Claude Limit: How Usage and Rate Limits Work
Claude limit explained: five-hour and weekly usage limits across claude.ai and Claude Code, the error messages you see, and how API rate limits work.
Josh Pocock8 min read
AI Coding
Claude limits
Stride AI Academy
A Claude limit is one of two things. On a Claude plan, it's a usage limit: a session allowance that resets five hours after your first message, plus a weekly limit, shared across claude.ai, the desktop app and Claude Code. On the API, it's a rate limit: a cap on requests and tokens per minute for your organization, which returns a 429 error when you go over it.
This guide covers how the plan limits work, what counts against them, the messages you see, how to check your usage and what to do when you hit a limit, then the API rate limits, usage tiers and headers. Everything here was checked on November 21, 2026 against the Help Center articles on usage and length limits, error messages and Claude Code with your plan, Claude's pricing page, Claude Code's costs and errors docs, and the API docs on rate limits and errors.
How Claude usage limits work
Paid plans have two kinds of usage limit, per the usage limits article:
- A session limit. The window starts when you send your first message and resets five hours later.
- Weekly limits. These apply across all models. Your reset day and time are fixed to your account, and you get your full allowance each cycle. Max and Ultra also have a separate weekly limit for Fable.
Verified Ultra accounts have no five-hour limit, only the weekly one.
Anthropic doesn't publish message counts. The pricing page states each plan's allowance relative to the others:
| Plan | Five-hour limit | Weekly limit |
|---|---|---|
| Free | Limited | Limited |
| Pro | More than Free | More than Free |
| Max | 5x Pro | 5x Pro |
| Ultra | None for verified accounts | 10x Pro |
For what each plan includes, see the Claude free tier and Claude Max plan guides. For prices, see Claude Code pricing.
What counts against your limit
The usage limits article lists four things that decide how fast you use your allowance:
- Conversation length and complexity. Long chats send more with every message.
- The features you use, such as Research, web search and connectors.
- The model. Heavier models use the allowance faster.
- The effort level you select.
One pool covers everything. The Claude Code article says limits "are shared across Claude and Claude Code," and IDE usage counts too.
Usage counts against the session and weekly allowances at the same time. Claude Code's errors docs warn that one burst of heavy activity, such as a large workflow fanout, can use up the weekly allowance before the five-hour window resets.
In Claude Code, the docs list what makes usage climb in a long session: the full conversation goes out with every request, the first message after a break can miss the cache, and subagents, scheduled tasks and agent teammates each send their own requests. Our guide to Claude Code context management covers the habits that keep this down.
The messages you see
In the Claude apps, the error messages article lists these:
- Warning: "Approaching session limit" or "Approaching weekly limit," sometimes with a percentage.
- Limit reached: "5-hour limit reached - resets [time]" or "You've reached your weekly limit. It resets [day] at [time]."
- With usage credits on: "5-hour limit resets [time] - continuing with usage credits."
In Claude Code, the messages look like this:
You've hit your session limit · resets 3:45pm
You've hit your weekly limit · resets Mon 12:00am
You've hit your Opus limit · resets 3:45pm
You've hit your Sonnet limit · resets 3:45pmThe session and weekly limits cover all models, so switching models won't help. The Opus and Sonnet limits only cover that model family, so /model to a different family keeps you working.
Two messages are not your plan limit. "Due to unexpected capacity constraints, Claude is unable to respond" in the apps means temporary high demand. "Server is temporarily limiting requests (not your usage limit)" in Claude Code is a short throttle that Claude Code retries automatically.
How to check your usage
In the Claude apps, go to Settings > Usage. It shows the time left in your current session and your next weekly reset.
In Claude Code, run:
/usageOn Pro, Max, Ultra, Team and Enterprise plans, /usage shows your plan usage bars and a breakdown of what used them: skills, subagents, plugins and individual MCP servers, with suggested fixes. Press d or w to switch between the last 24 hours and the last 7 days. The breakdown only counts sessions on that machine, not claude.ai or other devices.
To watch your allowance all the time, the errors docs point to the rate_limits fields in a custom status line.
What to do when you hit a limit
From the usage limits article and Claude Code's errors docs:
- Wait for the reset. The message tells you when.
- Let Claude Code continue for you. In an interactive session signed in with a subscription, Claude Code can wait and resume the interrupted task after the reset (v2.1.234 or later).
- Switch model families if the message names Opus or Sonnet.
- Turn on usage credits to keep working at standard API rates. Run
/usage-creditsin Claude Code, or use Settings > Usage. Our guide to Claude credits explains how they're billed and capped. - Upgrade your plan. New limits apply right away.
One trap: if ANTHROPIC_API_KEY is set in your environment, Claude Code uses that key instead of your plan. Run /status to check which credential is active.
Usage limits vs length limits
A length limit is different. It's the context window: how much Claude can work with in one chat. The usage limits article says newer models support up to 1M tokens on paid plans, and others support 500K or 200K. If you see "Your message will exceed the length limit for this chat," start a new chat or use a project. Usage limits cap your total use across all chats. Length limits cap one chat.
Claude API rate limits
The API works differently. There's no five-hour window. Per the rate limits docs, the API has two types of limit:
- Spend limits: a maximum monthly cost for your organization.
- Rate limits: requests per minute (RPM), input tokens per minute (ITPM) and output tokens per minute (OTPM), set per model.
Limits apply to the whole organization. The API uses a token bucket, so capacity refills continuously rather than resetting at fixed times. Short bursts can still trip a limit: 60 RPM might be enforced as 1 request per second.
Usage tiers
Your organization is placed on a tier automatically based on usage history and account standing. New organizations may start in an Evaluation tier with lower limits. The standard tiers carry these monthly spend caps:
| Tier | Monthly spend cap | Opus 5.5 limits (RPM / ITPM / OTPM) |
|---|---|---|
| Start | $500 | 4,000 / 2,000,000 / 400,000 |
| Build | $1,000 | 10,000 / 5,000,000 / 1,000,000 |
| Scale | $200,000 | 13,000 / 10,000,000 / 2,000,000 |
A Custom tier has no spend cap and is arranged with sales. Check your tier and exact limits at the Rate limits page in the Claude Console (platform.claude.com/usage/limits), and use Request tier increase there to ask for more.
Cached input doesn't count toward ITPM. Only uncached input tokens and cache writes do, so prompt caching raises your real throughput.
429 errors
A 429 rate_limit_error means one of three things, per the API errors docs:
- You went over a rate limit. The response usually has a
retry-afterheader. Wait that many seconds. - You hit a ramp limit. A sharp jump in usage can trigger a 429 even under your limits. Increase traffic gradually.
- You reached your tier's monthly spend cap. This 429 has no
retry-afterheader, anderror.details.error_codeisenforced_spend_limit_reached. Requests fail until the next month or a tier increase.
A spend limit you set yourself returns a 400, not a 429. A 529 overloaded_error means the API is busy for everyone.
Rate limit headers
Every API response tells you where you stand:
retry-after
anthropic-ratelimit-requests-limit
anthropic-ratelimit-requests-remaining
anthropic-ratelimit-requests-reset
anthropic-ratelimit-input-tokens-limit
anthropic-ratelimit-input-tokens-remaining
anthropic-ratelimit-input-tokens-reset
anthropic-ratelimit-output-tokens-limit
anthropic-ratelimit-output-tokens-remaining
anthropic-ratelimit-output-tokens-resetThe reset values are RFC 3339 timestamps. There are also anthropic-ratelimit-tokens-* headers, which show the most restrictive limit in effect.
If you run Claude Code with an API key, the costs docs suggest per-user rate limits by team size, from 200k to 300k tokens per minute for 1 to 5 users down to 10k to 15k for 500 or more. Scripted claude -p runs hit these limits fastest, so keep concurrency modest.
Learn to build with Claude Code
The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.
FAQ
What is the Claude limit?
On a Claude plan, it's a usage allowance that resets five hours after your first message, plus a weekly limit. On the API, it's a cap on requests and tokens per minute, plus a monthly spend cap set by your usage tier.
How long does the Claude limit last?
The session limit resets five hours after the first message of the window. The weekly limit resets at a fixed day and time assigned to your account. The limit message tells you the exact reset time.
Do Claude Code and claude.ai share the same limit?
Yes. The Help Center says limits are shared across Claude and Claude Code, and IDE and desktop use count too. Heavy coding leaves less for chat in the same window.
How do I check my Claude usage?
Open Settings > Usage in the Claude apps to see time left in your session and your next weekly reset. In Claude Code, run /usage for plan usage bars and a breakdown of what used them.
What are Claude API rate limits?
They cap requests per minute, input tokens per minute and output tokens per minute for each model, set by your organization's usage tier. Going over returns a 429 error, usually with a retry-after header.
Can I keep using Claude after hitting the limit?
On Pro, Max, Ultra and Team, you can turn on usage credits and keep working at standard API rates up to a spend limit you set. Otherwise, wait for the reset or upgrade, which applies straight away.