Plan limits and reset rules
How each subscription tiers its quota, and what triggers a reset window.
Read this before trusting the numbers
Vendors change limits without notice and their documentation is often vague. This table is a manually maintained summary. Treat it as a starting point, then check the official source.
| Provider | Plan | Rolling window | Reset behaviour | Official source |
|---|---|---|---|---|
| OpenAI Codex | ChatGPT PlusNot yet human-verified | Short rolling window, plus a longer cap | The short window rolls continuously. Vendor-issued resets are separate and announced without a schedule. | OpenAI pricing / in-product usage panel |
| OpenAI Codex | ChatGPT ProNot yet human-verified | Larger short rolling window, plus a longer cap | Same structure as Plus with higher ceilings. Resets are issued at the vendor's discretion. | OpenAI pricing / in-product usage panel |
| OpenAI Codex | ChatGPT BusinessNot yet human-verified | Workspace-level allocation across seats | Governed by workspace settings; individual behaviour differs from personal plans. | OpenAI business documentation |
| Claude Code | Claude ProNot yet human-verified | Session-scale rolling window | Session window rolls continuously. Vendor-issued resets are not on a published schedule. | Anthropic pricing |
| Claude Code | Claude MaxNot yet human-verified | Larger session window plus an outer recurring cap | Outer cap clears on a rolling basis; vendor-issued resets add on top. | Anthropic pricing |
| Grok | X Premium+ / SuperGrokNot yet human-verified | Tier-dependent; revised frequently by the vendor | Check xAI's current documentation — this is the tier that changes most often. | xAI |
| Gemini | Google AI Pro / UltraNot yet human-verified | Per-model daily-style limits rather than one account pool | Limits are organised by model, so one model can be exhausted while others remain usable. | Google One plans |
The verified-on date per row is the last time a human checked it. Nothing in this table is scraped automatically.