Claude.ai usage limits: the five-hour budget explained

If you have used Claude for a few days, you have probably seen the message that says you have reached your limit. That is not a bug. It is the budget talking. This article explains how Claude.ai’s usage limits work: the five-hour session limit, the weekly limits on paid plans, what drains the budget, how long one conversation can get, and how to make everything last longer. It applies to claude.ai in a browser.

The five-hour session limit

The Free plan has a session-based usage limit that resets every five hours. Anthropic is open about the fine print: the number of messages you can send varies with demand, and Claude notifies you when you reach your limit or when a prompt exceeds the available context window. A busy hour can therefore feel tighter than a quiet one.

Check where you stand

1. Click your initials in the lower left corner and select Settings.
2. Open Usage.
3. Read your current session usage and the time of the next reset. On paid plans the weekly limit and its reset are shown there too.

If a message will not send, this screen is the first place to look.

What paid plans add

Paid plans run two limits at once: a session limit that resets every five hours, and weekly limits that reset once every week. Paid plans do not remove the five-hour window; the weekly limit sits on top of it. Max comes in two sizes, one with five times Pro’s usage and one with twenty times, for long sessions and agentic work. Prices vary by region, so check Anthropic’s pricing page for yours.

What drains the budget

What you do changes how fast the budget drains. Anthropic names these factors:

• Conversation length and complexity. Longer threads consume more.
• The features you use. Tools such as web search and Research consume usage, as does creating artifacts and running multi-step tasks.
• File attachments. Attachment size counts. A folder of full-resolution photos is an expensive way to ask one question.
• The model you chat with. Model choice is one of the listed factors.
• The effort level. Higher effort means more thorough responses that take longer and use more tokens, so you reach your limit faster.

Your usage is pooled

Everything you do across claude.ai, Claude Code, and Claude Desktop counts toward the same limit, so a long session in one place leaves less for the others. If you spend your budget in Claude Code, you will feel it in the browser too.

How long one conversation can get

Length limits are separate from usage limits. The context window is how much a single chat can hold, and it depends on the model. On paid plans the newest models support up to a 1M token context window, while others support 500K or 200K tokens. A portion of the window is always reserved for Claude’s response, so the longest possible conversation is slightly smaller than the number in the spec sheet. We found no published figure for the Free plan specifically.

If you have code execution enabled, Claude manages long chats for you: when a conversation approaches the context window, it summarizes earlier messages to continue. Chats that trigger this summarizing use more of your budget, so when you are close to your cap, start a new chat instead of pushing the old one further.

Eight ways to make the budget last

1. Batch related questions into one message instead of sending them one at a time.
2. Review your prompt before you send it, so you do not spend a turn on a clarification.
3. Put documents you reuse in a project. Project content is cached, and cached content counts less against your limits when reused.
4. Remember the cache expires. Come back to a project after a long break and your first message counts that content in full again.
5. Use chat search and memory (paid plans for search) rather than re-pasting context you already gave Claude. Article 5 covers both.
6. Turn off tools you do not need. Connected apps and web search use a lot of tokens; switch off what a conversation does not use.
7. Lower the effort level for routine tasks, and switch thinking off when you do not need it. Article 2 explains both settings.
8. Watch Settings > Usage instead of guessing.

Frequently asked questions

Is the five-hour limit the same for everyone on the Free plan?

No. The number of messages you can send varies with demand. Anthropic states this explicitly rather than promising a fixed number.

Do paid plans remove the five-hour limit?

No. Paid plans run a five-hour session limit and weekly limits together. The next reset time for both is in Settings > Usage.

Does everything I do count against the same limit?

Yes. Usage is pooled across claude.ai, Claude Code, and Claude Desktop, so a long session in one place leaves less for the others.

What is the difference between a usage limit and a length limit?

A usage limit is your budget over time, five hours or a week. A length limit is the context window: how much a single conversation can hold. You can hit either one on its own.

Wrapping up the series

That is the whole picture: a session limit that resets every five hours, weekly limits on top for paid plans, pooled usage across every surface, and a context window that caps how long one chat can run. Watch Settings > Usage, batch your questions, and start a fresh chat when a thread gets long. This is the last article in the Claude-on-the-web beginner series. For running AI on your own machine, the Agentic AI series collects everything else on this site.