ChatGPT Usage Limits: Where to Check Them and What Changes in December

In short: The ChatGPT usage limit is not a fixed number of messages. It is an allowance that depends on your plan, the model you pick and the size of the task. On Plus, Work and Codex have a 5-hour limit and a weekly limit, and you can see what is left on the Settings → Usage screen. As of 3 November 2026, the OpenAI Help Center says that from December, ChatGPT and Codex will share one token-based allowance on Plus, Pro and Business, while Instant chat will not draw from it.

When a usage warning appears in the middle of a long Codex task, work stops unless you know which limit you hit and when it comes back. And ChatGPT does not have just one limit. Chat, Work and Codex, file uploads and voice each run on their own caps, and in December ChatGPT and Codex start sharing one allowance.

A usage limit sets how much work your plan lets you do in a given period. OpenAI's Understanding ChatGPT and Codex usage limits article states plainly that the allowance does not map to a fixed number of messages or tasks. How much you can do depends on the model, the task and your settings. A long, multi-step task can use far more of it than a short question.

OpenAI changes its limits often. For your own account, the Settings → Usage screen always has the final word.

ChatGPT's limits fall roughly into three groups.

• Everyday chat (Instant): According to OpenAI's ChatGPT Free Tier FAQ, daily text chats are unlimited for Free and Go users, subject to abuse guardrails. On the Business plan, the access OpenAI calls "virtually unlimited" also applies to Instant chat only.
• Thinking and Pro models: The GPT-6 and other models in ChatGPT article says Plus has the Medium and High thinking levels, but not Extra High or Pro. If you reach the thinking limit, ChatGPT may continue with another available model.
• Work and Codex: Long research, document and coding tasks draw from a shared allowance. On Plus, that allowance is capped by a 5-hour window and a weekly window.

File uploads, image generation, voice chat and data analysis have their own limits on top of these. OpenAI's Uploading files and audio to ChatGPT article caps uploads at 80 files every 3 hours, and at 3 files a day on the Free plan. OpenAI also says it may lower these caps at busy times. According to the ChatGPT Voice article, Plus gets 3 hours of Live voice chat in a rolling 24-hour window.

The "Context window full" warning is not a usage limit. According to OpenAI, it only means that the conversation's context window is full, and you can move to a new chat right away. We explain what a context window is in What is an LLM.

From December 2026, ChatGPT and Codex will share one allowance on the Plus, Pro and Business plans, and that allowance will be measured in tokens. OpenAI announced the change in the Understanding ChatGPT and Codex usage limits article without giving an exact date. According to the announcement, this is what changes.

• The allowance on Plus, Pro and Business will be shared between ChatGPT and Codex.
• Usage will be measured in tokens. How much you use will depend on the model and the task, and more advanced models and complex tasks will generally use more.
• Instant chat will not draw from the shared allowance.
• You will see your combined remaining allowance and reset times on the Settings → Usage screen.
• When you reach the limit, you will be able to wait, upgrade your plan or buy credits for eligible features.

OpenAI writes that it expects most users will not have to wait for a reset more often than they do today. A token here is one of the small pieces a language model splits text into. We measured why Turkish text takes more tokens in What is a token in AI?

You check your ChatGPT usage limits on the Settings → Usage screen. According to the Understanding ChatGPT and Codex usage limits article, it shows your remaining allowance and your reset times. Codex has two more places to look.

• Codex app: According to the Using Codex with your ChatGPT plan article, buying credits and the usage settings are on the Usage & Billing screen.
• Codex CLI: Type /status in an open session to see your usage.

The file upload quota does not appear on these screens. The Uploading files and audio to ChatGPT article says ChatGPT does not currently show how much of your rolling upload quota you have used. You can see your saved file storage on the Settings → Storage screen.

When a ChatGPT limit resets depends on the type of limit.

• The 5-hour window (Work and Codex): A new window starts with the first message you send to Work or Codex after the previous window ends. You can use up the window's allowance before the 5 hours are over.
• The weekly window: It sets how much work you can do over your weekly usage period. When both windows apply, you need allowance left in both to continue, so you can hit the 5-hour limit while you still have weekly allowance.
• Thinking and other models: ChatGPT shows when your allowance resets, when that information is available.
• Voice chat: Live usage is measured in a rolling 24-hour window.

There are three ways to keep going without waiting for a reset.

1. A paid instant reset: According to OpenAI's Paid weekly Work and Codex rate limit resets article, eligible personal Plus and Pro accounts can buy an instant reset from the usage settings in the ChatGPT desktop app, or from the prompt that appears when they reach the weekly limit. The option varies by account and billing country, and no prompt appears when you reach only the 5-hour limit. Once payment goes through, both the 5-hour and the weekly allowance reset at once. The reset cannot be saved for later, and it changes your weekly schedule. The new weekly period starts with your first Work or Codex request after the reset, and the next automatic weekly reset comes 7 days after that. The option is not available on Free, Go, Business, Enterprise or Edu.
2. A banked reset: If your account has a banked reset, you can use it from the Settings → Usage screen in the desktop app. A full banked reset renews both of Codex's windows, the 5-hour and the weekly one, and changes your weekly reset date. OpenAI also applies automatic resets from time to time. On 7 September 2026, the limits of eligible Plus, Pro and Business accounts were reset automatically because the broader GPT-6 Astra launch was delayed. No reset raises your limit permanently.
3. Credits: On personal plans, once the plan's included allowance runs out, eligible features can keep running by drawing on credits you have bought.

You cannot get your ChatGPT limit reset by writing to support, because OpenAI support does not reset ChatGPT or Codex limits. Switching models in a shared allowance does not bring the allowance back, and starting a new chat in Codex does not reset the limit either. Moving from the Free plan to Plus, Pro or Business, on the other hand, does reset your Free plan limits.

Codex usage comes out of an allowance shared with ChatGPT Work, and on Plus it is capped by a 5-hour window and a weekly window. Codex is OpenAI's coding agent. According to the Using Codex with your ChatGPT plan article, Codex is available on every ChatGPT plan, Free and Go included. Codex Cloud, which runs in the cloud, is available on eligible Plus accounts, every Pro tier, Business, Enterprise, Healthcare and Education, but not on Free or Go. We explain how AI agents work in What is an AI agent? and how they split up a job in One agent or many?

These are the points to know about Codex limits.

• Shared pool: Codex, ChatGPT Work, ChatGPT for Excel and Workspace Agents, where your plan has them, use the same allowance and credit pool.
• What raises usage: the model, where the task runs, complexity, context, thinking level, speed and tools. Cloud tasks can use more than local messages, and Fast mode also uses up the included allowance faster.
• What does not count: Auto-review checks do not count against the limit.
• Separate limits: ChatGPT's file upload, image generation and voice limits do not apply to Codex.
• Hitting the limit mid-task: If you reach the limit during a running turn, Codex can finish that turn within fair-use limits. After that, your options are credits, a reset, an upgrade or waiting.

For Plus, OpenAI estimates 5 to 45 local Codex and Work messages per 5-hour window with GPT-6 Astra, and 350 to 3,000 with GPT-6 Luna. The ranges in the table come from the Managing usage with GPT-6 Astra in Work and Codex article.

OpenAI writes that these ranges are estimates, not guaranteed message counts, and that weekly limits may also apply. The ranges are wide because the same model uses little on a small fix and a lot on a long task. According to the same article, large inputs and outputs, higher thinking levels, Fast mode and multi-step tasks use more. A lower thinking level makes the allowance last longer.

ChatGPT Plus has no single fixed limit. Work and Codex have 5-hour and weekly windows, chat has per-model limits, and voice and files have separate caps. According to OpenAI's What is ChatGPT Plus? article, Plus costs $20 a month and offers higher model limits and priority access compared with the Free plan. It is billed monthly, with no annual option. The Help Center gives no price in Turkish lira. The same article says Plus subscriptions may have limits such as message caps, especially at busy times, and that these may change with system conditions. API usage is not included in Plus and is billed separately.

This is what applies on Plus today.

• Chat: GPT-6 is rolling out to eligible Plus users with the Medium and High thinking levels, so it may not show on every account yet. Extra High and the Pro level are not included in Plus, and GPT-6 Pro is not available in Chat.
• Work and Codex: GPT-6 Astra is available in Work and Codex. The 5-hour and weekly windows apply, and GPT-6 Astra allows an estimated 5 to 45 local messages per 5 hours.
• Voice: 3 hours of Live voice chat per 24 hours.
• Projects: up to 25 files per project.
• Extra usage: Eligible accounts can buy credits or an instant reset.

ChatGPT credits let you keep using eligible features on personal plans after the included allowance runs out. According to OpenAI's Using Credits for Flexible Usage in ChatGPT (Personal plans) article, your plan's included allowance is used first, and credits only kick in once you hit the limit. Credits are valid for 12 months after purchase and do not work as API credits. We cover the monthly API credits that come with a ChatGPT Business plan in our ChatGPT Business API credits guide.

On ChatGPT Business, limits depend on the seat type. On a Standard seat, the Work and Codex allowance runs with a 5-hour limit, and on a Premium seat the allowance is higher and there is no 5-hour limit. OpenAI's ChatGPT Business models and limits article lists these differences between the two seats.

For estimated Work and Codex messages on Business Standard seats, two OpenAI articles give different numbers. The Managing usage with GPT-6 Astra in Work and Codex article gives 15 to 160 messages per 5 hours for GPT-6.1 Sol, while the ChatGPT Business models and limits article gives 15 to 150 messages for GPT-6.1 Sol and GPT-6 Sol together on Standard seats.

In Chat, Business seats get a limited number of GPT-6 Pro messages, while Plus has no GPT-6 Pro. Instant chat access on Business is, in OpenAI's words, virtually unlimited. When the included allowance runs out, credits the workspace has bought kick in, if spending controls allow it. Instant resets cannot be bought on Business, Enterprise or Edu. If you want the same comparison for Anthropic, we compare Claude's plans and usage limits in Claude pricing 2026.

If you are buying ChatGPT Business for your team, your seat and credit policy should answer four questions that follow from the Help Center's rules.

• Who gets a Premium seat? Premium gives a higher Work and Codex allowance with no 5-hour limit. It makes sense for people who use Codex or long Work tasks every day. Model limits in Chat are defined separately, so Premium does not mean higher limits everywhere.
• Who approves credit spending? Once the included allowance runs out, extra usage is charged to workspace credits. Set the spending controls and decide who can use credits in advance. If your teams use models through the API rather than ChatGPT seats, the same approval question applies there. RuneLab's RuneGate routes that traffic through one OpenAI-compatible endpoint, gives each team a monthly budget, allowed model tiers and a request limit, stops requests before they reach the provider once the budget runs out, and shows spending per team in real time. It does not manage ChatGPT seat limits.
• How do you prepare for December? From December, Chat, Work and Codex will draw from the same allowance, while Instant chat will not draw from it. Since advanced models and complex tasks will use more, write a rule on which team uses which model and which thinking level.
• What happens when you downgrade a seat? When a seat moves from Premium to Standard, Standard limits apply from the effective date.

The key question behind a model and thinking-level rule is which task needs which mode. RuneLab's Enterprise AI Assistants for Productivity course opens with that question. It works through the difference between standard chat, reasoning and deep research modes, and how to choose a model by task type, in ChatGPT Enterprise, Gemini and Claude. Its last module covers human review of critical output and a shared working standard for the team.

In ChatGPT, check the Settings → Usage screen, which shows your remaining allowance and reset times. In the Codex CLI, you can type /status in an open session. The file upload quota does not appear on that screen.

When a ChatGPT limit resets depends on the type of limit. The 5-hour window in Work and Codex starts again with the first message you send after the previous window ends, and the weekly limit applies separately. For other limits, ChatGPT shows the reset time when that information is available.

OpenAI support does not reset ChatGPT or Codex limits. Eligible personal Plus and Pro accounts can buy an instant reset, and if your account has a banked reset, you can use that too. Moving from the Free plan to a paid plan resets your Free limits.

ChatGPT Plus has no fixed message count. For Work and Codex on Plus, OpenAI estimates 5 to 45 local messages per 5 hours with GPT-6 Astra, and 350 to 3,000 with GPT-6 Luna. These are estimates, and a weekly limit may also apply.

Right now, Codex and ChatGPT Work share the same allowance. According to OpenAI, from December 2026 ChatGPT and Codex will also share one allowance on the Plus, Pro and Business plans. Instant chat will not draw from it.

Free ChatGPT has unlimited daily text chats, subject to abuse guardrails. File uploads (3 files a day), image generation, voice and data analysis have separate limits. Codex is available on the Free plan too, but Codex Cloud is not.

Work and Codex in ChatGPT have 5-hour and weekly windows, not a daily limit. Voice chat is measured in a rolling 24-hour window. The Free plan does have a daily cap on file uploads: 3 files a day.

First check the Settings → Usage screen to see which limit you hit and when it resets. Your options are to wait, buy credits or an instant reset where eligible, or upgrade your plan. Switching models does not bring back allowance in a shared pool.

• OpenAI, Understanding ChatGPT and Codex usage limits: what the limit is, the December 2026 change, the Settings → Usage screen, the 5-hour and weekly windows, and what does not reset a limit
• OpenAI, Managing usage with GPT-6 Astra in Work and Codex: estimated message ranges per 5 hours on Plus and what raises usage
• OpenAI, GPT-6 and other models in ChatGPT: thinking levels on Plus, the GPT-6 rollout to Plus and the model switch at the thinking limit
• OpenAI, Using Codex with your ChatGPT plan: which plans include Codex and Codex Cloud, the shared pool, Usage & Billing and /status
• OpenAI, ChatGPT Business models and limits: Standard and Premium seat differences, Instant chat access, the Standard seat message range and seat downgrades
• OpenAI, What is ChatGPT Plus?: the Plus price, billing, limits at busy times and separate API billing
• OpenAI, Paid weekly Work and Codex rate limit resets: instant resets, banked resets and the automatic reset of 7 September 2026
• OpenAI, Using Credits for Flexible Usage in ChatGPT (Personal plans): the order in which credits are used, how long they last and how they differ from API credits
• OpenAI, ChatGPT Free Tier FAQ: daily chat and other limits on the Free plan
• OpenAI, Uploading files and audio to ChatGPT: file upload caps, the upload quota and Settings → Storage
• OpenAI, ChatGPT Voice: Live voice chat time

• ChatGPT Business and Pro API credits
• Claude pricing: Pro, Max, Ultra and Team
• What is an LLM? Large language models explained
• One agent or many? Multi-agent systems
• Course: Enterprise AI Assistants for Productivity