ChatGPT Pro 500: What $500 a Month Buys, and Why It Costs the Same per Unit as Plus

OpenAI launched a $500-a-month ChatGPT plan on 29 September 2026. It's called Pro 500, it sits above the existing $100 and $200 Pro tiers, and it adds two things: the largest usage allowance OpenAI sells to one person, and access to Ultrafast, a faster version of its newest model.

That's the headline. Divide the price by the usage, though, and you get the more useful finding. Pro 500 charges exactly what Plus charges per unit of usage. The one plan on OpenAI's ladder that ever gave you a bulk discount was Pro 200, and on the same day OpenAI cut that plan's allowance for new subscribers.

Everything below comes from OpenAI's own pages, read on the day of the launch: the DevDay recap, the help article on Pro tiers, the ChatGPT release notes, the pricing page and the help article on usage in ChatGPT Work and Codex.

The Short Answer

Plan	Price	Usage, per OpenAI	Ultrafast	Cost per Plus-sized allowance
Plus	$20/mo	1x (the baseline)	No	$20
Pro 100	$100/mo	About 5x Plus	No	$20
Pro 200	$200/mo	About 20x Plus for existing subscribers. New subscribers get less	No	$10, rising for new subscribers
Pro 500	$500/mo	25x Plus	Yes	$20

The last column is my arithmetic. The multipliers are OpenAI's. Pro 500's 25x comes from the DevDay recap, which says the plan "offers our highest usage allowance at 25 times the ChatGPT Plus allowance". The 5x and 20x come from OpenAI's usage tables, which still label the two older tiers "Pro 5x" and "Pro 20x".

So Pro 500 is for exactly one kind of person: someone whose work already stalls at Pro 200's limits, or who loses real money waiting for long agent runs to finish. Everyone else is better off on a cheaper rung.

What $500 Buys

25 Times the Plus Allowance

This is the main thing you're paying for. OpenAI's help article on Pro tiers says only that "Pro 200 includes more usage than Pro 100" and that "Pro 500 offers the highest included usage of the three plans". The DevDay recap is the one place that puts a number on it, at 25 times Plus.

Two limits on what that number means. It isn't a message count. OpenAI's own usage guide for ChatGPT Work and Codex says the allowance "does not correspond to a fixed number of messages or tasks", and what it covers depends on the model, the task and the settings. And it isn't unlimited. The help centre says "a Pro subscription does not include unlimited use of the GPT-6 Pro model", and every model keeps its own allowance. Everyday text chat is listed as unlimited on every ChatGPT plan, Free included, so that part of the product doesn't change as you climb the ladder.

Ultrafast

Ultrafast is OpenAI's new "premium speed tier". The recap describes it as "up to 8× faster token generation (300 tokens per second) in Codex". On ChatGPT plans it runs OpenAI's newest model, Astra, inside ChatGPT Work and Codex, and you choose it from the model picker.

Three details matter before you pay for it:

• It's only on Pro 500 among personal plans. OpenAI's help article is explicit that "buying credits on Pro 100 or Pro 200 does not unlock Ultrafast".
• It spends your allowance, then your credits. Ultrafast "uses your plan's included usage first, then draws from your credit balance after that allowance is used". Faster work runs through the allowance faster too.
• It lives in Work and Codex, not in ordinary chat. If most of what you do in ChatGPT is conversation, Ultrafast doesn't touch it.

Everything Else Is Shared

The rest of the Pro card applies to all three tiers: OpenAI's Pro reasoning model, maximum deep research, the largest memory and context, and dots, OpenAI's new always-on agents, which were rolling out to Pro users outside the EEA, Switzerland and the UK at launch. OpenAI's pricing grid lists one set of context figures for Pro as a whole, so nothing there suggests Pro 500 reads more of a document at once than Pro 100. It gets more usage and more speed.

The Per-Unit Math: Pro 500 Charges Plus's Rate

Here's the ladder divided out, using OpenAI's own multipliers:

• Plus: $20 for 1x. $20 per unit.
• Pro 100: $100 for 5x. $20 per unit.
• Pro 200: $200 for 20x. $10 per unit.
• Pro 500: $500 for 25x. $20 per unit.

Three of the four rungs charge the same rate. Only Pro 200 was ever cheaper per unit, at half price. Anthropic's ladder has the same shape, and I found the same thing in Claude Pro vs Max: the middle rung charges the base plan's rate, and only the $200 rung discounts anything.

Now look at the step from Pro 200 to Pro 500 on its own. You pay $300 more for 5x more, which is $60 per extra Plus-sized allowance, three times what Plus charges. Part of that premium is Ultrafast. The rest is the price of being at the very top.

The picture gets worse for anyone signing up for Pro 200 now. New subscribers get "a lower usage allowance than previously offered with Pro 200", which OpenAI says reflects "our increasingly efficient models". It hasn't published the new figure. Whatever it is, the half-price rung just got less of a bargain, on the same day a $500 rung appeared above it.

What Changed for Pro 200

The $200 tier had a strange September. OpenAI paused new sign-ups on 10 September, and on 29 September it reopened them at the same $200 price, with the smaller allowance above. If you already had Pro 200:

• If your subscription was active at the eligibility cutoff, or during the seven days before it, you keep the previous allowance through 29 October 2026, as long as the subscription stays active.
• After that date you move to the lower allowance, and the price stays $200.
• Keeping the old allowance "does not upgrade your plan or add Ultrafast".
• If your subscription lapsed in the seven days before the cutoff, subscribing again before 29 October gets you the previous allowance back for that period.

OpenAI says only affected users will get an email. If you're on Pro 200 and wondering whether to move up, the decision point is 29 October, not today.

"No Usage Caps", According to One Page

OpenAI's Pro plan page says Pro "removes limits" and offers "no usage caps". Further down the same page, its own FAQ answers the question "What happens when I reach my GPT-6 Astra usage limit?" The help centre is plainer: every model on Pro has its own allowance, and at a model's limit "it may be unavailable until its allowance resets".

That gap matters most at $500. If you're paying for headroom, find the limit you actually hit before you pay, because it's a per-model limit. On Pro 200, for example, reaching the weekly limit on OpenAI's Pro model switches you automatically to a lighter thinking setting. On Pro 100, both of OpenAI's Pro models share one weekly allowance. Neither of those is fixed by buying a bigger multiple of the same thing unless that's the specific wall you keep hitting.

Who Should Pay $500

• Pay it if you run long jobs in ChatGPT Work or Codex every day, you already hit Pro 200's limits, and the waiting costs you more than $300 a month. Ultrafast is the only thing on the ladder that buys speed rather than volume.
• Stay on Pro 200 if you're grandfathered on the old allowance and rarely hit it. You have until 29 October to see whether the new allowance changes that.
• Stay on Pro 100 or Plus if you mostly chat. Everyday chat is unlimited on every tier, and the $500 plan's extras sit in Work and Codex.
• Consider a team seat instead only if other people need to share your work. For one person, a Premium Business seat still needs a second paid seat, and it changes who owns your chat history.

The full ladder, from Free and the $6 Go plan up, is in ChatGPT Go vs Plus vs Pro, and every tier's price, including Business, Enterprise and the API, is in how much ChatGPT costs.

Before You Buy More of One Vendor

Most people who look at a $500 plan got there by running into limits on a $20 one. Before paying for 25 times the allowance, it's worth asking which work actually needs OpenAI's top model, and which of it just needs a good model. ChatGPT's limits sit on specific features such as uploads, images, research and the Pro models, and a lot of everyday drafting and checking runs just as well on another lab's model.

That's what izzedo chat is for. It puts OpenAI's GPT models in one thread with Claude, Gemini, Grok, DeepSeek and more, so the jobs that don't need OpenAI's top tier go somewhere else, and a second model can check the ones that do. Paid plans start at $6 a month, and the free plan needs no card. It won't give you Ultrafast. For most people, though, the problem isn't speed. It's that one vendor's limits are standing between them and the work.

The Bottom Line

ChatGPT Pro 500 costs $500 a month for 25 times Plus's allowance and Ultrafast, OpenAI's fastest way to run its newest model in ChatGPT Work and Codex. Per unit of usage it charges exactly what Plus does, and the step up from Pro 200 costs three times Plus's rate for each extra allowance. It's worth that if speed on long agent jobs is what you're short of. If you're short of room on everyday work, you're paying the top rung's premium to solve a problem that a cheaper plan, or a second model, would fix.

---

Hitting ChatGPT's limits on work that doesn't need OpenAI's top model? Try izzedo chat free. No credit card required.