You've hit your usage limit (with quota remaining)
Community-reported mitigations, in order: sign out and back in so the CLI refreshes its entitlement state; redo the math against the official rate card, because limits apply per rolling 5-hour window and a full dashboard balance can hide an exhausted window; and if you're genuinely out, buy top-up credits where your plan offers them. There's no official fix for the dashboard/CLI mismatch itself — it's tracked as openai/codex issue #30041.
Affected versions: not pinned to a version range; tracked in issue #30041
The usage dashboard says you have quota left; the CLI and the app refuse to run anything and report that you’ve hit your usage limit. This mismatch is tracked on the openai/codex issue tracker as issue #30041, filed as “Codex CLI quota / entitlement issue: dashboard shows remaining quota but CLI and app says usage limit reached.” Community round-ups point to further reports of the same behavior (issue #16909 among them), though #30041 is the one we’ve verified directly.
Symptoms
- The CLI and the app both block with a usage-limit message.
- The usage dashboard, checked at the same time, shows remaining quota or credits.
- The mismatch is the defining trait — if the dashboard also shows zero, you’ve simply hit your limit.
What causes it?
There is no official root cause. Two explanations circulate in reports, and they aren’t mutually exclusive. The first is a genuine entitlement or quota-sync bug — the CLI’s view of your account disagreeing with the dashboard’s, which is what issue #30041 describes. The second is window math: since April 2, 2026, Codex bills through API-token-aligned credits, and message limits apply per rolling 5-hour window. On GPT-5.6 Sol, Plus allows 15–90 local messages per 5-hour window and Pro 20x allows 300–1,800, per the live pricing docs. A healthy-looking overall balance can coexist with an exhausted current window — which looks exactly like this bug but isn’t one.
How do you fix it?
These are community-reported mitigations, not an official fix:
- Sign out and sign back in. Re-authenticating forces the CLI to refresh its entitlement state, and it’s the first thing to try when the dashboard and CLI disagree.
- Redo the window math. Before calling the limit phantom, check what you’ve run inside the current 5-hour window. The official rate card documents the credit rates — on Sol, 250 credits per 1M input tokens, 25 per 1M cached, 1,500 per 1M output.
- Wait for the window to roll. If the math says you spent it, the reset comes with the rolling 5-hour window, not with the dashboard’s balance.
- Top up credits. Rate-card coverage describes purchasable top-up credits under the post-April-2026 pricing model — if you’re genuinely out, that’s the paved path.
- Add your report to issue #30041. If your dashboard and CLI genuinely disagree, that’s the live thread tracking it.
Our living usage-limits reference tracks the per-plan numbers as they change.
Which versions are affected?
The reports aren’t pinned to a version range, and OpenAI hasn’t marked the entitlement mismatch fixed in a release note we can point to. We’ll update this page when issue #30041 resolves.
Facts on this page verified July 9, 2026.