How do I check my Codex usage limits?

Open the usage dashboard at chatgpt.com/codex/settings/usage: OpenAI's pricing page sends you there for your current limits and reset times. Inside a Codex CLI session, type /status, the command the pricing page names for remaining limits. /usage shows how many tokens you used by day, week or in total, and /statusline can add rate limits to the footer. Codex also records the latest reading in its session files on your Mac.

What limits apply

As of October 2026, OpenAI's Codex pricing page says:

• The number of messages you can send depends on the model and on the size and complexity of your tasks. The page gives estimates of local messages per five-hour period for Plus and Standard Business, and says they are not fixed limits.
• Pro plans currently have no five-hour limit. Weekly limits may also apply.
• Local messages and cloud chats share your plan's usage allowance, and cloud tasks may use more of it than local messages.

Method 1: the usage dashboard

Go to chatgpt.com/codex/settings/usage. It is the page the pricing documentation points to for current limits and reset times.

Method 2: /status in a running session

In any Codex CLI chat, type /status. The pricing page says: "If you want to see your remaining limits during an active Codex CLI session, you can use /status." The slash-command reference describes /status differently: a summary of the active model, approval policy, writable roots and current token usage, with no mention of rate limits. If your /status output shows no limits, use the dashboard or the status line below.

Method 3: /usage for token activity

Type /usage to open a usage menu. You can show your ChatGPT token activity, or redeem an earned reset if one is available. To jump straight to a view, type /usage daily, /usage weekly or /usage cumulative.

If the session has no Codex account authentication, Codex shows a sign-in requirement instead.

Method 4: rate limits in the footer

Type /statusline. A picker lists the footer items: model, context stats, rate limits, git branch, token counters, session id and more. Toggle the rate-limit items, reorder, and confirm. The footer updates right away, and the choice is saved to tui.status_line in config.toml.

Method 5: the last reading in the session files

Codex writes every session to a file under ~/.codex/sessions/YYYY/MM/DD/, named rollout-<timestamp>-<conversation_id>.jsonl. As of October 2026, the Codex source code keeps token_count events in those files, and each one can carry a rate-limit snapshot with a primary and a secondary window. Each window holds:

• used_percent: the percentage of the window already consumed (0 to 100), not what is left;
• window_minutes: the length of the rolling window;
• resets_at: when the window resets, as a Unix timestamp in seconds.

To print the latest token_count line, searched in your five most recent sessions (a session you just opened has no reading yet):

grep -h '"token_count"' $(ls -tr ~/.codex/sessions/*/*/*/rollout-*.jsonl | tail -5) | tail -1

The rate limits sit under payload.rate_limits in that line.

We found no user documentation for this file format, only the source code, so it can change between Codex versions.

When do the limits reset?

The dashboard shows your reset times. In the session files, resets_at gives the reset time of each window as a Unix timestamp. If /usage offers an earned reset, you can redeem it from that menu.

Limits of these methods

• The CLI commands need a session. /status, /usage and /statusline all run inside the Codex TUI.
• The two OpenAI pages describe /status differently. Don't rely on one line of its output being there in every version.
• Several terminals, several readouts. With three Codex sessions open, each footer is its own readout, and none of them shows your Claude Code usage.

With Kanbell

Kanbell is a Mac workspace for terminal coding agents. It runs on Apple Silicon Macs and needs git 2.45 or later.

• Codex usage read locally. For Codex CLI, Kanbell shows whether each agent is working, waiting or done, when it waits for permission, and your usage, read locally on your Mac.
• Claude Code in the same app. For Claude Code, Kanbell shows your subscription usage for the session and the week. Gemini CLI exposes no quota, so Kanbell shows none for it.
• Read-only, on your Mac. Kanbell reads, locally and read-only, the session files these tools write: state, model, token counts and subscription usage. That information comes from the agents' own files and may be incomplete or late.
• Your own subscription. Kanbell runs your agent, it doesn't replace it. Usage stays on your account. There is no Kanbell account and no Kanbell server.

For the side-by-side setup itself, see running Codex CLI and Claude Code side by side.