# Anthropic API rate limit tiers: how to move up

> Anthropic API rate limit tiers set a monthly spend cap and per-minute limits. The tiers in October 2026, how to move up, and what Claude Code hits first.

Source: https://shuriken.sh/guides/anthropic-api-rate-limit-tiers

Anthropic API rate limit tiers set a monthly spend cap and per-minute limits. The tiers in October 2026, how to move up, and what Claude Code hits first.

By Shuriken Updated 27 October 2026 5 min read

**Contents**

Short answer: Anthropic API rate limit tiers decide two things for your organisation: how much it may spend a month, and how many requests and tokens it may send per minute. As of 22 October 2026 the tiers are Evaluation, Start, Build, Scale and Custom. You move up automatically as you build usage history, or faster with Request tier increase on the Console's Rate limits page.

Every number below is from Anthropic's [rate limits page](https://platform.claude.com/docs/en/api/rate-limits), read on 22 October 2026. It changes; check the page and your own Console before you plan around a figure.

## The two kinds of limit #

Anthropic's page starts with this, and it explains most confusing 429s:

- Spend limits set "a maximum monthly cost an organization can incur for API usage".
- Rate limits set how many requests and tokens you can send over a period: requests per minute (RPM), input tokens per minute (ITPM) and output tokens per minute (OTPM), per model.

Both are set per organisation, not per API key. Every key, every teammate and every Claude Code session on the same organisation shares them.

## The tiers as of October 2026 #

Many guides still describe numbered tiers that you reach by buying credits and waiting. Anthropic's page, as read on 22 October 2026, describes it differently: organisations are "placed on a tier automatically based on usage history and account standing and can move to a higher tier over time". New organisations may start in an Evaluation tier with lower limits while that history builds.

| Tier | Monthly spend cap | Claude Opus 5.5 RPM | ITPM | OTPM |
| --- | --- | --- | --- | --- |
| Evaluation | Not listed | below standard, rising with history | below standard | below standard |
| Start | $500 | 1,000 | 2,000,000 | 400,000 |
| Build | $1,000 | 5,000 | 5,000,000 | 1,000,000 |
| Scale | $200,000 | 10,000 | 10,000,000 | 2,000,000 |
| Custom | None; arranged with Anthropic | arranged | arranged | arranged |

Sonnet 5.5 and Haiku 5.5 get the same numbers as Opus 5.5 in every tier. Fable has its own, lower row, and Fable 5.1 and Fable 5 draw on one shared limit:

| Tier | Fable RPM | ITPM | OTPM |
| --- | --- | --- | --- |
| Start | 1,000 | 500,000 | 100,000 |
| Build | 2,000 | 1,500,000 | 300,000 |
| Scale | 4,000 | 4,000,000 | 800,000 |

## How to move up a tier #

- Open platform.claude.com/usage/limits in the [Claude Console](https://platform.claude.com/). The Rate limits page shows your tier and the limits you have now.
- Keep sending traffic. Tiers rise with usage history and account standing. Ramp up gradually, because Anthropic also has acceleration limits: a sharp jump in traffic can get 429s on its own.
- Click Request tier increase on the same page to ask for higher rate limits or a higher spend cap. Anthropic support can raise limits too when it's urgent.
- For a Custom tier past Scale, contact sales from the Rate limits page.

You can also go the other way: on the Billing page, set your own spend limit below the tier's cap, and on a workspace's Limits page, set lower rate limits for that workspace so one project can't starve another.

## Which 429 is which #

Each limit fails differently, and the fix is different.

| What you see | What it means | What to do |
| --- | --- | --- |
| 429 rate_limit_error with a retry-after header | A per-minute limit | Wait that many seconds and retry; smooth out bursts |
| 429 rate_limit_error , error_code: enforced_spend_limit_reached , no retry-after | Your tier's monthly spend cap | Retries fail until 00:00 UTC on the 1st, or until you move up a tier |
| 400 invalid_request_error , "You have reached your specified API usage limits" | A spend limit you set | Raise or remove it on the Billing page |

On a Pro or Max plan, a 429 in Claude Code is a different thing; our guide to the [Claude Code 429 error](https://shuriken.sh/guides/claude-code-429-rate-limit-error) covers that side.

Responses also carry anthropic-ratelimit-* headers with the limit, what's left and when it refills. Limits use a token bucket, so capacity refills continuously rather than at the top of the minute, and Anthropic notes that 60 RPM "might be enforced as 1 request per second".

Shuriken moves a running chat to another of your own accounts when one reaches its limit, on every device.

[Download](https://shuriken.sh/download)

## Caching raises your real ceiling #

On most current models, ITPM counts fresh input tokens and cache writes. Cache reads are left out. Anthropic works an example on the rate limits page: a 2,000,000 ITPM limit with an 80% cache hit rate lets you process 10,000,000 input tokens a minute.

Cache reads are cheaper too, as our guide to [prompt caching in Claude Code](https://shuriken.sh/guides/prompt-caching-in-claude-code) explains. On Claude Opus 5.5 they cost [$0.20 per million tokens against $4 for fresh input](https://platform.claude.com/docs/en/about-claude/pricing). Claude Code caches heavily on its own, so most of a long session's input is cache reads that don't count toward ITPM.

## What heavy Claude Code API users hit first #

On a key, Claude Code has no 5-hour or weekly window, so the API's limits are the only ones. For one person on Start, the monthly spend cap is the limit most likely to bite.

The arithmetic uses Anthropic's [cost docs](https://code.claude.com/docs/en/costs): about $13 per developer per active day across enterprise deployments, and 90% of users under $30. $500 / $13 is about 38 active days. $500 / $30 is about 17. A heavy user on Start can reach the cap with days of the month left, and from then on every request returns enforced_spend_limit_reached until the 1st. Parallel agents get there faster, because each runs its own context.

- The first time you sign Claude Code in with a Console account, it creates a workspace called "Claude Code". For organisations with custom rate limits, Anthropic says that workspace's traffic counts toward the organisation's overall limits; a workspace rate limit keeps it from crowding out production.
- For teams of one to five, Anthropic suggests 200,000 to 300,000 tokens per minute and 5 to 7 requests per minute per user.
- Fable's ITPM on Start is 500,000, a quarter of Opus 5.5's 2,000,000, so a long Fable session with a big context is the likeliest to see a per-minute 429.

If you'd rather not depend on API limits at all, a Pro or Max plan swaps them for 5-hour and weekly windows. Our guide to [API key or subscription for Claude Code](https://shuriken.sh/guides/claude-code-api-vs-subscription) compares the two with your own numbers, and [what to do when you reach a Claude limit](https://shuriken.sh/guides/claude-limit-reached-what-to-do) covers the plan side.

## How Shuriken treats an API key #

In Shuriken, an API key is a source, like a Claude sign-in. It doesn't take chats on its own. It runs a chat when you pick it with Use, or as a last resort if you turn on Switch to API keys under Sources › Switch settings (off by default): then keys take chats only once every subscription for that agent is out.

The Spend cap per fallback setting, on by default at $25, stops that run ("Past this, chats wait for a subscription"). It counts everything the run spends with that provider, across every key it uses, not per key. When a subscription has room again, chats go back to it and the next run starts from a fresh cap. We wrote up the reasoning in [API keys as a last resort](https://shuriken.sh/blog/api-keys-last-resort).

That cap sits on top of Anthropic's limits, not instead of them: your organisation's spend cap and per-minute limits still apply to the key. To see what your turns would cost at API prices before you add a key, [download Shuriken](https://shuriken.sh/download) and check its Analytics page.

Try the free tool: [Claude limit reset tracker](https://shuriken.sh/tools/limit-resets/claude)

Share

[X](https://x.com/intent/post?url=https%3A%2F%2Fshuriken.sh%2Fguides%2Fanthropic-api-rate-limit-tiers&text=Anthropic%20API%20rate%20limit%20tiers%3A%20how%20to%20move%20up)[LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fshuriken.sh%2Fguides%2Fanthropic-api-rate-limit-tiers)[Hacker News](https://news.ycombinator.com/submitlink?u=https%3A%2F%2Fshuriken.sh%2Fguides%2Fanthropic-api-rate-limit-tiers&t=Anthropic%20API%20rate%20limit%20tiers%3A%20how%20to%20move%20up)[Reddit](https://www.reddit.com/submit?url=https%3A%2F%2Fshuriken.sh%2Fguides%2Fanthropic-api-rate-limit-tiers&title=Anthropic%20API%20rate%20limit%20tiers%3A%20how%20to%20move%20up)[E-mail](mailto:?subject=Anthropic%20API%20rate%20limit%20tiers%3A%20how%20to%20move%20up&body=https%3A%2F%2Fshuriken.sh%2Fguides%2Fanthropic-api-rate-limit-tiers)

## Sources

- [Claude API docs: Rate limits](https://platform.claude.com/docs/en/api/rate-limits) (accessed 22 October 2026)
- [Claude Code docs: Manage costs effectively](https://code.claude.com/docs/en/costs) (accessed 22 October 2026)
- [Claude API docs: Pricing](https://platform.claude.com/docs/en/about-claude/pricing) (accessed 22 October 2026)

## Frequently asked questions

**What are the Anthropic API usage tiers?**

Anthropic's rate-limits page, read on 22 October 2026, lists Evaluation, Start, Build, Scale and Custom. The spend cap is $500 a month on Start, $1,000 on Build and $200,000 on Scale; Custom limits are agreed with Anthropic.

**How do I get a higher Anthropic API tier?**

Anthropic moves organisations up automatically based on usage history and account standing. To go faster, use Request tier increase on the Rate limits page of the Claude Console, or contact Anthropic support.

**Do cached tokens count toward Anthropic's rate limits?**

For most current models, no. Only uncached input tokens and cache writes count toward input tokens per minute. Cache reads are billed at a much lower rate and don't count, so prompt caching raises your effective throughput.

**How do I tell a spend-cap 429 from a rate-limit 429?**

Both have the type rate_limit_error. The spend-cap one has the error code enforced_spend_limit_reached and no retry-after header, and retrying fails until the 1st of next month or a higher tier.

**Do Claude Pro or Max plans have API rate limit tiers?**

No. Tiers apply to Claude Console organisations that pay per token. Pro, Max and Team plans have 5-hour and weekly usage limits instead.

Source switching

One account runs dry. The next one takes over.

Add the Claude, Codex and Gemini accounts you own. At your switch point, your chats move to another of them and keep going.

[Download](https://shuriken.sh/download)

[Other platforms](https://shuriken.sh/download#main)

## Keep reading

- [How to check your Claude Code usage: Guide 27 Oct 2026 6 min read To check Claude Code usage, type /usage or open Settings › Usage on claude.ai. Both read one pool: your 5-hour session and your weekly limit.](https://shuriken.sh/guides/check-claude-code-usage)
- [Claude extra usage: how pay-as-you-go overflow works: Guide 27 Oct 2026 5 min read Claude extra usage, now called usage credits, keeps you working past your plan's limit at API rates. Where to turn it on, the cap, and when it pays.](https://shuriken.sh/guides/claude-extra-usage-explained)
- [Every Claude usage limit change in 2026, dated: Guide 27 Oct 2026 5 min read Claude usage limits changed in 2026 more often than the plans did: peak hours, doubled 5-hour limits, a summer weekly boost, and a new weekly level. Dated.](https://shuriken.sh/guides/claude-usage-limit-changes-2026)
- [How to cap Claude Code spending for a team: Guide 27 Oct 2026 6 min read Limit Claude Code spending for a team with per-member usage-credit limits, Console workspace limits, a gateway's daily caps or --max-budget-usd per run.](https://shuriken.sh/guides/limit-claude-code-spending-for-a-team)
- [How to switch Claude accounts without logging out: Guide 27 Oct 2026 7 min read Three ways to use several Claude accounts in Claude Code without /logout and /login: one config per account, a switcher CLI, or a harness.](https://shuriken.sh/guides/switch-claude-accounts-without-logging-out)
- [Claude's 5-hour limit, explained: Guide 27 Oct 2026 6 min read Claude's 5-hour window starts with your first message and resets five hours later. An idle account has no window running, so it saves nothing up.](https://shuriken.sh/guides/claude-5-hour-limit-explained)

## Free tools

- [Claude limit reset tracker](https://shuriken.sh/tools/limit-resets/claude)
- [Codex limit reset tracker](https://shuriken.sh/tools/limit-resets/codex)