You've hit your usage limit. Upgrade your plan or add credits to continue, or try again later.
Codex usage on ChatGPT plans is metered in rolling windows (roughly 5-hourly plus weekly). Heavy agent sessions exhaust the window; it clears on reset. Codex CLI can also run on your own API key or any custom OpenAI-compatible provider — which removes the plan meter entirely.
Cloud tasks and parallel runs multiply usage silently — three parallel tasks is three windows' worth of burn.
Codex CLI supports custom providers natively: add a [model_providers] entry with base_url https://api.stdcmpt.com/v1 and model "standardcompute", and Codex runs at a flat monthly price — no 5-hour window, no weekly cap. Setup guide: /integrations/codex-cli.
Yes — windows are rolling and clear automatically; the message usually indicates when. Weekly caps reset weekly.
It's unmetered by the plan but billed per token — heavy use gets expensive. A flat-rate OpenAI-compatible provider gives unmetered *and* fixed-cost.