Live

Are we in
peak hours?

Real-time monitoring of Claude's peak hours. Optimize your usage and get the most out of your subscription.

๐ŸŸข
OFF-PEAK
All clear โ€” use Claude without restrictions

Your 5-hour session runs at normal pace. Great time for heavy tasks, long conversations, and intensive Claude Code sessions.

--:--:--
Your local time
Next change in
--:--:--
Tips for right now
Reference schedule (weekdays only)
๐Ÿ—“๏ธ Weekends have no peak hours โ€” normal limits apply all day
Alternative: OpenAI Codex
๐Ÿค–
RECOMMENDED COMPLEMENT
OpenAI Codex โ€” your Plan B during peak hours

Codex is OpenAI's coding agent, bundled with ChatGPT plans. Its limits are completely independent from Claude, so you can use it without affecting your Claude quota.

๐Ÿ’ฐ

Plans & pricing

Plus โ€” $20/mo: Includes Codex with GPT-5.x model access. Great for individual devs and moderate use.
Pro โ€” $200/mo: Much higher limits, ideal for intensive all-day usage.
Business โ€” $30/user/mo: For teams, with a guarantee your data won't train the model.
There's no standalone Codex subscription โ€” it comes bundled with ChatGPT.

๐Ÿ“Š

Limit system

Codex uses a credit-based system tied to the model and task complexity. Limits reset in 5-hour windows (similar to Claude). Local tasks (CLI/IDE) consume fewer credits than cloud tasks. You can switch to the Mini model to stretch your limits by 2.5x to 3.3x.

๐Ÿš€

Active promotion (limited time)

Codex currently has doubled rate limits across all paid plans, and is temporarily available on ChatGPT Free and Go plans too. Great moment to try it as a Claude complement.

โš™๏ธ

Available models

GPT-5.4: Most powerful, for complex reasoning and coding tasks.
GPT-5.3-Codex: Good cost/capability balance, ideal for GitHub code reviews.
GPT-5.1-Codex-Mini: Cheapest per task, perfect for routine work. No cloud support.
You can also use the API with per-token billing if you prefer pay-as-you-go.

๐Ÿ”—

Where to use it

Codex works on the ChatGPT web app (Codex Cloud), as a VS Code IDE extension, and as a local CLI tool. It also integrates with GitHub for automated code reviews, Slack, and Linear. Visit codex.openai.com to get started.

What does this mean?
๐Ÿ“Š

Tokens, not clock time

Claude measures your 5-hour session in tokens, not real minutes. During peak hours, each message "costs" more tokens, so your session runs out faster.

๐Ÿ‘ฅ

Affects ~7% of users

Pro subscribers are the most impacted. Max 20x users barely notice any difference (~2% only). API customers are not affected at all.

๐Ÿ“ˆ

Weekly limits unchanged

Your total weekly quota doesn't change. It's only redistributed: less capacity during peak hours, more during off-peak. The total remains the same.