OpenAI sells ChatGPT six ways. Four are for individuals (Free, Go, Plus, Pro), two are for organisations (Business and Enterprise), and then there is the API, which is not a plan at all but a meter. The prices below come from our catalogue at the time of writing; the live version is on /pricing/chatgpt-pricing.
What each plan is actually worth
Free ($0) is not a trial. It has a current model, search, file upload and voice, with limits that reset over time. If you use ChatGPT a few times a week, this is your plan and you should not be paying anything.
Go ($8) is the tier for people who hit the free limits regularly but mostly do light work: emails, questions, short drafts. It buys more usage, not a different class of tool.
Plus ($20) unlocks the full model picker, higher limits, Projects and the advanced features OpenAI ships first. If ChatGPT is part of your working day, this is the default. It is also the plan most likely to be overpaid for by people whose real usage fits Go or Free.
Pro ($200) is ten times Plus. It exists for people who run out of Plus usage every day and need the longest-thinking Pro modes. If you have never hit a Plus limit, Pro is not for you. Our catalogue notes that OpenAI has positioned Pro at different price points during rollouts, so confirm on the official pricing page. See also ChatGPT Pro vs Claude Max.
Business ($25/user monthly, $20/user annual, 2-user minimum) adds a shared workspace, admin controls, connectors and the promise that your data is not used for training. Enterprise is custom-priced with SSO, compliance and analytics. Both are covered on the official business pricing page.
Subscription or API?
The API bills per token with no monthly fee. At GPT-5.4 mini prices ($0.75 in, $4.50 out per 1M tokens at the time of writing), $20 buys millions of tokens, far more than a typical chat user generates in a month. But the API has no chat interface, no memory, no Projects, and you have to build or buy a front end. For most non-developers the subscription wins on convenience even when the API would win on price.
One more option: the API is also how ChatGPT-style features get into other products. If you are a developer, lesson 7 shows the quickstart. If you just want to compare a subscription against per-token spend for your own volume, the subscription vs API guide does the maths.
Knowledge check
You use ChatGPT roughly ten times a week for short emails and never see a limit message. What should you pay?