Which has the more generous limits, Claude or Codex?

It depends on the plan and the model, and both change their allowances over time. Comparing percentages on the same afternoon is more useful than any published table; both report exact percentages through their usage endpoints.

Do Claude and Codex limits reset at the same time?

No. Each window resets a fixed time after it opened, and they open whenever you first use that provider, so the resets are unrelated.

Can I see both in one place?

Yes. AI Cooldown shows every Claude and Codex account you connect on one page with countdowns to each reset.

updated September 21, 2026

Claude vs Codex: how the usage limits compare

Both work the same way at the core: a five-hour window that opens on first use and a rolling seven-day window, each reported as a percentage with an exact reset time. Claude adds per-model weekly windows (Fable, Opus, Sonnet); Codex adds per-model availability and purchasable credits. Neither resets at a fixed time of day, and both are per account, which is why people who use both keep more than one.

Side by side

	Claude (Pro, Max)	Codex (Plus, Pro, Business)
Short window	5 hours, opens on first message	5 hours, opens on first request
Long window	7 days, rolling, all models	7 days, rolling
Per-model limits	Separate weekly windows, e.g. Fable, Opus, Sonnet	None as windows; models can be temporarily unavailable
Reset rule	Fixed length from when the window opened	Fixed length from when the window opened
Reported as	Percent used plus reset timestamp per window	Percent used plus reset timestamp per window
Extra usage	Paid extra usage on some plans	Credits on some plans
Check in the CLI	/usage in Claude Code	/status in the Codex CLI
Plans, list price	Pro $20, Max 5x $100, Max 20x $200 per month	Plus $20, Pro $200 per month; Business per seat

Where they differ in practice

Claude's per-model windows are the thing that surprises people. You can be at 30% on the shared weekly window and still be stopped because the Fable window is full. The answer is to switch model, not to wait. Codex does not split its windows by model, but a model can be switched off for your plan for a while, which the usage endpoint reports separately from the windows.

Codex's credits and Claude's extra usage are both a way to keep going past the included windows for money, on the plans that offer them. Neither changes the windows themselves.

Using both

If you switch between Claude Code and Codex depending on the task, the useful view is both providers at once: which one has room right now, and how long until the other is back. AI Cooldown puts a Go or Wait verdict for each provider at the top, ranks your accounts by headroom, and lists every upcoming reset on one timeline. Everything comes from the two official usage endpoints, so the numbers match the CLIs.

see your exact numbers

Want the exact reset for your accounts?

Connect your Claude and Codex accounts to AI Cooldown and get every window with a live countdown, which one is limiting you, and which account to use next. Free, open source, and it reads the same official usage endpoints the CLIs use.

Connect an account

Questions people ask

Which has the more generous limits, Claude or Codex?: It depends on the plan and the model, and both change their allowances over time. Comparing percentages on the same afternoon is more useful than any published table; both report exact percentages through their usage endpoints.
Do Claude and Codex limits reset at the same time?: No. Each window resets a fixed time after it opened, and they open whenever you first use that provider, so the resets are unrelated.
Can I see both in one place?: Yes. AI Cooldown shows every Claude and Codex account you connect on one page with countdowns to each reset.