ChatGPT Context Window by Plan: Free, Go, Plus, Pro, Business, Enterprise

ChatGPT's context window depends on your plan and on whether you use GPT Instant or a reasoning model. OpenAI lists a GPT Instant context window of 27K tokens on Free, 54K on Go, Plus and Business, and 128K on Pro and Enterprise. Reasoning models get 256K tokens on Go, Plus, Business and Enterprise and 400K on Pro, while Free is listed as "varies". The part you can fill with your own text is smaller than the total, because instructions, memory and reasoning share the window.

ChatGPT context window by plan

Source: the plan comparison on OpenAI's ChatGPT pricing page, checked October 11, 2026. The page-count column is OpenAI's own approximation of how much text you can enter.

Why you can paste less than the full window

The context window is shared. OpenAI says ChatGPT uses one window to understand your request, track the conversation, retrieve relevant information and write the response, and that part of it goes to system instructions, including tools and personality, to memories if you have them turned on, and to internal processing such as reasoning. That is why the pricing page lists a smaller "input maximum" next to each total, and why OpenAI calls that figure an approximation that can change with the features you use and what is stored in memory.

In practice, a long conversation, large pasted documents and enabled memory all compete for the same space.

GPT Instant versus reasoning models

The pricing page gives each plan two figures: one for GPT Instant, the fast everyday mode, and a larger one for reasoning. On every paid plan the reasoning window is the bigger of the two, so for long documents or long analytical threads a reasoning model has more room. The gap is widest on Go, Plus and Business, where reasoning has 256K tokens against 54K for Instant.

Pro is the only plan that raises both figures, to 128K for Instant and 400K for reasoning. Enterprise matches Pro's 128K Instant window but keeps reasoning at 256K, the same as Plus and Business.

Files, Codex and the API

Uploaded files have their own limits. A single document can hold up to 2 million tokens, far more than any plan's window, and the per-file and upload limits are listed in ChatGPT file upload limits by plan.

The windows in the table describe ChatGPT conversations. ChatGPT Work, Codex and the OpenAI API follow the limits of the model you choose. OpenAI's Enterprise rate card, for example, prices Work and Codex input above 272K tokens at a separate long-context rate, which means those tools can use more context than a chat window allows.

What to do when a conversation gets too long

1. Start a new chat for a new task. A fresh conversation frees the space taken by earlier turns.
2. Ask for a summary before you switch. Paste the summary into the new chat instead of the whole history.
3. Use a reasoning model for long material. On every paid plan it has a larger window than GPT Instant.
4. Upload long documents instead of pasting them. Files have their own size limits, listed on the file upload page.
5. Move to Pro if you regularly need more. Pro has the largest windows: 128K for Instant and 400K for reasoning.

Which plan fits

• Free for short questions and small documents; the 27K Instant window is roughly 12 pages.
• Go, Plus or Business for most long documents, especially with a reasoning model and its 256K window.
• Pro when you work with very long documents or codebases in chat and need the 400K reasoning window.
• Enterprise when a company needs the larger 128K Instant window across a managed workspace.

For prices and everything else each plan includes, see ChatGPT pricing. To compare Plus and Pro on usage as well as context, see ChatGPT Plus vs Pro.