Twenty dollars a month sounds like a simple number until you try to work out what it buys. ChatGPT’s plan lineup has multiplied since 2023, the model names keep shifting, and the details that actually matter (message caps, file limits, whether your chats train the model) live in help-centre pages rather than the pricing table. This ChatGPT pricing review walks through each tier, says where the money is well spent, and flags the costs that don’t appear on the checkout page.
What ChatGPT costs right now
These are the US dollar prices listed on OpenAI’s site, before tax. Currency and local pricing vary, sometimes a lot.
- Free: $0, with a daily allowance on the top model
- Go: a cheaper regional tier, roughly the equivalent of $8 a month, sold in selected countries
- Plus: $20 a month, or $240 if you pay annually
- Pro: $200 a month
- Team: $25 per user a month billed annually, $30 month-to-month, two-seat minimum
- Enterprise: negotiated, almost always a twelve-month contract
Worth noticing: the $20 Plus price has not moved since launch, even though the underlying model has been replaced three or four times. That is genuinely unusual in software, and it tells you where OpenAI thinks the growth is coming from. It isn’t the individual subscriber.
The free tier is not a demo anymore
Free ChatGPT runs the flagship model with a cap. In my own use I get somewhere between eight and twelve substantial prompts before it hands me off to a lighter model, and the counter resets the next day. You get file uploads, image generation, and a few web searches. It is enough to answer the question of whether you need ChatGPT at all.
The honest test is a two-week trial on free, not a five-minute poke. If you finish those two weeks without hitting the ceiling once, do not upgrade. If there are reviews of free tools you’d rather read first, there’s a solid roundup of the best free AI tools judged by what reviewers actually write that covers how the alternatives compare at zero cost.
ChatGPT Plus at $20: the plan most people should weigh
Plus lifts the cap, unlocks Projects, custom GPTs, Tasks, Deep Research in limited amounts, and a better voice mode. The jump in usable hours is real. The ceiling still exists, though. On a heavy writing day I have hit the top-model limit on Plus by mid-afternoon and spent the evening on the fallback.
Where Plus earns its keep
Long documents are the strongest case. I routinely feed in 60-page PDFs and ask for a structured comparison, and Plus handles that without the truncation you get on free. Projects are quietly the most useful feature for anyone with recurring work: you set the context once and stop re-explaining your job every morning.
Where Plus falls short
There is no admin console, no shared workspace, and no way to centrally manage who on your team is using what. Chat history is on by default and used for training unless you turn it off in settings, which plenty of people never do. For solo work, fine. For a business handling client data, that is a compliance conversation you need to have before you swipe the card.
Pro at $200: a specialist’s tool
$200 a month is ten times Plus, and the pitch is extended reasoning, near-unlimited access to the heaviest model, longer Deep Research runs, and better video generation. If you are a consultant billing $150 an hour and the reasoning mode saves you three hours a month, the maths is easy. If you are a student, it is not.
My read is that Pro is priced to be a business expense rather than a personal one. It also has a real switching cost buried in it: once you get used to extended reasoning on hard problems, dropping back to Plus for a month feels like losing a colleague.
Team and Enterprise: where the real bill lands
Team looks like a bargain at $25 per seat annually, and it is, if every seat belongs to someone who uses it daily. The problem is the two-seat minimum plus an annual commitment, which makes it easy to buy five licences for a six-person team and quietly pay for three people who log in once a fortnight. At $25 a seat that’s $900 a year in licences, and nobody notices because it’s one line on a card statement.
What you get for it is the stuff Plus lacks: a shared workspace, an admin console, and the guarantee that your data is excluded from training. Enterprise adds SSO, compliance tooling and a support contract, at a price you have to ask for. Comparing seat-based plans properly is its own exercise, and this value-for-money review of AI tools for business is a useful second opinion before you sign a twelve-month deal.
The costs that never make the pricing page
Three things reliably inflate what you actually pay.
Tax and currency. UK and EU buyers add roughly 20% VAT on top of the dollar figure. Buyers in India or Brazil may see rupee or real pricing that looks nothing like the US number. If you are budgeting, add the tax before you compare tiers.
Annual versus monthly. Team is $25 or $30 depending on commitment. That’s $60 per seat per year for the convenience of paying monthly, which is fine for a trial and wasteful for a permanent team.
The second subscription. Almost nobody runs one AI tool. A typical freelancer I know pays for ChatGPT Plus, Claude Pro and a transcription service, which is $50 a month before any of the niche tools. The pattern of stacked subscriptions and renewal creep is covered well in this breakdown of what you actually pay after the first AI invoice, and it’s the number worth tracking rather than any single line item.
How it stacks up against the alternatives
Claude Pro and Gemini’s premium tier both sit at the same $20 mark. Perplexity Pro is also $20. So the price is the market rate, not a differentiator, and the decision comes down to which model handles your particular work. In blind comparisons on long-document editing, I find Claude steadier; on anything involving images or voice, ChatGPT is ahead.
Purpose-built writing tools cost more and do less, generally. Jasper starts around $39 a month for a single seat, and if you want a detailed breakdown of whether that premium is justified, there’s an honest Jasper AI review covering its features, cost and who should skip it. My short version: pay for a general assistant before you pay for a specialist one.
How I’d decide, in practice
Buy one month of Plus, in cash, on a card you can cancel. Keep a rough note of every day you open it and what you used it for. At the end of the month, three outcomes are possible, and only one of them is an upgrade.
If you used it fewer than ten times, cancel and stay on free. If you used it most days and hit the model cap more than twice, Plus is earning its $20 and you should keep it. If you hit the cap and the work is billable, try Pro for a single month and see whether the extended reasoning actually changes your output or just makes you feel better equipped.
For teams, the calculation is different and more brutal. Count the people who used it last week, not the people who asked for a licence. Seat-based pricing punishes optimism, and the annual discount is only a discount if the seats stay filled.
What I pay for, and what I dropped
I keep Plus and I have never once been tempted by Pro, because my hard problems are editing problems rather than reasoning problems. I dropped a $30-a-month writing assistant after two months when I realised I was pasting its output into ChatGPT for a second pass anyway. The one upgrade I would make tomorrow is Team, the moment a second person needs access to the same project files, purely for the data-training exclusion and the shared workspace.
That’s the shape of it. The headline number is $20, and for a lot of people that’s the right amount to spend. The trap isn’t the price of Plus. It’s the four other subscriptions that quietly attach themselves to it, and the team licences nobody audits. If you take one thing from this ChatGPT pricing review, make it this: check what you spent on AI last quarter, not what the plan page says it costs.