Claude Code Usage Limits Reset: When It Happens and What to Do (2026)

Last updated: November 7, 2026. By Veer. Checked on November 7, 2026 against Anthropic’s help articles on how usage limits work, Claude Code on Pro, Max and Ultra and limit resets, the Claude pricing page, and the Claude Code docs on errors, interactive mode and commands. This is a researched explainer from public documentation, not a hands-on test.

If you use Claude Code every day, you’ve probably seen the message that ends your flow: “You’ve hit your session limit · resets 3:45pm.” The good news is that Claude Code usage limits reset on a predictable schedule, and you can see exactly when. The less obvious part is that there are two clocks running at once, and the weekly one doesn’t work the way most people assume.

This guide explains both resets, how to check your own times, what changed with the November 2026 plan rename, and your real options when you’re locked out.

When do Claude Code usage limits reset?

The five-hour session limit resets five hours after the first message of that session, and the weekly limit resets at a fixed day and time Anthropic assigns to your account. Both limits are shared across claude.ai, Claude Desktop and Claude Code, so a long chat in the browser eats into your terminal budget too.

Here’s how Anthropic describes it in its usage limits help article:

• The five-hour period “starts when you send your first message, and your limit resets in full five hours later.”
• Weekly limits “reset at a fixed time each week that is assigned to your account. Your reset day and time stay the same regardless of when you start using Claude or when your subscription begins.”

So if you start working at 9:10am and burn through your session allowance by 11am, you’re back at 2:10pm, not at midnight and not “five hours from when you ran out.” The weekly limit is different: if your account resets on Monday at 12:00am, it resets then whether you started coding on Saturday or Wednesday.

Source: Anthropic help center and Claude Code errors reference, checked November 7, 2026.

One thing to keep in mind: every five-hour window also draws down your weekly allowance. Once the week is used up, a fresh five-hour window doesn’t help. You wait for the weekly reset.

How to see your exact reset time

Type /usage inside Claude Code, or open Settings > Usage on claude.ai; both show each limit with the time it resets. The limit message itself also carries the reset time.

You have four places to look:

1. The limit message. Claude Code’s errors reference lists the exact wording, for example You've hit your session limit · resets 3:45pm or You've hit your weekly limit · resets Mon 12:00am.
2. /usage in Claude Code. On Pro, Max, Ultra, Team and Enterprise it shows your plan usage bars along with a breakdown of what’s using them. /cost and /stats are aliases for the same screen, per the commands list.
3. Settings > Usage on claude.ai or Claude Desktop. It shows “Current session” time remaining and your next weekly reset.
4. Your status line. If you script a custom status line, Claude Code passes rate_limits.five_hour.resets_at and rate_limits.seven_day.resets_at (Unix timestamps) so you can keep a countdown on screen, according to the status line docs.

Times show in your own time zone. Don’t assume your weekly reset is midnight, because each account gets its own slot.

Limits by plan after the November 2026 rename

On Pro and Max you have both a five-hour and a weekly limit; on verified Ultra you only have the weekly one. Anthropic renamed its plans at the end of October 2026: the old Max 5x is now just “Max” and the old Max 20x is now “Ultra,” at the same prices, as The Next Web reported on November 1.

US prices from claude.com/pricing; UK in-app prices from the Claude App Store listing; both checked November 7, 2026. App Store prices can differ from buying on the web.

Anthropic doesn’t publish message or token counts for any plan. The multiples are relative to Pro, and how fast you hit a limit depends on the model, effort level, context size and tools you use. Newer Ultra accounts keep a temporary five-hour limit until Anthropic verifies them, and Anthropic says it will email you when that’s lifted.

If you’re weighing plans, our Claude Max 5x vs 20x guide goes deeper on the price math.

What happens when you hit the limit mid-task

Claude Code stops new requests until the reset time, but it can finish the step it’s on and, since v2.1.234, pick the task back up automatically after the reset.

Two features soften the blow:

• Wrap-up allowance. If you hit the five-hour limit partway through a response, Claude Code may keep working briefly to finish that step, showing “Usage limit reached · wrapping up.” Anthropic’s Claude Code plan article says this is small and capped, counts toward your weekly limit, and applies once per week on Pro but at every five-hour limit on Max. It’s still rolling out.
• Automatic continue. In interactive sessions signed in with a Claude subscription, Claude Code waits in the open session and resumes when the limit resets. You’ll see Continuing automatically at 3:45pm · Esc to cancel. It’s on by default and needs v2.1.234 or later, per the interactive mode docs. To turn it off, use /config and switch off “Continue automatically at usage limit,” or set autoContinueAtUsageLimit to false.

Leaving a terminal open overnight so a task resumes at 2am is handy, but it is unattended execution. Only do it in a repo where you’re comfortable with the permissions you’ve granted.

Can you reset Claude Code limits early?

Not on demand. Your options are a free “limit reset” if Anthropic has offered you one, usage credits, a plan upgrade, or switching to API billing.

• Free limit resets. Anthropic occasionally gives eligible plans a one-time reset. You use it from Settings > Usage on the web or in Claude Desktop with the “Reset for free” button. It isn’t available inside Claude Code or on mobile, but because limits are shared, the reset applies in Claude Code too. Your weekly limit still resets on its usual day, according to Anthropic’s limit reset article.
• Usage credits. Turn them on with /usage-credits (it opens Settings > Usage in your browser) and your work continues on paid credits instead of stopping. /rate-limit-options shows the wait, credits and upgrade choices in one menu.
• Upgrade. Anthropic says a new plan’s limits apply right away.
• Switch model, sometimes. Switching models with /model doesn’t restore access after a session or weekly limit. It does help when the message is a model-specific one like “You’ve hit your Opus limit.”
• API billing. You can sign in with a Claude Console account and pay standard API rates for a heavy sprint.

Watch out for one trap: if an ANTHROPIC_API_KEY environment variable is set, Claude Code uses that key instead of your subscription and you get billed at API rates. Anthropic flags this in its help center. If you’d rather never be offered API credits, log out and sign back in using only your Pro, Max or Ultra credentials.

How to make your limits last longer

Use a lighter model and lower effort for routine work, keep context lean, and don’t spend your Claude budget in the browser if you need it in the terminal.

Anthropic’s own advice in its usage help article is to pick lower effort for routine tasks, turn off tools and connectors you don’t need, and start fresh conversations rather than letting long ones get summarized. In Claude Code, the /usage breakdown points at what’s actually eating your allowance, such as a bloated CLAUDE.md file or a chatty MCP server, and newer versions can apply some fixes with a single key press.

A few habits that follow from how the limits work:

• Start heavy sessions early in a five-hour window so the reset lands while you’re still working.
• Use Sonnet 5.5 for everyday edits and save Opus for hard problems. Our Sonnet 5.5 vs Opus 5.5 comparison covers the trade-off.
• If you also pay for ChatGPT, a second agent with its own separate limits is a real option. See how to use Claude Code and Codex together.

Verdict

Claude Code’s limits are less mysterious than they feel at 3pm on a deadline. The session clock starts with your first message and runs five hours. The weekly clock is a fixed slot on your account. Check both with /usage, let auto-continue pick work back up, and keep usage credits as a safety valve rather than a habit. If you hit the weekly wall most weeks, that’s the signal to move up a plan. Compare costs first in our Claude Code vs Codex breakdown and our best AI subscription guide.

FAQ

How often do Claude Code usage limits reset?

The session limit resets five hours after the first message of each session, and the weekly limit resets once a week at a fixed day and time assigned to your account. Verified Ultra accounts have only the weekly limit.

When does the Claude Code weekly limit reset?

At the same day and time every week, set by Anthropic for your account. It doesn’t depend on when you subscribed or started using Claude. Check the exact time with /usage in Claude Code or Settings > Usage on claude.ai.

Is there a command to reset Claude Code limits manually?

No. You can’t reset limits on demand. If Anthropic has given your account a free limit reset, you redeem it with the “Reset for free” button in Settings > Usage on the web or in Claude Desktop. Otherwise you can wait, turn on usage credits with /usage-credits, or upgrade.

Can Claude Code continue automatically after the limit resets?

Yes. From version 2.1.234, interactive sessions signed in with a Claude subscription wait and continue the task on their own when the limit resets. You can turn this off in /config or with the autoContinueAtUsageLimit setting.

Do claude.ai chats count against my Claude Code limits?

Yes. Anthropic says usage across claude.ai, Claude Desktop, mobile and Claude Code all draws from the same limits, so heavy chatting reduces what’s left for coding.

Does switching models get around a usage limit?

Not for the session or weekly limit, which apply across all models. It does help with a model-specific limit such as “You’ve hit your Opus limit,” where moving to another model family with /model keeps you working.