OpenAI API error 429 rate limit exceeded

You're making API calls and hitting a wall with "Rate limit reached" or error code 429. Your application stops working, your users are locked out, and you're left wondering why OpenAI suddenly cut you off.

OpenAI API error 429 rate limit exceeded

You're making API calls and hitting a wall with "Rate limit reached" or error code 429. Your application stops working, your users are locked out, and you're left wondering why OpenAI suddenly cut you off.

What's actually happening

Error 429 means you've exceeded OpenAI's rate limits — the maximum number of requests or tokens you can process within a specific time window. OpenAI sets these limits per model, per organisation, and they vary based on your usage tier.

The full error typically looks like: `RateLimitError: 429 Rate limit reached for requests` or `Rate limit reached for tokens`. OpenAI tracks three separate limits: requests per minute (RPM), tokens per minute (TPM), and tokens per day (TPD). Hit any one of these thresholds and your next API call fails instantly with a 429 response.

Your usage tier determines your limits. Free tier users get minimal allowances — often 3 RPM and 40,000 TPD for GPT-4o. Paid users on higher tiers get progressively higher limits, sometimes reaching millions of tokens per minute. You can check your current tier and limits at platform.openai.com/settings/organization/limits. The page shows exact numbers for each model you're using.

How to fix it

1. Check your current usage and limits

Log into platform.openai.com and go to Settings > Organization > Limits. This page shows your tier, your specific RPM/TPM/TPD limits for each model, and your current usage. If you're hitting 100% on any metric, that's your bottleneck.

2. Implement exponential backoff in your code

When you get a 429 error, wait before retrying. Start with a 1-second delay, then double it with each failed attempt: 1s, 2s, 4s, 8s. Most API libraries support this automatically — OpenAI's Python SDK includes built-in retry logic with `max_retries=2` by default. Increase this to 5 or higher for production apps.

3. Batch requests and reduce frequency

If you're hitting RPM limits, combine multiple prompts into single requests where possible. Instead of 100 separate completion calls, send 10 requests with batched prompts. For TPM limits, reduce your `max_tokens` parameter — GPT-4o doesn't need 4,000 output tokens for a simple classification task.

4. Request a rate limit increase

Go to platform.openai.com/settings/organization/limits and click "Request increase" next to the model you need. You'll fill out a form explaining your use case and desired limits. OpenAI typically responds within 24-48 hours. They're more likely to approve increases if you have consistent payment history and realistic usage projections.

If the increase request doesn't appear, you might need to upgrade your usage tier first by spending more over time — tiers unlock automatically based on cumulative spend and account age.

If that doesn't work

Sometimes 429 errors come from temporary platform-wide issues, not your actual usage. Check status.openai.com to see if there's an ongoing incident affecting API availability. During major outages, even users well below their limits see 429 responses.

If you're certain you're not exceeding limits and there's no outage, contact OpenAI support at help.openai.com. Include your organisation ID (found at platform.openai.com/settings/organization/general), the exact timestamp of failed requests, and the complete error response. OpenAI support typically replies within 1-3 business days for API issues. For detailed steps on reaching support effectively, see how to contact OpenAI support.

Don't expect instant relief — rate limit increases require manual review. Plan your application architecture to handle 429 errors gracefully from day one rather than treating limits as a problem to solve later.

Questions people actually ask

Q: Why am I getting 429 errors when I just signed up?

A: New free accounts have extremely low limits — often 3 requests per minute for GPT-4o. Add a payment method and make a successful payment of at least $5 to unlock tier 1 limits, which are significantly higher.

Q: Can I pay to skip rate limits entirely?

A: No. Every tier has limits. You can increase limits by requesting higher thresholds or by upgrading tiers through continued spending, but no amount of money removes rate limiting completely.

Q: How long does a rate limit ban last?

A: Rate limits reset every minute for RPM/TPM and every 24 hours for TPD. You're not banned — you're delayed. Wait for the time window to roll over and your quota resets automatically.

What to remember

  • Check platform.openai.com/settings/organization/limits to see your exact rate limits and current usage percentage
  • Add exponential backoff retry logic to every API call — don't assume requests will always succeed
  • Request rate limit increases through the platform with specific use cases, not vague descriptions
  • Free tier limits are minimal — add payment to unlock tier 1 (200+ RPM for most models)
  • Rate limits reset automatically every minute or day depending on the metric

---

*openai-support.com is an independent resource, not affiliated with OpenAI Inc.*

Related help