What to do when Codex hits its usage limit: 9 ways to keep coding in 2026

Disclosure: Enjoy is our product. We checked every price and limit below against OpenAI’s, Anthropic’s, Ollama’s and OpenCode’s own pages on October 27, 2026, and we say where another way is the better pick, including the free ones that don’t need Enjoy.

The ways to keep coding, compared

When Codex runs out, it says “You’ve hit your usage limit.” In the Codex CLI, the message goes on to say when to “try again,” and on Plus it suggests you “Upgrade to Pro” or “purchase more credits.” OpenAI’s usage guide explains the two limits behind it: “Depending on your plan, usage limits may apply over a five-hour window and a weekly window. Where both apply, you need allowance remaining in both to continue.” Codex shares them with ChatGPT Work, and Pro plans have only the weekly one. What to do depends mostly on which limit you hit and how far away the reset is.

Way	Cost, checked October 27, 2026	How soon you’re back	Best for
Wait for the reset	Free	At the reset: up to five hours, or up to a week for the weekly limit	A five-hour limit that resets soon
Use a free reset	Free, when OpenAI gives you one	Right away	Anyone with one in Settings > Usage
Pay as you go	Credits, at the price shown for your account; or an API key, where GPT-6.1 Sol is $2 in and $10 out per million tokens	Right away	Finishing the task in front of you
Buy an instant reset	The price shown at checkout	Right away	A weekly limit you want back today
Upgrade to Pro	Pro 100 $100 a month; Pro 200 $200; Pro 500 $500	Right away	Running out most weeks
A second ChatGPT account	Another plan: Plus is $20 a month	Right away	Projects that belong on another plan you have
Claude Code, on a Claude plan	Claude Pro $20 a month, or $17 billed yearly; Max from $100	Right away, in a new conversation	A second lab’s agent
A local model, inside Codex	Free, on your own computer	Once it’s downloaded	Routine work with no limit at all
Enjoy	Free; Enjoy Cloud adds your phone, other computers and teammates (pricing)	Right away, in the same conversation	Moving a conversation to another account or agent

Prices are each company’s US list prices, before tax. Waiting, a free reset and a local model cost nothing extra; the other ways add to what you pay for ChatGPT.

Wait for the reset

Best for: a five-hour limit that resets soon, when you can step away.

A task that’s running when the limit hits can carry on. Codex’s pricing page says “If you reach your usage limits during an active turn, the agent will be able to continue working on that turn, subject to fair use limits.” After that, you wait until the time the message gave.

To see your limits, run /status in the Codex CLI, which shows “your remaining limits during an active Codex CLI session,” or open Settings > Usage in ChatGPT. On Plus, “A new window starts when you send your first message in Work or Codex after the previous window ends,” and OpenAI warns that “You may reach the five-hour limit before five hours have passed, even if you still have weekly usage remaining.”

Price: free.

Watch for: on Pro, the limit you hit is the weekly one, which can be days away: “Pro 100, Pro 200 and Pro 500 currently have no five-hour usage limit in Work and Codex.” And OpenAI’s message counts “are not fixed message limits.” Cloud tasks “may use more of your allowance than local messages.”

Use a free reset, if you have one

Best for: anyone OpenAI has given one to.

OpenAI hands out free resets now and then, including through referrals. Its help center calls a banked reset “A free, one-time reset saved to your account for later use.” A full one restores “The 5-hour and weekly allowances” on Plus, and “The weekly allowance” on Pro. In the desktop app, open Settings > Usage & billing and select “1 reset available” or “Full reset.” “In Codex CLI, enter /usage and select Redeem reset.” You can also turn on Allow your assistant to use resets, which lets Codex use one when your five-hour or weekly allowance has “10% or less remaining.”

Price: free.

Watch for: “Free banked and automatic resets are discretionary promotional benefits, with no guaranteed availability or schedule.” To use one from Usage settings, “at least one allowance it covers must have less than 80% remaining,” and a full reset “changes your weekly reset date.” Resets expire, and support won’t “extend expiration dates.”

Pay as you go: credits or an API key

Best for: finishing the task in front of you today, without changing plans.

“ChatGPT Plus and Pro users who reach their usage limit can purchase additional credits to continue working without needing to upgrade their existing plan,” says Codex’s pricing page. When Codex stops, “you’ll see an option to add credits,” per OpenAI’s credits article, or buy them in Settings > Usage in ChatGPT or Usage & Billing in the Codex app, where you can also turn on automatic reload. “Your plan’s included usage is used first. After you hit plan limits, usage draws from your credit balance.”

The other route is an API key: “All users may also run extra local chats using an API key, with usage charged at standard API rates.” In the CLI, Codex’s authentication docs sign in with printenv OPENAI_API_KEY | codex login --with-api-key.

Price: “Credit purchase prices and applicable discounts depend on your plan or agreement,” so the price is the one shown for your account. Credits are spent by the token: GPT-6.1 Sol uses 50 credits per million input tokens and 250 per million output, GPT-6.1 Terra half that, and GPT-6 Astra five times as much. On an API key, OpenAI’s API pricing lists GPT-6.1 Sol at $2 per million input tokens and $10 per million output, and Terra at $1 and $5.

Watch for: “Credits are non-refundable unless required by law or contract,” and they’re “valid for 12 months from purchase.” On Free and Go, buying credits is “currently available to a limited group.” An API key has “No cloud-based features (GitHub code review, Slack, etc.),” and “When you sign in with an API key, Codex uses standard API pricing instead of included ChatGPT plan credits,” until you sign in with ChatGPT again.

Buy an instant reset

Best for: a weekly limit you want back today, when you’ll use the whole week it starts.

“Customers on ChatGPT Plus and Pro plans can buy an instant reset from Usage settings in ChatGPT Desktop before reaching a limit, or from an in-app offer after reaching the weekly limit,” says OpenAI’s help center. It “immediately restores both 5-hour and weekly usage.”

Price: OpenAI’s help center doesn’t list one. You “Review the price and payment method in checkout.”

Watch for: it isn’t extra usage. It “pulls your normal weekly allowance forward rather than adding a separate usage entitlement,” and “Your next automatic weekly reset is scheduled 7 days after” your first request afterward. It “cannot be saved for later,” it’s “not available on Free, Go, Business, Enterprise, or Edu plans,” and “Generally, we do not offer refunds for weekly limit reset purchases.”

Upgrade to Pro

Best for: running out most weeks, not once in a while.

Pro comes in three sizes, and none has a five-hour limit, so you can spend the week’s allowance when you like. OpenAI’s help center only ranks them: “Pro 200 includes more usage than Pro 100. Pro 500 offers the highest included usage of the three plans.” The multiples come from OpenAI’s Thibault Sottiaux, who listed “Plus = 1X Pro 100 = 5X Pro 200 = 10X” on September 29, and from OpenAI’s email to Pro 200 subscribers, which gives Pro 500 “25 times the usage of Plus.”

Price: Pro 100 is $100 a month, Pro 200 $200 and Pro 500 $500, against Plus’s $20. “Pro is billed monthly.”

Watch for: each plan works out to $20 for every Plus-sized allowance, so a bigger plan buys more usage, not cheaper usage. On October 30, existing Pro 200 plans drop from 20 to 10 times Plus. The week still has a limit: “Your plan still has an included usage allowance.” And Ultrafast, only on Pro 500, “uses your included Work and Codex usage first at 8× the Standard rate.”

A second ChatGPT account

Best for: projects that belong on another plan you already have, such as your employer’s.

OpenAI’s help center explains how to “stay signed in to 2 ChatGPT accounts at the same time” on ChatGPT’s website, with “chats, memory, history, billing, and workspaces fully separate.” The Codex CLI keeps its sign-in with the rest of its state under CODEX_HOME, which, per Codex’s docs, “Sets the root for Codex state, including config, auth, logs, sessions, skills.”

Price: a second plan at full price, such as Plus at $20 a month.

Watch for: OpenAI’s terms say you may not “circumvent any rate limits or restrictions,” and OpenAI hasn’t said whether a second paid plan bought for more Codex usage counts. If you need more usage, changing your plan is the safest route. Account switching “is not yet supported in Codex desktop,” and a conversation started under one account isn’t in the other’s history.

Claude Code, on a Claude plan

Best for: a second lab’s agent, with limits of its own.

Anthropic’s limits are separate from OpenAI’s, so Claude Code keeps working when Codex can’t. One instruction file covers both: Claude Code “can read AGENTS.md as your project instructions, so a repository already set up for other coding agents works without adding a CLAUDE.md,” per its docs.

Price: Claude’s pricing lists “Claude Code included” with Pro, at $20 a month or $17 a month billed yearly, and with Max, from $100. Claude’s Free plan leaves it out.

Watch for: Claude has limits too: “Every plan has usage limits that reset on a rolling five-hour session window, and paid plans add weekly limits on top,” and “Claude Code all draws from the same pool” as your Claude chats. Claude Code doesn’t see what Codex was doing, so start it with where things stand. Our Claude Code vs Codex guide compares the two.

A local model, inside Codex

Best for: routine work with no limit and no bill, on a computer with plenty of memory.

Codex itself can run a model on your computer. Its docs say “Codex can run against a local ‘open source’ provider such as Ollama or LM Studio when you pass --oss.” Ollama’s Codex guide sets it up with ollama launch codex, or codex --oss -m and a model’s name, and you “Type /model inside Codex to switch between your installed, tool-capable Ollama models.” One to try is Recursive’s Flint, “a 39.8 billion-parameter mixture-of-experts model from Recursive, built for coding agents,” free under Apache 2.0. Our Flint memo covers what it needs. If you’d rather use another agent, OpenCode’s docs say “You can configure opencode to use local models through Ollama.”

Price: free, on your own computer.

Watch for: Flint’s download is 23 GB, and Ollama says “Use a context window of at least 64k tokens for Codex.” What runs well depends on how much memory your computer has.

Enjoy

Best for: moving a conversation to another account or agent without starting over.

Enjoy runs Codex, Claude Code, Grok Build, Antigravity, Muse Code, and local models through OpenCode, on Mac, Windows and Linux, on the plans you already have. When Codex runs out of usage:

• It uses your free resets. When OpenAI has given your account one, the waiting conversation offers Use a reset, and the conversation continues right away.
• Or it moves to another account. If you’ve signed in another ChatGPT account with usage left, Enjoy switches to it. Codex accounts “share conversations, so a conversation usually carries on where it left off after you switch.”
• Or it waits and carries on. With no account left, the conversation waits for the reset, a weekly one included, then continues where it stopped.
• Or another agent picks it up. Select Stop waiting, choose Claude Code or another agent, and it gets the conversation so far.

It also shows every signed-in account’s limits, credits and free resets, and when a wait is long, it links to OpenAI’s own pages to upgrade or buy credits.

Price: Enjoy is free on your computer, with no account. Enjoy Cloud, a monthly subscription, adds your phone and any browser, your other computers and teammates. See pricing.

Watch for: Enjoy doesn’t change OpenAI’s limits, and each account is a plan you pay for, so the caution about a second account above applies here too. Enjoy runs the Codex CLI, which OpenAI includes from Plus. Waiting needs “Your computer must be on, with Enjoy open, when the usage resets,” and “Waiting doesn’t keep the computer awake.” Enjoy isn’t open source, and the phone, other computers and teammates need Enjoy Cloud.

When to stay on your plan and wait

If the five-hour window resets within the hour, waiting is free. If you run out once a month, a bigger plan is money spent for a rare afternoon: wait, use a free reset, or pay for that day with credits. Move to Pro when it happens most weeks.

Spending less helps too, and the model matters most. On Plus, OpenAI estimates 15 to 160 local messages every five hours with GPT-6.1 Sol, 35 to 285 with GPT-6.1 Terra and 5 to 45 with GPT-6 Astra, and its usage guide says “Astra at Low effort can outperform Sol at High effort.” Fast mode “uses your included allowance at 2.5 times the rate.” Codex’s pricing page also suggests you “Reduce the size of your AGENTS.md” and “Disable MCP servers when you don’t need them.” What switching models can’t do is bring usage back after a limit: “Switching models does not restore allowance in a shared usage pool.”

How we compared

We read OpenAI’s Codex pricing, docs, help center and terms, Codex’s source on GitHub, Anthropic’s pricing and Claude Code docs, and Ollama’s and OpenCode’s pages on October 27, 2026, and every one is linked below. We quote each company’s own words for its limits and leave out what it hasn’t documented. Plans and limits change often, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.