First, check how you signed in.
Codex supports ChatGPT sign-in for subscription access and API-key sign-in for usage-based access. API-key use is billed through your OpenAI Platform account, so it should not be read as the same balance as your included subscription allowance. See OpenAI’s authentication guide.
- Confirm the account and workspace. In the CLI,
codex login statusshows the active authentication method. - Open the official usage dashboard. Open the Codex usage dashboard. Check the selected account, each remaining allowance and its reset time.
- Check during a CLI session. Enter
/statusin an active session. The official command reference explains the status view.
Keep a CLI reading visible
Use /status for a session check. In supported Codex CLI versions, /statusbar configures the persistent status bar and its available items. Check the installed version’s options for model, context and usage windows. Context capacity concerns the current conversation; it is separate from account allowance.
Which limits apply, and when do they reset?
The current OpenAI plan documentation distinguishes these cases:
- Plus and Standard Business: OpenAI gives estimated local-message ranges per five-hour period. These are estimates, not guaranteed message counts.
- Pro: Pro plans currently have no five-hour limit. This does not mean unlimited usage; check the other limits shown for your account.
- Weekly allowance: weekly limits may also apply. Local messages and cloud chats draw on your plan’s allowance.
Read the remaining amount and reset time together for each window. A shorter window resetting may still leave a weekly limit in effect. Use the times in your account dashboard instead of assuming a daily reset at midnight.
Model choice, context, reasoning and tool use affect how much a task consumes. A remaining percentage cannot tell you exactly how many more prompts will fit. For lighter work, try a suitable smaller model and include only the files the task needs.
Read the reset type before using it.
| Type | Effect on the window | Expiry and history |
|---|---|---|
| Scheduled | The applicable window renews on the account schedule. Read its supplied reset time. | This is a window renewal, not a saved reset balance. |
| Banked | Receiving a reset stores it; applying an eligible reset restores the affected allowance. A window already at 100% is not restarted by that banked reset. | Expiry follows the grant/account terms. The official 30-day banked activity history is not an expiry rule or the Available balance. |
| Automatic / global promotion | Applies only to the accounts and limits covered by the event. An announcement alone does not prove your eligibility. | Check the event terms. Paid and global events are outside the banked activity history. |
| Paid instant reset | For eligible accounts it applies immediately and cannot be stored or stacked. After application, the first subsequent Work or Codex request starts the new weekly period. | Review the official offer and receipt. If the allowance did not recover, check the account and contact support before buying again. |
Eligibility and controls belong to OpenAI. Refreshing TokenVive only fetches supported readings; it does not redeem or buy resets, manage banked balances or add credits.
Still blocked, or using allowance faster?
- Match the sign-in and workspace. Compare the account used by the editor with the official dashboard. API-key billing and ChatGPT subscription access have separate limits.
- Identify the actual limiting window. Read every applicable short, weekly and model-specific notice. A remaining short window does not clear an exhausted weekly allowance.
- Read the error category. A rate-limit message, context-capacity warning, authentication failure and network error require different checks. Keep the exact non-private wording and time.
- Check current service information. Use OpenAI Status for reported incidents and the official help links below for account issues. For unexpected consumption, inspect the model, context, reasoning, tool use and any task still running in the original client.
TokenVive displays supported account data. Stopping tasks, resolving access restrictions, billing disputes and refunds must be handled in the relevant service.
Looking for past use?
A remaining percentage answers what is available now. Codex history in TokenVive separates daily token totals across all models from message counts with available model breakdowns. Those fields can cover different dates and update separately.
Keep supported Codex data close at hand.
TokenVive lets you view supported Codex usage limits and reset information alongside your other connected AI services. Connect Codex in Settings on iPhone, finish the provider’s sign-in flow, then check the account and update time in the usage view.
Codex also supports message and token analysis in TokenVive, subject to the history and fields available to your account. Those records are separate from remaining usage: TokenVive does not turn a percentage into a fabricated message or token count.
After opening TokenVive on both paired devices, your Apple Watch can check supported usage limits with the credentials synced from iPhone. Sign-in and expired-session reconnection stay on iPhone. Explore AI tracking and supported services.
To display a specific account and window on your Home Screen or Lock Screen, follow the iPhone AI usage widget tutorial.
TokenVive displays provider data; it cannot increase your Codex allowance or change OpenAI’s reset schedule. AI connection and refresh features require active TokenVive membership or remaining trial access.
Choose the usage limit you want to monitor.
Use the arrow beside your Codex account to expand its usage-limit options. Select the available period you want to keep in view, then check its label and reset time. Switching the selected period changes which allowance the ring displays; it does not transfer allowance between periods.
When two readings look different.
- Account: compare the same personal or work account.
- Window: compare the same limit period, rather than a short-term allowance with a weekly one.
- Freshness: a failed TokenVive refresh preserves the last successful result and its timestamp. Try a manual refresh or reconnect if prompted.
- Device: background transfers and widget updates may arrive later than a foreground refresh.
If the result still looks wrong, use the official account view to check the current provider state and follow TokenVive’s connection troubleshooting.
Official sources
This guide was checked against these OpenAI documents on . Plan rules can change.
The reset-type explanation was checked separately on .
- Pricing, usage limits and the usage dashboard
- CLI commands and the status view
- ChatGPT and API-key authentication
- Banked and automatic Codex resets
- Paid instant resets and the next weekly period
TokenVive is an independent app. Codex and OpenAI names identify the services discussed here.
Reset details and the status-bar discussion were read November 15, 2026. This page was edited November 16; that is not a new test of every plan.