You've hit your Claude Code or Codex limit: what you can do now

Claude's 5-hour window opens on your first message. In Anthropic's words, "The period starts when you send your first message, and your limit resets in full five hours later." The weekly limit resets at a fixed time each week that is assigned to your account, which you can see under Settings > Usage on claude.ai. On the Ultra plan, verified accounts have no 5-hour limit, only the weekly one.

You do not have to stay at the keyboard. On a Claude subscription, Claude Code waits in the open session and continues the task on its own after the limit resets, up to three times in a row. It does not start that wait by itself when the reset is more than 24 hours away, which a weekly limit often is; /rate-limit-options lets you choose what to do instead.

Codex on ChatGPT Plus and Business Standard seats has a 5-hour window, and OpenAI's pricing page says "Pro plans currently have no five-hour limit"; a weekly limit can apply on top. /status shows when each window resets, and a turn that was already running when you hit the limit can finish.

The full picture is in how Claude Code's 5-hour and weekly limits work and how Codex's limits work.

• If the message names one model, such as "You've hit your Opus limit", switch model. Claude Code's error reference says the Opus and Sonnet limits apply only to that model family, so `/model` and a model outside the family keeps you working. The session and weekly limits are shared across all models, so switching does not help with those.
• You can pay for what you use past the limit. On Claude Pro, Max and Ultra, /usage-credits turns on usage credits, and Anthropic says "your work continues on usage credits instead of stopping". On ChatGPT Plus and Pro, OpenAI says you can buy credits to keep working without changing plan.
• If you pay for both Claude Code and Codex, carry on in the other one. Writing a short note of where the task stands before you switch saves explaining it all again; keep coding after the limit has the two prompts, tried both ways.
• Some people keep several accounts with the same vendor and move between them when one runs out. That may break the vendor's terms of service, so read them before you try it.

Windows usually empty fast for a few reasons: a long conversation that is re-read with every request, coming back to it after a break, compacting a large conversation, and many subagents at once. Why your Claude Code limit ran out so fast goes through each cause, sets out what each kind of work costs against the 5-hour window, and says what to change.

Before you start again, run /usage in Claude Code. On a subscription it also shows a breakdown of what counts against your plan limits: recent usage split by skills, subagents, plugins and MCP servers, with a suggested change for anything behind ten per cent or more. It covers this machine only, for the last day or the last week. To see which turns used this window up, open Session Forensics, a free page that reads your Claude Code and Codex session files in your browser. It lists the limits you hit in the last seven days and the turns behind each one, with no sign-up, and your files are never uploaded. It needs no allowance, so you can use it while you wait. For Claude Code its 5-hour figure is an estimate, because Claude Code's files do not record the meter. If you would rather have your own agent do it, a prompt for your own agent reads your session logs and reports what used the most.

What people do when they run out collects the habits people describe for making a window last, from starting it early to stopping near the end, with what each costs, and the tools that help covers what to install or switch on.

Read the message to see whose limit it was, because that decides who can change it. "You've hit your session limit" or "You've hit your weekly limit" is your seat's own allowance, and it comes back by itself. "You've hit your individual spend limit", "You've hit your org's monthly spend limit" or "You've hit your team's shared budget" is a cap your company set, and only an admin can raise it (Claude Code's error reference). On a Claude Team or Enterprise plan, /usage-credits sends a request to your admins.

How to ask your company for more Claude Code budget has a short request to fill in and what approvers say they want to see. What happens when your team's Claude credits run out explains the organisation's limit and what an admin can change.