ChatGPT's price tiers themselves have not changed — the free tier, Plus at $20 per month, and Pro at $200 per month are still the three paid and unpaid options listed in our AI tool database — but what each tier unlocks has shifted as OpenAI moved its newest flagship models behind the paid plans.
According to our AI tool database, the free tier now runs on GPT-4o mini, while ChatGPT Plus at $20 per month is the level where you get access to OpenAI's flagship assistant, described as GPT-5.5 with a 1M context window, native image generation, and the Codex coding agent.
So if you clicked this because your bill looks the same but the product feels different, that is the real story: the pricing held steady, and the value inside each tier moved.
To understand why, you have to separate two things that get blurred together in most discussions: the subscription price and the model access. OpenAI sells one product with several doors, and the model you reach depends on which door you walk through. The free door leads to a smaller, faster model that handles everyday questions well but struggles with long documents or complex coding.
The $20 door leads to the flagship. The $200 Pro door leads to the same flagship with far higher usage ceilings for people who hit limits constantly. This is a common pattern across the category — our database shows Claude priced at $0 free, $17 per month on annual billing or $20 monthly for Pro, and from $100 per month for Max, and Google Gemini's Advanced tier at $19.99 per month.
The tiers are not really about features so much as about how much of the good model you get to use.
A concrete example makes this clearer. Say you write product descriptions and paste in a 40-page supplier PDF to pull specs from. On the free tier, running GPT-4o mini, you would likely hit trouble with that much text at once and get summaries that miss details buried deep in the document.
On Plus at $20 per month, the 1M context window means that same PDF fits comfortably, and the flagship model is better at holding the whole thing in view while it answers. Nothing about your subscription price changed to make that possible — the model behind the $20 tier did. That is the distinction worth holding onto when you read headlines about ChatGPT pricing: check whether the article is talking about the sticker price or about which model sits behind it.
There are real limits to this way of thinking, though. Our database snapshot is dated, and AI pricing and model lineups move faster than almost any other software category — the ChatGPT figures above come from a snapshot recorded at verification time, and the specific model names attached to each tier can change without the price changing at all.
Treat any published price as a starting point, not a contract, and confirm current details on OpenAI's own pricing page before you commit. Also worth knowing: a $20 Plus subscription is per person, so a small team of five is looking at five separate seats, which is a very different budget conversation than one person deciding whether to upgrade.
And if your work is mostly short questions and quick drafts, the free tier may genuinely be enough — paying for Plus mainly buys you the flagship model, longer context, and image generation, not a fundamentally different kind of assistant. For a broader look at where free tiers stop being useful, see our guide to the best free AI writing helper options and where they hit their limits.
The honest summary: ChatGPT's pricing did not change, but the gap between free and paid widened, and that gap is what you are actually paying to close.