insufficient_quota: the 429 that is not a rate limit

What the error actually says

On the official OpenAI API a billing 429 names its cause in error.code: credit_balance_exhausted (no prepaid credits left), organization_spend_limit_exceeded or project_spend_limit_exceeded (a hard monthly spend limit you set was reached), or organization_usage_limit_exceeded (the monthly usage limit OpenAI assigns your organization). Per OpenAI's error-codes guide (checked 2026-10-09), the broader error.type can still be insufficient_quota. Whatever the code, waiting or backoff does not clear it. An illustrative body for a depleted prepaid balance:

{
  "error": {
    "message": "...",
    "type": "insufficient_quota",
    "code": "credit_balance_exhausted"
  }
}

Why you are seeing it

• Prepaid credits are used up — error.code credit_balance_exhausted, the most common case on pay-as-you-go accounts.
• No payment method is attached, or the card on file was declined when you tried to buy credits, so no new credit arrived.
• A hard monthly spend limit on the organization or the project has been reached — organization_spend_limit_exceeded or project_spend_limit_exceeded; spend alerts alone never block requests.
• The organization reached the monthly usage limit OpenAI assigns to it — organization_usage_limit_exceeded, separate from any limit you set.
• The key belongs to a project whose own spend limit is exhausted, even if another project of yours can still run — per OpenAI, other projects continue unless their own limit or the organization's is reached.

Billing 429 vs rate-limit 429

The official API uses 429 for both. They need opposite responses, so read error.code before changing anything:

Fix it on the official API

• Read error.code first; it names which of the four billing causes you hit.
• credit_balance_exhausted: open the billing settings and check the prepaid credit balance; if it is depleted, add credits, and fix the payment method if a purchase failed.
• organization_spend_limit_exceeded or project_spend_limit_exceeded: raise or remove the hard limit in the organization limits or the project's settings, or wait for the monthly reset; a conservative monthly cap silently reintroduces this error mid-month.
• organization_usage_limit_exceeded: request a higher approved usage limit, or contact OpenAI support.
• Retrying does not clear these codes: after you raise or remove a limit, traffic resumes once the update propagates (per OpenAI's spend-limits guide, checked 2026-10-09).

When the real blocker is payment, not usage

For many developers the loop is unbreakable: the fix requires attaching a card the platform accepts, and no such card exists — regional cards get declined and the account sits at $0 forever. Router One removes that dependency with a prepaid wallet. Top up with a card or Alipay through one hosted checkout, or with USDT/USDC on six chains (Tron, BSC, Ethereum, Polygon, Base, Arbitrum). No US credit card required. Balance-out returns a clear 402 (never a misleading 429), and each API key can carry its own maxSpend cap so one runaway project cannot drain the wallet.

Point your client at the gateway

Your request code does not change — swap the base URL and key, pick any model ID from the /models catalog, and the same call works:

curl https://api.router.one/v1/chat/completions \
  -H "Authorization: Bearer sk-your-router-one-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<model-id-from-/models>",
    "messages": [{"role": "user", "content": "hello"}]
  }'

FAQ