OpenAI API error 429 rate limit reached

You're building something with the OpenAI API, and suddenly every request returns "Error 429: Rate limit reached for requests." Your application stops working, and you're not sure if you hit a wall or just need to wait.

OpenAI API error 429 rate limit reached

You're building something with the OpenAI API, and suddenly every request returns "Error 429: Rate limit reached for requests." Your application stops working, and you're not sure if you hit a wall or just need to wait.

What's actually happening

Error 429 means you've exceeded OpenAI's rate limits — the maximum number of requests or tokens you can process in a given timeframe. OpenAI sets these limits based on your account tier and usage history to prevent abuse and manage server load.

There are three types of rate limits you might hit:

• RPM (Requests Per Minute) — how many API calls you can make per minute
• TPM (Tokens Per Minute) — how many tokens you can process per minute across all requests
• RPD (Requests Per Day) — daily request cap, mainly for free tier users

Free tier accounts get severely restricted limits (around 3 RPM for GPT-4). Paid accounts start with higher limits that increase automatically as you spend more. If you're on a new paid account and hitting 429s immediately, you're likely maxing out your initial TPM limit — which might be 90,000 TPM for GPT-4 depending on your tier.

The error response usually includes headers showing your current limits: `x-ratelimit-limit-requests`, `x-ratelimit-remaining-requests`, and `x-ratelimit-reset-requests`. Check these in your API logs to see exactly what you hit.

How to fix it

1. Implement exponential backoff in your code

When you get a 429 error, don't immediately retry. Wait a few seconds, then try again. If it fails again, double your wait time. Here's the pattern:

```

First retry: wait 1 second

Second retry: wait 2 seconds

Third retry: wait 4 seconds

```

Most API libraries support this natively. In Python's OpenAI library, you can catch `RateLimitError` and handle it explicitly.

2. Check your current rate limits

Go to platform.openai.com/settings/organization/limits. You'll see your exact RPM, TPM, and RPD limits for each model. If you're hitting these regularly, you need higher limits — not just better retry logic.

3. Batch your requests more efficiently

If you're making hundreds of small requests, you're wasting your RPM limit. Combine multiple operations into single requests where possible. For example, process multiple texts in one API call rather than looping through individual calls.

4. Request a rate limit increase

In your usage limits dashboard, click "Request increase" next to the model causing problems. Fill out the form explaining your use case and expected volume. OpenAI typically responds within 1-2 business days for straightforward requests.

Be specific: "I need 500 RPM for GPT-4 to handle 10,000 customer support tickets daily" gets approved faster than "I need more because I'm hitting limits."

5. Upgrade your account tier if you're on free

Free tier limits are deliberately restrictive. Add a payment method at platform.openai.com/settings/organization/billing. Your limits increase immediately when you move to Tier 1 (paid). You don't need to spend a specific amount first — just having payment details upgrades you.

If you continue getting 429 errors alongside other API failures, you might be dealing with multiple issues simultaneously.

If that doesn't work

Sometimes rate limit increases get denied or delayed. If you need immediate help, contact OpenAI support through the help widget at platform.openai.com. Include:

• Your organization ID (from settings)
• The specific model hitting limits
• Your current limits and requested increase
• A brief use case description

Response times vary — typically 1-3 business days for limit requests. While waiting, implement request queuing in your application so failed requests retry automatically rather than dropping completely.

If you're seeing charges you didn't expect while debugging rate limits, check our guide on unexpected OpenAI charges.

Questions people actually ask

Q: Will my rate limits increase automatically over time?

A: Yes. As you use the API consistently and spend more, OpenAI gradually raises your limits. This happens automatically — no action needed. Most accounts see increases after spending $50-100.

Q: Does error 429 count against my usage or bill?

A: No. Failed requests due to rate limiting don't consume tokens or incur charges. You only pay for successful API calls that return responses.

Q: Can I get 429 errors even with exponential backoff?

A: Yes, if your baseline request volume exceeds your limits. Backoff helps with temporary spikes, but if you're constantly hitting limits, you need a tier upgrade or rate increase — retry logic alone won't solve it.

Q: How long do rate limit windows last?

A: One minute for RPM and TPM limits, rolling. If you max out your TPM at 12:00:30, you'll have capacity again at 12:01:30. It's not a hard reset at the top of each minute.

What to remember

• Error 429 means you've hit RPM, TPM, or RPD caps — check platform.openai.com/settings/organization/limits for exact numbers
• Implement exponential backoff in your code before requesting higher limits
• Free tier limits are extremely low — add payment details to upgrade to Tier 1 immediately
• Request increases through the limits dashboard with specific use case details
• Rate limit errors don't cost money — only successful API responses count toward your bill

---

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