Home / How to check your context window in ChatGPT, Claude and Gemini

How to check your context window in ChatGPT, Claude and Gemini

None of the three big chat apps shows you how full your context window is. Here are three ways to find out, from quickest to most hands-off.

Method 1: paste it into a token counter

Copy the conversation, paste it into the free token counter, and read the estimated tokens and the share of each window it uses. It runs in your browser and sends nothing. Good for a one-off check.

Method 2: use a live context meter

ContextMeter adds a small percent pill to ChatGPT and Claude chats and updates it as you go. Open it for tokens used, room left, and the biggest messages. It also reads long chats that load in pieces, and can hand a full chat off to a fresh one. See how it works. We are working on Gemini support and will introduce it soon.

Method 3: use a rule of thumb

Quality slips before the hard limit, so do not wait for an error. A simple habit is to wrap up at around 50 percent and hand off by 70 percent. Count tokens, not messages: one large paste can outweigh fifty short messages. See when to start a new chat.

Know your window size

To turn tokens into a percent you need your window: it varies by plan and model. See context window sizes by plan or the guides for ChatGPT, Claude and Gemini.

Exact counts for developers

For exact numbers, use the provider's own tokenizer or token-counting endpoint. Chat apps add hidden text such as system prompts and memory, so a count of the visible messages is always a lower bound.

Frequently asked questions

How do I check my context window in ChatGPT?

ChatGPT has no built-in meter. Estimate the size with a token counter, or use a browser extension such as ContextMeter that shows the percent of the window used while you chat.

How do I check my context window in Claude?

Claude's chat does not show a running meter. Paste the conversation into a token counter, or use an extension. Claude Code is different: it has its own commands for checking context.

How do I check my context window in Gemini?

The Gemini app has no running meter. Paste text into a token counter to estimate it. Google AI Studio shows token counts for prompts. ContextMeter support for Gemini is coming soon.

Can I get an exact token count?

Only from the provider's own tokenizer. OpenAI publishes one for GPT models, and the Claude and Gemini APIs have token-counting endpoints. For deciding when to start a new chat, an estimate is accurate enough.

Last reviewed 2026-10-08. Numbers change often; see the sources linked on this page.