Claude Code vs Codex usage limits: how each one counts
Both use a five-hour window with weekly limits on top, but the checks, the sharing rules and the fallbacks differ.
By ClaudeGPT Si Newsroom, Editorial team
2 min read
Claude Code vs Codex usage limits come down to the same basic shape: both meter subscription use in a rolling five-hour window with weekly limits on top, but each vendor counts, reports and extends them in its own way. If you run both agents, the simplest way to keep the two sets of limits in view is UsageAtlas, a free, open-source desktop tracker that shows each tool's windows and reset times next to a combined total.
We leave the actual allowances out on purpose. Both companies change them, and both say usage depends on the model and the task.
How Claude Code counts
Anthropic's Pro and Max plans have a five-hour session limit and a weekly limit, and that usage is shared between the Claude apps and Claude Code. Work in the chat app and work in the terminal draw from the same pool.
Inside a session, run /usage. On Pro, Max, Team and Enterprise plans it shows plan usage bars plus a breakdown of what counts against your limits. When you hit a limit, usage credits let you keep going, and you can manage them with /usage-credits. One trap: if ANTHROPIC_API_KEY is set in your shell, Claude Code bills that key instead of your subscription, so your plan limits stop being the thing to watch.
How Codex counts
OpenAI's pricing page gives estimates per five-hour period for Plus and Standard Business, says Pro currently has no five-hour limit, and notes that weekly limits may also apply. Local messages and cloud chats share the same allowance. OpenAI is explicit that these are estimates, not fixed message counts.
In the CLI, /status shows the chat ID, context usage and rate limits. The full picture lives on the Codex usage dashboard at chatgpt.com. When you run out, Plus and Pro users can buy credits, and anyone can run extra local chats with an API key billed at API rates.
Side by side
| Claude Code | Codex | |
|---|---|---|
| Short window | Five-hour session limit | Five-hour period on Plus and Business |
| Longer window | Weekly limit | Weekly limits may apply |
| Shared with | Claude apps | Codex cloud chats |
| In-session check | /usage |
/status |
| Web view | Settings > Usage on claude.ai | Codex usage dashboard |
| Past the limit | Usage credits | Credits or an API key |
Where the two collide
The trouble starts when you use both. Claude's window might reset at 3pm while Codex's weekly allowance is nearly spent, and neither tool tells you about the other. Developers end up opening two commands and a browser tab to answer one question: which agent still has room right now?
Seeing both limits at once
UsageAtlas reads Claude's five-hour and weekly windows and Codex's session and weekly windows, and shows remaining capacity with the local reset time for each. You can set a threshold per window and get one desktop alert per reset cycle. A tool filter shows Claude or Codex alone; with no filter you see the combined tokens and estimated cost for the day.
It runs on Windows, macOS and Linux, needs no account for local tracking, and does not read prompts or responses. Provider data can lag, so if a deadline depends on the exact figure, check /usage or the Codex dashboard as well.
If you only use one of the two tools, its own command is enough. The tracker pays off once both are in your week.