How to raise your Anthropic rate limits

Where do you see your current tier?

On the Rate limits page in the Claude Console. New organizations can start in an Evaluation tier below the standard limits. It lifts on its own as you build usage history, so a low limit in your first week doesn’t mean anything is wrong with the account.

How do you ask for a higher tier?

1. Open the Rate limits page in the Claude Console.
2. Click Request tier increase. The same flow covers higher rate limits and a higher monthly spend cap.
3. If it’s urgent, contact Anthropic support as well. Support can raise limits too.
4. If you need more than the Scale tier, contact sales from the same page.

Anthropic also moves organizations up automatically over time. It doesn’t publish the thresholds.

Tier	Monthly spend cap
Start	$500
Build	$1,000
Scale	$200,000
Custom	None, arranged with your account team

Rates as listed on Anthropic’s Rate limits page, checked November 2, 2026.

Per-model numbers are on our AI API rate limits page.

Are you on Claude Platform on AWS?

Then Request tier increase isn’t available to you. Contact your Anthropic account representative or support instead. Tell them which models you need raised and your peak input and output tokens per minute for each one. Add roughly what share of your input is cached or repeated.

That’s a good shape for a limit request to any vendor, AWS or not.

What should you do about the spend cap?

Treat it as a harder stop than a rate limit. Once you reach the cap, API usage pauses until 00:00 UTC on the first of the next month. The 429 you get carries no retry-after header, and retries keep failing until access comes back. On the Messages API, error.details.error_code is enforced_spend_limit_reached, which is how your error handling can tell it apart from an ordinary rate limit.

Moving to a higher tier restores access. Request the increase before you get close to the cap, not after.

Can you get more out of the limits you have?

Start with prompt caching. On most Claude models only uncached input counts toward input tokens per minute. Anthropic’s own example is an organization with a 2,000,000 limit and an 80 percent cache hit rate, handling 10,000,000 total input tokens a minute.

The Message Batches API has its own limits, so anything that can wait should go there. If several teams share one organization, set workspace limits so one job can’t use up everyone’s quota.

Sharp jumps in usage can trigger acceleration limits, which return 429s of their own. Ramp new traffic up gradually.