ChatGPT Pro 500 is OpenAI’s new top consumer plan. It costs $500 per month, gives you 25 times the ChatGPT Plus allowance, and is the only plan that includes Ultrafast, a speed tier that runs up to 8x faster in Codex. OpenAI announced it at DevDay on September 29, 2026, and it was available immediately.
The plan arrived alongside a less welcome change. From October 30, 2026, the $200 Pro plan gets a smaller allowance, which widens the gap between the two top tiers.
This guide covers the price, what the $500 buys, how it compares with Plus and Pro 200, and who should skip it. It is part of our OpenAI DevDay 2026 announcements coverage, and all figures are as of September 30, 2026.
How much does ChatGPT Pro 500 cost per month?
ChatGPT Pro 500 costs $500 per month, which is $6,000 over a year. That is 25 times the price of ChatGPT Plus at $20 and 2.5 times the price of Pro 200. OpenAI sells it as a consumer plan, so it is billed per person, not per workspace.
Plus is $20 per month, according to OpenAI’s ChatGPT Plus help article. OpenAI’s ChatGPT Pro tiers help page now lists three Pro plans: Pro 100 at $100, Pro 200 at $200 and Pro 500 at $500. For a wider look at every tier, see our ChatGPT pricing guide.
Regional prices and local taxes can differ from the US figure. Check the price shown at checkout before you commit.
What do you get with ChatGPT Pro 500?
ChatGPT Pro 500 gives you the highest usage allowance OpenAI sells to individuals, at 25 times the Plus allowance, plus exclusive access to Ultrafast. It also includes the Agents API with computer use in Codex and ChatGPT Work, which Plus and free users do not get.
OpenAI’s DevDay 2026 recap describes the plan as “our highest usage allowance at 25 times the ChatGPT Plus allowance” and says it is available now. Like the other Pro plans, it includes Pro models, Codex, deep research, image creation, memory and file uploads. Reports also say each Pro plan comes with one personal AI agent, called a dot, and that conversations with your dot do not count against ChatGPT usage limits.
Two things are missing from the announcement. OpenAI has not published a token count for the 25x figure, and it has not said how many weekly GPT-6 Pro chat messages the $500 plan includes. Treat “25x” as a ratio against a Plus allowance that is itself not published as a fixed number.
What is Ultrafast and how much does it cost?
Ultrafast is a premium speed tier that generates tokens up to 8x faster in Codex, about 300 tokens per second, and up to 6x faster in the API. It costs up to 6x the normal rate. On the API, GPT-6 Astra Ultrafast is $60 input and $300 output per million tokens.
The Astra Ultrafast rate is exactly six times the standard Astra list price, which is why the “up to 6x” claim is easy to check. GPT-6.1 Sol Ultrafast is listed as coming soon. On the subscription side, Ultrafast is limited to Pro 500 and Enterprise.
| Model tier (API) | Input per 1M tokens | Output per 1M tokens |
|---|---|---|
| GPT-6 Astra standard | $10 | $50 |
| GPT-6 Astra Ultrafast | $60 | $300 |
| GPT-6.1 Sol standard | $2 | $10 |
Speed is not the same as finishing sooner. Ultrafast raises the rate at which the model writes tokens, but a task that waits on tests, web pages or human approval still waits. Some third-party guides also say Ultrafast draws down your included usage faster than standard mode. I could not confirm an exact rate in OpenAI’s own pages, so treat that as unverified. For the new cheaper model, see our post on GPT-6.1 Sol pricing and benchmarks.
What does the Agents API with computer use add?
The Agents API now supports computer use, meaning an agent can click, type and navigate software the way a person would. It also brings Codex’s multi-agent tools, tool search, tool calling and context compaction into your own product. OpenAI runs the infrastructure, so you do not host anything yourself.
OpenAI lists availability through the API and through Codex and ChatGPT Work on Pro 500 and Enterprise. Our explainer on the OpenAI Agents API with computer use covers the developer side, and the earlier OpenAI Astra computer use post explains how the underlying capability works.
Codex itself is not exclusive to Pro. Codex cloud, which runs tasks remotely from any device, is included in Plus, Pro, Business and Enterprise plans, as our Codex cloud guide explains.
Can you run OpenAI agents inside AWS Bedrock?
Yes, in principle. OpenAI and Amazon announced Bedrock Managed Agents, which lets developers run OpenAI agents entirely inside AWS. That suits companies whose data and procurement already sit in Amazon’s ecosystem. Sources disagree on which ChatGPT plans this is tied to, so confirm before you buy.
OpenAI’s recap describes the AWS integration without a plan restriction. Decrypt’s DevDay roundup says it is limited to Pro 500 and Enterprise users. If Bedrock is the reason you are considering Pro 500, ask OpenAI or AWS for written confirmation first.
Test Ultrafast Before Committing
Ultrafast helps most when a human is waiting on a long agent run, such as a live coding session. If your work is queued overnight or waits on approvals, the extra speed buys you very little. Run a real workday on Pro 200 first and note where you actually sat idle.
What changed for ChatGPT Pro 200 on October 30?
From October 30, 2026, Pro 200’s Work and Codex allowance falls from 20x to 10x the Plus allowance, and weekly GPT-6 Pro chat messages fall from 200 to 100. The price stays at $200. Existing subscribers keep their current limits until October 29 and receive a one-time $2,500 usage credit.
Those details come from The Next Web’s report on the Pro 200 cut and Pro 500 launch. The credit expires at the end of the year. OpenAI also said it will not bring back the five-hour limit on Pro 200.
New Pro 200 sign-ups were paused in September and reopened on stage at DevDay. OpenAI has not said how the $2,500 credit converts into tasks, so you cannot yet judge how far it stretches. Some users on community forums called it a retroactive cut, and OpenAI’s stated reason was that its models are becoming more efficient.
Third-party guides also report Pro 100 at 5x Plus. OpenAI’s help page for the tiers gives no multipliers for Pro 100 or Pro 200, so treat 5x as reported, not confirmed.
ChatGPT Pro 500 vs Plus vs Pro 200: how do they compare?
Pro 500 is $500 with 25x the Plus allowance and Ultrafast. Pro 200 is $200 with 10x from October 30. Plus is $20 with the baseline allowance. Only Pro 500 has Ultrafast and computer use in the Agents API. The cells below say “not disclosed” wherever OpenAI has not published a figure.
| Feature (as of September 30, 2026) | Plus | Pro 200 | Pro 500 |
|---|---|---|---|
| Monthly price | $20 | $200 | $500 |
| Work and Codex allowance vs Plus | 1x (baseline) | 10x from Oct 30, 2026 (20x before) | 25x |
| Weekly GPT-6 Pro chat messages | Not disclosed | 100 from Oct 30 (200 before) | Not disclosed |
| Ultrafast | No | No | Yes |
| Agents API with computer use in Work and Codex | No | Not listed by OpenAI | Yes |
| Personal dot agent | Not listed | One dot, reported | One dot, reported |
| Codex cloud | Included | Included | Included |
| Fixed token allowance | Not disclosed | Not disclosed | Not disclosed |
The comparison uses the OpenAI DevDay recap, the tiers help page and the Next Web report. Our ChatGPT pricing 2026 guide tracks Plus, Business and Enterprise in more detail.
Who should buy ChatGPT Pro 500?
Buy ChatGPT Pro 500 if you hit Pro 200 limits several times a week, run agents for hours at a time, or lose real money waiting on slow generation. It suits heavy Codex users, solo developers and consultants whose billable time outweighs $300 of extra monthly cost.
A few specific profiles fit well:
- Developers who run several Codex tasks in parallel and are blocked by usage caps rather than by ideas.
- Freelancers who bill by the hour and can pass on the cost, where faster output shortens delivery.
- Power users who need computer use in Work and Codex, and do not want to build against the API.
- Anyone whose Pro 200 allowance drops on October 30 and who was already near the ceiling under the old 20x.
The test is simple. If you can name the hours the plan will save, and those hours are worth more than $300 a month, the upgrade pays for itself. If you cannot name them, it does not.
Who should skip ChatGPT Pro 500?
Skip ChatGPT Pro 500 if you use ChatGPT for writing, research or occasional coding. Plus at $20 or a Pro plan below $500 covers most people. You should also skip it if you cannot say what Ultrafast would speed up or if your team needs shared, governed access.
Three cases stand out. First, casual users will never touch 25x, so they pay for capacity they cannot use. Second, anyone whose limits are met at Pro 200 gains nothing from more headroom. Third, teams: OpenAI’s Pro tiers page prohibits sharing account credentials, so five people cannot split one $500 login.
There is also a transparency cost. OpenAI states the multipliers as ratios and publishes no fixed token or message numbers for most of them. You are buying capacity you cannot precisely measure, and the vendor can change the ratio, as the Pro 200 cut shows.
Is ChatGPT Pro 500 cheaper than paying for the API?
It depends on volume, and OpenAI does not meter subscription usage in tokens, so no one can give an exact break-even. As an illustration only: at Astra’s standard API prices, $500 covers roughly 67 heavy agent tasks. At Ultrafast prices it covers about 11. Sol covers over 300.
The table below is illustrative, not a forecast. It assumes one large agent task uses 500,000 uncached input tokens and 50,000 output tokens. Real tasks vary widely, and cached input is much cheaper. The prices are OpenAI’s published API list prices per million tokens.
| Model (illustrative task) | Cost per task | Tasks per $500 |
|---|---|---|
| GPT-6 Astra standard ($10 / $50) | $7.50 | About 67 |
| GPT-6 Astra Ultrafast ($60 / $300) | $45.00 | About 11 |
| GPT-6.1 Sol standard ($2 / $10) | $1.50 | About 333 |
Read this two ways. If you would run many hundreds of large tasks a month, a flat $500 can beat pay-per-token pricing, especially on Astra. If you run a few dozen, API billing on Sol is far cheaper and has no monthly floor. The subscription’s advantage is predictability, not a lower rate.
Ultrafast changes the maths. At six times the price, the API route becomes expensive fast, which is the main financial argument for having it bundled into a flat fee. For a deeper model comparison, read our GPT-6 Sol vs Luna vs Astra breakdown.
Price The Cheaper Model First
Before paying $500 for Astra capacity, rerun your heaviest job on GPT-6.1 Sol. OpenAI says it nearly matches Astra on agentic coding and computer use at one-fifth of the price. If Sol clears your quality bar, you may not need the top tier at all.
Should L&D teams budget for ChatGPT Pro 500?
Most L&D teams should not buy Pro 500 seats for everyone. Pilot one plan with one instructional designer, then move to Business or Enterprise for shared, governed access. Pro 500 is a personal plan, and 10 seats would cost $5,000 a month, or $60,000 a year.
The Pro 500 features are aimed at developers and agent builders. Course authoring, SME interview synthesis, learning objectives, storyboards, assessment drafts and Kirkpatrick Level 1 to 3 reporting rarely need 300 tokens per second. Most of that work fits comfortably in standard plans. Our guide to GPT-6 Astra for learning and development shows where the stronger models genuinely help.
There is a governance point too. OpenAI’s Business FAQ lists Standard seats at $25 per user per month, or $20 billed annually, and Premium seats at $125 monthly, or $100 annually. It also says OpenAI does not train on your workspace data. For learner records, compliance training content or SCORM and xAPI exports, a workspace with admin controls is safer than personal accounts. Check the data settings on any personal plan before you upload anything sensitive.
A practical budget approach:
- Give one power user Pro 500 for a 30-day pilot and track hours saved.
- Put the wider team on Business, and compare Premium seats against Pro 200 for heavy users.
- Ask your account manager about Enterprise if you need Ultrafast or the Agents API across the team.
If you also build training inside an LMS or LXP, remember that most of your cost will still be authoring time and review, not model tokens.
Is ChatGPT Pro 500 the same as Pro 200?
No. ChatGPT Pro 500 costs $500 per month and Pro 200 costs $200 per month. They are separate plans with different allowances, and only Pro 500 includes Ultrafast. The two are easy to mix up because Pro 200 was reopened and changed on the same day as the Pro 500 launch.
OpenAI’s DevDay recap names the plan Pro 500 and the Pro tiers page lists it at $500. If a summary attaches a $200 price to Ultrafast or the 25x allowance, check it against OpenAI’s page before you act on it. Our OpenAI DevDay 2026 recap keeps the corrected numbers in one place.
Conclusion
Your next step is to find out whether you actually hit limits. Check your current usage on Plus or Pro 200 for two weeks and write down each time you were blocked or waiting. If the answer is never, stay where you are.
If you are blocked several times a week, or Ultrafast would save billable hours, run a 30-day Pro 500 trial and measure it against the $500 fee. If you manage a team, start with one seat and a written test, then decide on Business or Enterprise.
Prices and allowances can change, as the Pro 200 cut shows, so recheck the current ChatGPT pricing before you commit.
FAQ
Q1. How much is ChatGPT Pro 500?
ChatGPT Pro 500 costs $500 per month, or $6,000 a year before any local taxes. OpenAI announced it at DevDay on September 29, 2026. It offers 25 times the ChatGPT Plus allowance and is the only plan with Ultrafast. Plus stays at $20 and Pro 200 stays at $200 per month.
Q2. What is the difference between ChatGPT Pro 200 and Pro 500?
Pro 500 has the highest usage allowance at 25 times Plus, plus Ultrafast and the Agents API with computer use in Work and Codex. From October 30, 2026, Pro 200 drops from 20x to 10x Plus and from 200 to 100 weekly GPT-6 Pro messages. The price gap is $300 a month.
Q3. What is ChatGPT Ultrafast?
Ultrafast is a premium speed tier for GPT-6 Astra. It generates tokens up to 8x faster in Codex, about 300 tokens per second, and up to 6x faster in the API. It costs up to 6x standard rates, which is $60 input and $300 output per million tokens on the API. Subscribers need Pro 500 or Enterprise.
Q4. Is ChatGPT Pro 500 worth it?
It is worth it only if you regularly hit Pro 200 limits or lose billable time waiting on slow agent runs. Casual users, writers and occasional coders will not use 25x Plus capacity. Test your real usage on a cheaper plan first, then run a 30-day Pro 500 trial and measure the hours saved.
Q5. What happens to existing ChatGPT Pro 200 subscribers?
Existing Pro 200 subscribers keep their current limits until October 29, 2026. From October 30, the Work and Codex allowance falls from 20x to 10x Plus and weekly GPT-6 Pro chat messages fall from 200 to 100. They also receive a one-time $2,500 usage credit that expires at the end of the year.
Q6. Can a team share one ChatGPT Pro 500 account?
No. OpenAI’s Pro tiers help page prohibits sharing account credentials and reselling access. Pro 500 is a personal plan billed per person. Teams that want shared, governed access should look at ChatGPT Business or Enterprise, where admins manage seats and OpenAI says it does not train on workspace data.
Q7. Does ChatGPT Pro 500 include API credits?
OpenAI has not said that it does. The subscription covers usage in ChatGPT, Work and Codex, and Ultrafast in those apps. The API is billed separately at per-token prices, such as $10 input and $50 output per million tokens for GPT-6 Astra. Check your account billing page for any credits.
