Claude Code hit its usage limit: what you can do
Checked 16 November 2026
On a Pro, Max, Ultra, Team or Enterprise plan, Claude Code draws from a usage allowance with a weekly limit and, on most plans, a five-hour session limit. When one runs out, Claude Code stops sending requests until the reset time in the message. You can wait, pay for more, switch to an API key, switch model if the limit is model-specific, or move the task to another agent.
This page is published by Bivy, which is one of the options below. Each claim about Claude Code comes from Anthropic's own docs and help center, and each claim about another tool from that tool's docs, linked under Sources. Limits and commands change often, so check the linked page before you rely on a detail.
How the limits work
- Two windows at once: Pro and Max plans have a five-hour session limit and a weekly limit. Usage counts against both at the same time, so a heavy burst can use up the weekly allowance before the session window resets. Ultra plans have a weekly limit, and verified Ultra accounts have no five-hour session limit.
- Shared with Claude: the limits are shared across Claude and Claude Code. Chat on claude.ai or in Claude Desktop uses the same allowance.
- Model limits: Anthropic's docs also list Opus and Sonnet limits that apply only to requests to that model family. Max and Ultra plans have a separate weekly limit for Fable.
- Plan size: Max ($100/month) gives five times Pro's usage every five hours and every week. For more than that, Anthropic points to the Ultra plan. Team and Enterprise members draw from a per-seat allowance that depends on the seat tier.
- What uses it up: every turn sends the whole conversation, your project context such as
CLAUDE.md, and files Claude has read. Long sessions, Opus, subagents and scheduled tasks use more. A529 Overloadederror is not your usage limit and doesn't count against it.
What the message means
| Message | What ran out | Does /model help? |
|---|---|---|
You've hit your session limit · resets 3:45pm | The five-hour session allowance, across all models | No |
You've hit your weekly limit · resets Mon 12:00am | The weekly allowance, across all models | No |
You've hit your Opus limit or Sonnet limit | The allowance for that model family | Yes, switch to a model outside that family |
You've hit your monthly spend limit and similar | Usage credits reached a spend cap you or your admin set | No; raise the cap or wait |
Credit balance is too low | Prepaid Console API credits, or Claude Code is using an API key you didn't mean to use | No; see the API key section |
See where you stand
Run /usage in Claude Code. On a subscription it shows plan usage bars, when each limit resets, and a breakdown of what used your allowance, such as skills, subagents, MCP servers and loops, with suggested changes. The breakdown is computed from this machine's history only. Settings > Usage on claude.ai shows the current session and weekly bars and when each resets. Claude Code also warns you before a window runs out, for example You've used 85% of your session limit.
Wait for the reset
The free option, and often the right one for a session limit. Claude Code blocks requests until the time in the message.
- Automatic continue: in an interactive session signed in with a claude.ai subscription, Claude Code v2.1.234 or later waits in the open session and continues the task after the reset. It prompts Claude to pick up where it stopped; it doesn't resend your last message. It's on by default. Turn it off in
/configwith Continue automatically at usage limit. - Keep the session open: exiting Claude Code, switching accounts, clearing or resuming another conversation ends the wait. If the computer slept more than about 30 minutes past the reset, you press Enter to continue.
- When it doesn't start on its own: for a reset more than 24 hours away, such as many weekly limits, and in Remote Control sessions. You can start it yourself from
/rate-limit-options. It isn't offered in background sessions,-pruns, or with an API key. - Desktop app: the Code tab's session-limit card has an Auto-continue when limits reset checkbox. The weekly-limit card doesn't.
- Limit resets: Anthropic occasionally gives eligible plans a limit reset. Depending on the type, it puts the five-hour session limit or the weekly limit back to full, or gives back the weekly usage spent on Fable models. Use it with “Reset for free” under Settings > Usage on the web, in Claude Desktop or in the Claude mobile apps, or on the limit message there. It isn't available in Claude Code in your terminal or IDE, and you can't undo it.
- Trade-off: permission prompts still apply, so an unattended task can stop on one. After three automatic continuations, Claude Code stops and you choose what to do in
/rate-limit-options. A weekly limit can be days away.
Buy usage credits
Keep using your subscription past the limit and pay for the extra. Anthropic calls these usage credits. They are available on Pro, Max and Ultra, and on Team and Enterprise if your admin turns them on.
- How: run
/usage-creditsin Claude Code. On Pro, Max and Ultra it opens Settings > Usage on claude.ai, where you enable credits, add a payment method and prepay. On Team and Enterprise it sends a request to your admins unless you have billing access. - Cost: standard API rates, charged separately from your subscription. You can set a monthly spend cap, turn on auto-reload and set usage alerts. Credits bought on or after 7 December 2026 generally expire 12 months after purchase; earlier ones don't expire.
- Trade-off: a long Claude Code session at API rates can cost more than you expect. You can also buy credits in the Claude mobile apps where in-app purchases are available; auto-reload is web-only.
Use an Anthropic API key instead
Bill Claude Code per token through the Claude Console instead of your plan. Anthropic's help center suggests this for intensive coding sprints.
- How: buy credits in the Claude Console, then run
/loginand sign in with the Console account, or setANTHROPIC_API_KEY. To go back to your subscription, run/loginagain and unset the variable. - Cost: standard API rates. There is no usage window, only your prepaid balance and any workspace spend limit.
- Watch out: while
ANTHROPIC_API_KEYis set, Claude Code uses it instead of your subscription. If you forget it in your shell profile, you keep paying API rates after the reset./statusshows which credential is active.
Switch model
Helps only with an Opus or Sonnet limit. Run /model and pick a model outside the family that ran out. Each model has its own prompt cache, so the next request re-reads the whole conversation.
To make the allowance last longer next time, Anthropic's docs suggest Sonnet for most coding, /clear between unrelated tasks, a shorter CLAUDE.md, and the recommendations in /usage.
Move the task to another agent
Use a plan you already pay for with another vendor. Your repository and files stay where they are; what you have to carry over is the conversation.
- Codex (OpenAI): included in ChatGPT Free, Go, Plus, Pro, Business, Edu and Enterprise, or used with an OpenAI API key. It has its own usage limits, and Plus and Pro users can buy credits to continue past them.
- OpenCode: open source, with a provider catalog from models.dev. You can sign in with a ChatGPT Plus or Pro plan or a GitHub Copilot account, or run local models through Ollama.
- Gemini CLI (Google): sign in with a Google account, a Gemini API key or Vertex AI. Google says that for unpaid-tier and Google One users, Gemini CLI was replaced by Antigravity CLI on 18 June 2026.
Carry the context over by hand
- Save the code state. Commit or stash what Claude Code changed, or at least note
git statusand the branch. The next agent reads the files, not the old conversation. - Write a handoff summary. If you see the 85% warning, ask Claude Code for a short summary: the goal, what's done, what's left, decisions made, and commands that test the work. Save it to a file in the repo. After the limit, write it yourself.
- Share project instructions. Codex and OpenCode read
AGENTS.md. OpenCode falls back toCLAUDE.mdwhen there's noAGENTS.md. Gemini CLI readsGEMINI.mdby default and can be set to readAGENTS.mdwithcontext.fileName. Claude Code readsAGENTS.mdwhen a repository has noCLAUDE.md, or through an@AGENTS.mdimport, so one shared file can serve all of them. - Start the other agent in the same directory and point it at the summary file.
Expect some loss. A summary doesn't carry every detail, tool result or rejected idea from the original conversation.
Bivy
An open-source (MIT) workspace that runs coding agents on your own machine or a Bivy cloud computer, and steers them from a phone, browser or terminal. Two of its features apply here. Its README says: "Hit a usage limit? Keep going. Fork the session, code and conversation included, to another agent, or let Bivy pick it up again when the limit resets."
- When a turn hits a limit: the chat offers Fork to another agent and Retry when the limit resets, and sends a notification. This works for Claude Code, Codex, OpenCode, Pi, Grok and other ACP agents.
- Retry at the reset: Bivy reads the reset time from the agent's limit message and re-sends the turn's last prompt then. The due time is saved, so a restart of Bivy doesn't lose it. A new prompt from you replaces the pending retry. The machine running the session still has to be on at that time.
- Fork: a fork copies the session's code and conversation to another agent. From Claude Code, a fork to Codex, Pi or OpenCode replays the conversation history into the new agent. A fork to any other agent, such as Gemini CLI or Aider, is a seeded continuation: a new session that starts from a continuation prompt built from the old one.
- Trade-off: you still need a plan or API key for the agent you fork to; Bivy doesn't supply model usage. Replayed history is not byte-identical to the original session. Bivy is one more tool to install, and Claude Code's own automatic continue already covers waiting in an open terminal.
Which should you pick?
- A session limit with a reset within a few hours: wait. Keep the session open and let Claude Code continue on its own.
- An Opus or Sonnet limit:
/modelto the other family. - You need Claude now and can pay: usage credits if you want to stay on your subscription login; a Console API key if you want a separate, per-token bill.
- A weekly limit days away, and you have a ChatGPT or other plan: commit, write a summary, and continue in Codex or OpenCode.
- You hit limits often and switch agents a lot: keep instructions in
AGENTS.md, or use a tool like Bivy that forks the session or retries at the reset for you. - You hit limits every week: check the
/usagebreakdown first, then consider a larger plan.
FAQ
When does my Claude Code usage limit reset?
The limit message shows the reset time, for example "You've hit your session limit · resets 3:45pm". Pro and Max plans have a five-hour session limit and a weekly limit. Run /usage in Claude Code, or open Settings > Usage on claude.ai, to see both and when they reset.
Does switching models get around the limit?
Only for a model-specific limit. Anthropic's docs say the session and weekly limits are shared across all models, so /model doesn't help. After "You've hit your Opus limit" or "You've hit your Sonnet limit", switching to a model outside that family keeps you working.
Can I keep using Claude Code after hitting the limit?
Yes, if you pay for it. On Pro, Max and Ultra you can turn on usage credits, billed at standard API rates, with /usage-credits. Or sign in with a Claude Console account or set ANTHROPIC_API_KEY, and Claude Code bills per token instead of drawing on your plan.
Does Claude Code continue on its own after the reset?
In an interactive session signed in with a claude.ai subscription, Claude Code v2.1.234 or later waits and continues the task after the reset, as long as you keep the session open. It doesn't start the wait on its own for a reset more than 24 hours away, such as a weekly limit.
How do I move the task to another agent without losing context?
Commit or note your git state, ask Claude Code for a written summary before you run out if you can, and keep project instructions in AGENTS.md, which Codex and OpenCode read. Then start the other agent in the same repository and give it the summary.
Sources
All checked on 16 November 2026.
- Anthropic, Claude Code docs: Usage limit errors, Wait for a usage limit to reset, Manage costs effectively, AGENTS.md support
- Anthropic, Claude help center: Use Claude Code with your Pro or Max plan, Manage usage credits for paid Claude plans, What is a limit reset?, Usage limit best practices, What is the Max plan?, Models, usage, and limits in Claude Code
- OpenAI: Codex pricing and limits, AGENTS.md in Codex
- OpenCode: Providers, Rules
- Gemini CLI: Authentication, GEMINI.md context files
- Bivy: README, fork fidelity matrix, in-session resume after a usage limit, agent support matrix
Want to try Bivy? Start on a cloud computer with nothing installed, or connect your own Mac or Linux machine.