The complete guide to AI coding assistant usage limits (2026)
Claude, Codex, and more all cap your usage differently. This guide explains every limit type, when they reset, and how to track them across providers.
Every major AI coding assistant now caps how much you can use per subscription — and each does it differently. If you've ever been blocked mid-task, this guide is the map: what the limits are, when they reset, and how to keep an eye on all of them at once.
The two limit shapes
Almost every provider uses some mix of a short rolling window (hours) and a longer window (weekly). Short windows stop marathon sessions; long windows cap total spend. The trap is that you can pass one and fail the other.
Claude (Code, Max, Pro)
A rolling 5-hour session limit plus weekly caps (all-models, and a tighter Opus cap on Max). See our Claude limits guide for the details.
OpenAI Codex / ChatGPT
Usage metered weekly against your ChatGPT plan, with separate buckets for standard and GPT-5.x-Codex models and reset dates that move. See our Codex limits guide.
Why a single dashboard beats per-tool apps
Most trackers are Mac-only menu-bar apps that read one tool on one machine. How Much Usage is a web dashboard that aggregates every account across providers with reset countdowns and alerts — so one glance tells you where everything stands. More providers coming.
Frequently asked questions
- Do all AI coding assistants have usage limits?
- Yes — Claude, Codex/ChatGPT, and others all cap usage per subscription, typically with a short rolling window plus a weekly cap.
- How do I track usage across different AI providers?
- How Much Usage shows Claude and ChatGPT/Codex on one dashboard, across every account, with reset countdowns and warnings.