Codex Usage: How to Check Your Limits and Resets
Codex usage is the share of your plan's quota you have already spent in the current window. You can check it three ways: run /status inside the Codex CLI, open the usage page in your ChatGPT account, or run a free local dashboard that shows the same 5-hour and weekly remaining percentages.
Run inside any active Codex CLI session for instant turn-by-turn percentages:
- ✓Shows 5h rolling & weekly quota %
- ✓Shows exact reset time down to the minute
- ✕Requires entering a CLI session
Check web dashboard directly in your browser without terminal access:
chatgpt.com/.../usageImportant note: Separate from metered API billing at platform.openai.com.
- ✓Official account status & plan tier
- ✕Requires login, no live background alert
A free terminal + local browser dashboard showing your real 5-hour and weekly remaining percentages:
- ✓Terminal + local browser dashboard, refreshes every 60s
- ✓~30-second install, no separate account or API key
- Runs locally; sends no usage telemetry back to us
What Counts as Codex Usage
Codex usage measures model inference calls initiated by the Codex CLI, IDE extension, or Codex Cloud background tasks. Local code compilation, terminal test execution, and file inspection performed on your machine consume zero plan quota.
- •CLI Prompt Turns: Each back-and-forth prompt sent to Codex.
- •Cloud Task Dispatches: Background agent workflows run on Codex Cloud.
- •IDE Context Completions: Full-file and multi-file code revisions in VS Code.
- •Local Command Execution: Shell commands run by Codex locally (e.g.
npm test,git status). - •Local File Reads: Grepping and traversing local disk files.
- •Status Inquiries: Running
/statusor checking config settings.
How to Check Your Codex Usage (3 Methods)
Depending on whether you are currently in a terminal session or managing team seats, you have three primary ways to inspect your remaining quota:
The /status Command in Codex CLI
Whenever you are inside an active Codex CLI session, type /status and press Enter. Codex immediately outputs your remaining 5-hour and weekly quota percentages and the reset times for those limits.
The ChatGPT Account Usage Dashboard
If you are away from the terminal, visit your ChatGPT Settings under the Codex tab at chatgpt.com/codex/settings/usage. This displays your overall subscription status and billing period. Note that this page is separate from the developer API dashboard at platform.openai.com/usage, which tracks pay-as-you-go metered token spend.
Local Codex Usage Dashboard
Instead of running /status manually, install Codex Usage, a free local tool. It reuses your already-authenticated Codex client to show 5-hour and weekly remaining percentages and reset times in your terminal, or in a local browser dashboard (codex-usage --web) that auto-refreshes every 60 seconds and tracks usage since you opened it.
Your Limits by Plan
Codex Quotas by Plan Tier
OpenAI publishes estimated message ranges for Plus and frames Pro as 5x–20x more usage than Plus, but exact weekly allowances aren't published. Last updated: 04 Sep 2026.
| Plan | 5-Hour Window | Weekly Window | Token Ceiling | Ceiling Behavior | Source Type |
|---|---|---|---|---|---|
| Free | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Paused until reset; may fall back to a lower-capability model | Not published |
| Plus ($20/mo) | Est. 15–280 local messages / 5h (model-dependent) | Not published by OpenAI | Not published by OpenAI | Prompts pause until reset, credits, or a paid instant reset | Official (estimate) |
| Pro ($200/mo) | 5x–20x more usage than Plus | Not published by OpenAI | Not published by OpenAI | Prompts pause until reset, credits, or a paid instant reset | Official |
| Business / Enterprise | Custom / seat-pooled (contract-specific) | Custom / seat-pooled (contract-specific) | Contract agreement | Admin visibility; behavior varies by contract | Official |
Looking for the current official reset timing or enterprise seat management? See the complete Codex usage limits table.
When Your Usage Resets
When Does Your Codex Usage Reset?
Codex enforces two distinct clock mechanisms simultaneously: a short rolling window (5 hours) for burst control, and a longer calendar window (weekly) for aggregate volume. Understanding which window you tripped determines how long you wait.
Rolling 5-Hour Window
Codex tracks a rolling 5-hour window rather than a fixed hourly reset. OpenAI does not publish the exact per-turn replenishment mechanics; your current remaining percentage and reset time are shown live in /status.
Weekly Plan Pool
A separate weekly window resets on its own cadence tied to your account. You can check your specific weekly reset anchor at any time by running /status.
Plus and Pro accounts draw from the same two windows but exhaust them at different rates. For your actual current numbers, use /status or the official ChatGPT usage page. See more on when your Codex usage resets.
What Happens When You Hit the Limit
What Happens When You Hit the Limit?
When you hit your limit, your session is temporarily paused. OpenAI does not delete your pending task, discard uncommitted code, or wipe conversation history.
A usage limit does not itself modify your local files, Git branch, or staged changes.
If you reach a limit during an active turn, Codex may finish that turn. Afterward, wait for the displayed reset or use any available credit, reset, or upgrade option.
If urgent, you can switch your CLI profile to metered API keys to bypass plan caps.
Stuck on a 429 or wondering how to bypass the cooldown? Read our comprehensive triage on why am I rate limited and what happens now.
How to Keep Usage Visible While You Work
Stop guessing your quota mid-refactor.
Codex Usage is a free local tool. It reuses your already-authenticated Codex CLI to show 5-hour and weekly remaining percentages and reset times, in your terminal or a local browser dashboard.
- No separate account, OAuth, or API key
- ~30-second install, no sudo
- Local browser dashboard refreshes every 60s, plus a manual Refresh button
- Runs locally; session tracking is in memory only, no telemetry sent back to us