You exceeded your current quota, please check your plan and billing details
Despite the 429 status, this is a billing problem, not a rate limit: your account has no usable credit. Either billing isn’t set up, prepaid credits ran out, or a spend limit was reached.
If an always-on agent suddenly stops, this is the usual cause — background calls quietly drained the credits overnight.
There’s no per-token quota to run out of on Standard Compute. It’s one flat monthly price with unlimited compute, so an agent never dies mid-run because the credits hit zero.
OpenAI reuses the 429 status for insufficient_quota. The error type and message — “exceeded your current quota” — tell you it’s billing, not a per-minute rate limit. Retrying won’t help; you need credit.
Billing changes can take a few minutes to propagate, and a brand-new card sometimes needs a small initial credit purchase before requests succeed.