Codex usage limit reached: what to do
The useful response is not to keep retrying. Identify the exhausted window, preserve remaining capacity elsewhere, and choose between waiting, conserving, or changing the task plan.
Identify what actually ran out
Read the exact message and usage counter in the official product. A five-hour limit is an immediate burst constraint; a weekly limit is a longer budget. Workspace policies or service incidents can produce different symptoms and should not be treated as quota exhaustion.
Check the account and workspace context before comparing with another user. Different subscriptions and organizational controls can expose different limits.
Choose the least disruptive response
If the reset is close and the work is not urgent, waiting is usually clearer than repeatedly starting failed runs. If weekly capacity remains but a short window is exhausted, organize prompts, code context, and acceptance criteria offline so the next session starts efficiently.
- Stop automatic retries that cannot succeed.
- Save local work and task state.
- Move low-cost preparation outside the blocked session.
- Use a banked reset only when the product offers one and urgency justifies it.
Can changing models solve it?
Model choice and reasoning effort can affect usage, but changing them does not guarantee access when the relevant account limit is already exhausted. Follow the options shown by the official interface and avoid assuming that a cheaper task bypasses a hard window.
When to troubleshoot instead of wait
If the displayed reset time has passed, the counter does not refresh, or the error looks like an outage, verify service status and the signed-in workspace. Contact official support for account-specific resolution. Never give a tracker your session cookie, API key, or login credentials.