Fix Claude API 429 Rate Limit Exceeded Error

An HTTP 429 Too Many Requests error from the Claude API indicates that you have exceeded your rate limits. Anthropic enforces limits on both Requests Per Minute (RPM) and Tokens Per Minute (TPM), as well as daily token limits. These limits scale based on your account's usage tier.

When your application triggers a 429 error, the API stops serving your requests until the rate limit window resets. Use the following practical steps to resolve and prevent these errors.

1. Inspect the Rate Limit Headers Before changing code, analyze the response headers returned by the Anthropic API. Every API response includes metadata about your current usage state. These headers tell you exactly why you were rate-limited and when you can retry.

Look for the following headers in your failed request response: * anthropic-ratelimit-requests-limit: The maximum number of requests allowed per minute. * anthropic-ratelimit-requests-remaining: The number of requests you have left in the current window. * anthropic-ratelimit-requests-reset: The ISO 8601 timestamp indicating when your request limit resets. * anthropic-ratelimit-tokens-limit: The maximum number of tokens allowed per minute. * anthropic-ratelimit-tokens-remaining: The remaining tokens you can use in the current window. * anthropic-ratelimit-tokens-reset: The ISO 8601 timestamp indicating when your token limit resets.

Logging these values in your application's error monitoring tool will help you identify whether your bottleneck is request frequency (RPM) or payload size (TPM).

2. Implement Exponential Backoff with Jitter Do not immediately retry failed requests in a rapid loop, as this will compound the rate-limiting block. Instead, implement an exponential backoff algorithm with randomized delay (jitter).

When your client receives a 429 status code, pause execution before retrying. Increase the delay exponentially with each successive failure. Adding "jitter" (random variation) prevents a cluster of threads from hitting the API at the exact same millisecond when the rate window resets.

Here is a conceptual Python example using the standard time and random libraries:

`python import time import random

def retry_with_backoff(api_call_func, max_retries=5): base_delay = 1.0 # start with a 1-second delay for attempt in range(max_retries): try: return api_call_func() except Exception as e: if "429" in str(e) and attempt < max_retries - 1: # Calculate exponential delay with randomized jitter delay = (base_delay * (2 ** attempt)) + random.uniform(0, 1) time.sleep(delay) else: raise e `

3. Reduce Token Consumption per Request If you are hitting the Tokens Per Minute (TPM) limit, you must optimize how much text you send and request from Claude.

• Trim System Prompts: Avoid passing large, repetitive system prompts with every single request. Keep system instructions concise.
• Limit Conversation History: In conversational applications, do not pass the entire chat history back to the API. Implement a sliding window that only sends the last 5 to 10 turns, or summarize older turns.
• Lower max_tokens: Set the max_tokens parameter only as high as necessary. Although you are billed for actual generated tokens, some systems use your max_tokens limit to reserve capacity, which can trigger rate limits prematurely.

4. Check Your API Usage Tier and Upgrade Anthropic dynamically assigns rate limits based on your lifetime spending tier. New accounts start on Tier 1, which has highly restrictive limits.

To view your current tier and upgrade: 1. Log in to the Anthropic Console. 2. Navigate to the Billing section. 3. Check your lifetime spend or deposit history. Upgrading to a higher tier requires prepaying a minimum balance (for example, depositing $40 or more to move from Tier 1 to Tier 2). 4. Once your payment clears, your rate limits will scale automatically within a few minutes. Check the Limits tab in the console to confirm your new limits.

When to escalate If your organization has upgraded to the highest self-service tier and you still regularly exceed the token or request limits, you must contact Anthropic directly. Go to the **Limits** page in your Anthropic Console and click the "Request Limit Increase" link to submit a formal request for custom enterprise limits. Provide detailed metrics on your current queries-per-second requirements to speed up approval.