OpenAI API error 429 rate limit how to fix

You're building something with the OpenAI API, and suddenly you hit a wall: Error 429: Rate limit reached. Your requests stop working, your app grinds to a halt, and you're stuck wondering what just happened.

OpenAI API error 429 rate limit how to fix

You're building something with the OpenAI API, and suddenly you hit a wall: `Error 429: Rate limit reached`. Your requests stop working, your app grinds to a halt, and you're stuck wondering what just happened.

What's actually happening

Error 429 means you've exceeded the rate limits OpenAI sets for your account tier. The API enforces three types of limits: requests per minute (RPM), tokens per minute (TPM), and requests per day (RPD). When you cross any of these thresholds, the API refuses new requests until your usage window resets.

Your exact limits depend on your usage tier. New free-tier accounts might get 3 RPM and 40,000 TPM for GPT-4, while paid accounts scale up based on how much you've spent historically. A Tier 1 account (you've paid $5+) gets 500 RPM and 10 million TPM for GPT-3.5-turbo. Tier 5 accounts (you've paid $1,000+) can hit 10,000 RPM and 30 million TPM.

The error message usually looks like this: `Rate limit reached for gpt-4 in organization org-xxxxx on requests per min (RPM): Limit 3, Used 3, Requested 1`. That tells you exactly which limit you hit and how much headroom you had left.

How to fix it

1. Check your current tier and limits

Go to platform.openai.com/settings/organization/limits. This page shows your exact RPM, TPM and RPD limits for each model. You'll also see which usage tier you're on and how much you need to spend to reach the next tier.

2. Implement exponential backoff

Add retry logic to your code. When you get a 429 error, wait a few seconds and try again—doubling the wait time with each failure. Most API libraries have this built in. For Python with OpenAI's library, wrap your calls in a try-except block and use `time.sleep()` with increasing delays.

3. Batch your requests properly

If you're processing multiple prompts, don't fire them all at once. Use the Batch API endpoint for non-urgent requests—it's 50% cheaper and doesn't count against rate limits. For urgent requests, queue them with a delay between each call to stay under your RPM limit.

4. Reduce your token usage

Check response lengths. If you're requesting `max_tokens: 4000` but only using 500, you're wasting tokens. Set `max_tokens` to what you actually need. Each API call counts both prompt tokens and completion tokens against your TPM limit.

5. Upgrade your usage tier

The fastest fix: spend more money with OpenAI. Your tier automatically increases as your total spend grows. Tier 1 ($5 spent) is a massive jump from free tier. Tier 2 ($50 spent) doubles most limits again. You can prepay by adding credits to your account—platform.openai.com/settings/organization/billing.

If you've already spent enough to qualify for a higher tier but limits haven't updated, wait 24-48 hours. OpenAI reviews spend history daily, not in real-time.

If that doesn't work

You might be hitting model-specific limits even if your account tier supports higher rates. Some newer models like GPT-4 Turbo have separate, lower limits during high-demand periods. Check the exact error message—it specifies which model triggered the 429.

For persistent issues beyond tier limits, you need to contact OpenAI's API support team. Go to help.openai.com and click "Messages" in the bottom-right corner. Include your organization ID (platform.openai.com/settings/organization/general), the specific error message with timestamp, and what you've already tried. Reference your current tier and explain your use case. Support typically responds in 1-3 business days for paid accounts, longer for free tier. For more guidance on reaching support effectively, see how to contact OpenAI support.

If you're seeing different API errors alongside the 429—like 401 authentication failures or 500 server errors—you might have multiple issues at play. Check the full OpenAI API error 429 and troubleshooting guide for solutions to authentication, quota and timeout problems.

Questions people actually ask

Q: How long until my rate limit resets?

A: One minute for RPM limits, 24 hours for daily limits. The clock resets on a rolling window—if you used 400 requests at 2:00 PM, those 400 free up at 2:01 PM.

Q: Will upgrading to ChatGPT Plus increase my API limits?

A: No. ChatGPT Plus ($20/month) and API access are separate products with separate billing. API limits only increase with API spend, not Plus subscriptions.

Q: Can I pay to immediately increase my tier?

A: Sort of. Add $100 to your API credits and your tier will update within 48 hours once the system recognizes the spend. You can't pay a fee to skip tier requirements.

What to remember

  • Check platform.openai.com/settings/organization/limits for your exact RPM and TPM thresholds
  • Implement exponential backoff in your code—retry failed requests with increasing delays
  • Use the Batch API for non-urgent requests to bypass rate limits entirely
  • Your usage tier increases automatically as you spend more—Tier 1 starts at just $5 total spend
  • 429 errors reset on a rolling window, usually within one minute for RPM limits

---

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

Related help