AI coding usage limits by plan: Claude Code, Codex and OpenCode Go
Claude Code, Codex and OpenCode Go all cap usage in a rolling five-hour window and a longer window behind it. The windows are measured differently and reset at different times, so here they are side by side.
Short answer. Claude Pro and Max have a five-hour session limit plus a weekly limit that resets at a fixed time assigned to your account. ChatGPT Plus has a five-hour Codex limit again since 25 August 2026, while ChatGPT Pro currently has none; weekly limits may apply to both. OpenCode Go limits each model by dollars spent: 20% of the monthly allowance per five hours, 50% per week, 100% per month.
Limits by plan
| Service | Plan | Price | 5-hour window | Longer window | Past the limit |
|---|---|---|---|---|---|
| Claude Code | Pro | $20/mo ($17/mo yearly) | Yes, session limit | Weekly, across all models | Usage credits at API rates |
| Claude Code | Max 5x | $100/mo | 5× Pro per session | Weekly | Usage credits at API rates |
| Claude Code | Max 20x | $200/mo | 20× Pro per session | Weekly | Usage credits at API rates |
| Codex | ChatGPT Plus | $20/mo | Yes (restored 25 Aug 2026) | Weekly may apply | Credits |
| Codex | ChatGPT Pro | from $100/mo | None at the moment | Weekly may apply | Credits |
| Codex | Business | $20/user/mo | Yes on Standard | Weekly may apply | Credits |
| OpenCode | Go | $10/mo | 20% of the monthly limit | Weekly 50%, monthly 100% | Free models or Zen balance |
| OpenCode | Go Plus | $40/mo | 20% of the monthly limit | Weekly 50%, monthly 100% | Free models or Zen balance |
Claude limits are shared between claude.ai and Claude Code. OpenAI publishes Codex allowances as message ranges per model, because the real count depends on model, context, reasoning and tools. OpenCode Go limits are per model and in dollars, so cheaper models allow more requests.
Claude Code
Every Claude plan has a session limit that resets every five hours; the window opens with your first message. Pro and Max add a weekly limit across all models that resets at a fixed time each week assigned to your account. Max 5x gives five times the Pro per-session allowance and Max 20x twenty times.
Where to check: run /usage in Claude Code, or open Settings → Usage on claude.ai, which also shows the next reset. At the limit you can wait, move to a higher plan, or turn on usage credits to keep working at standard API rates. Anthropic occasionally hands out a free limit reset that refills the session or weekly limit at once; it is used from Settings → Usage, not from the terminal. Claude Code limits in detail →
Codex
Codex usage comes with a ChatGPT plan. OpenAI restored the five-hour limit for Plus on 25 August 2026; Pro plans ($100, $200 and $500) have no five-hour limit for now, and weekly limits may apply on top. The five-hour timer starts with your first message after the previous window ends, not on the clock.
Where to check: type /status in a Codex CLI session to see what is left in both windows, or open the usage dashboard in ChatGPT. Past the limit, credits are charged per million tokens. Codex limits in detail →
OpenCode Go
OpenCode Go ($10) and Go Plus ($40) give every model its own monthly dollar allowance. Five hours may use 20% of it and a week 50%. When a window runs out, OpenCode’s free models still work, and “Use balance” falls back to Zen credits. How OpenCode Go limits work, in detail →
Seeing them all in one place
Each service only shows its own numbers, and only when you ask: a slash command, a settings page, a console. If you switch between agents when one runs dry, that is three places to look. A dozen Mac apps now collect them for you, from menu bar trackers like CodexBar to notch apps; here is how they compare. CapaTheNotch puts them where you already look, in the MacBook notch.
CapaTheNotch: your limits in the notch
Any two of Claude Code, Codex and OpenCode Go at a glance: what’s left in the five-hour and weekly windows and when each resets, with one notification when a window drops below 10%. It reads Codex and Claude Code without their credentials. Free and open source, for Apple Silicon Macs on macOS 14 or later.
Download for macOSSources
- Claude plans and pricing (Anthropic)
- What is the Max plan? (Anthropic)
- Using Claude Code with your Pro or Max plan (Anthropic)
- What is a limit reset? (Anthropic)
- ChatGPT and Codex pricing (OpenAI)
- OpenAI restores 5-hour Codex limits for ChatGPT Plus (9to5Mac)
- OpenCode Go (OpenCode docs)
FAQ
Do claude.ai chats count against my Claude Code limits?
Yes. On Pro and Max plans, usage limits are shared across Claude and Claude Code, so chatting on claude.ai and coding in Claude Code draw from the same allowance.
When does the Claude weekly limit reset?
At a fixed time each week that Anthropic assigns to your account. It is not the same day for everyone. Settings → Usage on claude.ai shows your next reset.
Does ChatGPT Pro have a five-hour Codex limit?
Not at the moment. OpenAI’s pricing page says Pro plans have no five-hour limit; Plus and Standard Business do. Weekly limits may still apply.
What happens when OpenCode Go runs out?
You can keep using OpenCode’s free models, or turn on “Use balance” in the OpenCode console to continue on Zen credits.
Is there one place to see all three?
Each service only shows its own limits. CapaTheNotch, a free open-source Mac app, puts any two of Claude Code, Codex and OpenCode Go side by side in the MacBook notch.