pro. Pro has no five-hour window today, and OpenAI has not said whether that changes.
- Third-party coverage varies. Some outlets report different update dates. Use OpenAI's own article as the reference.
Alternatives to ChatGPT and Codex for report and slide work
If your concern is the deliverable side, such as reports, slides, and spreadsheets, other tools have their own limits and prices. These come from each vendor's page on November 4.
| Tool | How usage is described on the vendor page | Price, per the vendor |
|---|---|---|
| Claude | Free has "Limited usage every 5 hours"; Max has 5x Pro usage; Ultra has "No 5-hour limit" | Pro $17 per month with annual subscription, $20 if billed monthly; Max $100; Ultra $200 |
| Gemini app | Google AI Pro gives four times more Gemini access than non-AI subscribers | Google AI Pro $19.99/mo or $199.99/year; Ultra $99.99/mo |
| Powerdrill Bloom | Report and slide tasks from data files; scheduled tasks by plan | Free plan; Pro $13.27/mo |
Claude sets limits by plan and window. Its pricing page also lists Claude Design, Slides, and Docs on the Free plan.
The Gemini app creates and edits files in the chat. Google AI Pro is billed monthly or yearly, and Ultra is billed monthly only.
For a wider list, see our roundup of ChatGPT Work alternatives for reports, slides, and spreadsheets.
Where Powerdrill Bloom fits
Powerdrill Bloom focuses on the deliverable side of this work. You upload Excel, CSV, or PDF files and get analysis, reports, and slides.
The AI Report Generator page says, "Simply upload your data file, and let the AI do the rest." The Excel to PPT tool says it can "Create professional PowerPoint slides in seconds from any topic, Excel, CSV or PDF." The homepage adds that "Every number comes back with the page, the row and the figure behind it."
This split can help in December. Keep Codex for code, and move recurring reports to a separate tool. Your coding days then stop competing with your monthly deck. The pricing page lists a Free plan at $0.00 with 1 scheduled task, and Pro at $13.27/mo with 20 scheduled tasks.
Try Powerdrill Bloom with the report you rebuild most often.
Frequently asked questions
When do ChatGPT and Codex start sharing usage limits?
OpenAI says the change starts in December 2026. It has not named a specific date. The shared allowance applies to Plus, Pro, and Business plans.
Does Instant chat count toward the shared usage limits?
No. OpenAI states that "Instant chat will not draw from your usage." Thinking models, Work tasks, and Codex sessions do draw from the pool.
How do I check my remaining ChatGPT and Codex usage?
Open Settings, then Usage. OpenAI says you will see your combined remaining usage and reset times there. The same page already shows Work and Codex usage today.
What happens when I reach my limit?
You can wait for a reset, upgrade your plan, or buy credits. Some Plus and Pro accounts can also use banked resets or buy an instant reset. Check the options shown in Settings, then Usage.
Are Free and Go plans affected?
OpenAI's article names Plus, Pro, and Business only. Codex is included on Free and Go with their own usage limits. OpenAI has not said whether those plans change in December.
Sources: OpenAI Help, Understanding ChatGPT and Codex usage limits · OpenAI Help, Managing usage with GPT-6 Astra in Work and Codex · OpenAI Help, Using Codex with your ChatGPT plan · OpenAI Help, ChatGPT Business models and limits · ChatGPT release notes · Codex pricing · Claude pricing · Google AI plans · Google Workspace Updates, Create, preview, and edit files directly in the Gemini app. Details read from vendor pages on November 4, 2026.