ChatGPT Pro Max Could Cost $500 a Month: What the Leak Actually Reveals

Key Takeaways

• OpenAI appears to be testing a higher ChatGPT subscription tier internally associated with promax and chatgptpromax. Third-party analysis of ChatGPT web assets has surfaced those identifiers, but OpenAI has not announced a public product called ChatGPT Pro Max.
• The widely circulated $500-per-month price remains unconfirmed. The strongest available leak evidence does not reveal a final price, launch date, usage multiplier, or complete feature list.
• OpenAI currently sells ChatGPT Pro in $100 Pro 5X and $200 Pro 20X tiers. Both provide the same core Pro capabilities, with the primary difference being usage allowance.
• New purchases and upgrades to Pro 20X have been temporarily paused since September 10, 2026, while existing $200 subscribers and the $100 Pro tier remain available.
• A genuine $500 plan would need to offer substantially more than extra chat messages. The clearest opportunity is heavier GPT-6 Astra, Codex, ChatGPT Work, agentic workflows, and long-running tasks, where compute consumption rises rapidly with context, reasoning, tools, and task duration.
• OpenAI DevDay 2026 takes place on September 29 in San Francisco, making the timing of the leak notable, although there is currently no official evidence that Pro Max will be announced there.

Is ChatGPT Pro Max Really Coming?

Evidence for a new high-end ChatGPT tier is stronger than a typical social-media rumor, but it is important to separate what has been observed from what has been inferred.

Reports examining ChatGPT-related JavaScript assets have identified internal-looking strings including promax, PROMAX, and chatgptpromax. These identifiers appear alongside other subscription-related values, suggesting that OpenAI has at least experimented with an additional plan or entitlement level.

That does not automatically mean a finished product called ChatGPT Pro Max is ready to launch.

Internal product identifiers can represent:

• unreleased experiments;
• abandoned pricing tests;
• temporary engineering names;
• A/B testing variants;
• entitlement groups rather than purchasable plans;
• products whose public names change before release.

The most defensible conclusion is therefore narrower: OpenAI appears to have developed or tested a higher Pro-related tier internally.

A separate report citing TestingCatalog connects the possible Pro Max tier with OpenAI's work on supporting long-running ChatGPT tasks, reinforcing the idea that this could be aimed at heavy agent users rather than ordinary chat users. However, that reporting does not establish a final price or feature set.

Is the $500 ChatGPT Pro Max Price Confirmed?

No.

As of September 25, 2026, OpenAI has not publicly announced a $500 ChatGPT subscription.

The company's official ChatGPT pricing page continues to list the established consumer plan structure, including Free, Go, Plus, and Pro. The public Pro page describes the plan as starting from a lower price rather than presenting a new $500 Pro Max option.

The code leak itself is also important because of what it does not contain. Available screenshots and analysis do not establish:

• a $500 checkout price;
• a 50X, 60X, or 100X allowance;
• a launch date;
• supported countries;
• an exclusive model;
• unlimited Astra usage;
• unlimited Codex agents;
• a final customer-facing name.

In other words, the $500 figure should currently be described as a rumored price, not an announced OpenAI price.

That distinction matters for SEO content because headlines stating that OpenAI has already launched or confirmed a $500 plan would currently overstate the available evidence.

What ChatGPT Pro Costs Today

OpenAI has already moved beyond a single flat Pro subscription.

Its official documentation currently describes two Pro tiers:

OpenAI says the existing two Pro tiers contain the same core advanced capabilities, including Pro models, Codex, deep research, image creation, memory, and file uploads. Their major difference is the included usage allowance.

This is important because it provides a plausible framework for Pro Max.

A simplified internal product ladder could eventually resemble:

Plus
  ↓
Pro 5X
  ↓
Pro 20X
  ↓
Pro Max

But there is currently no verified evidence establishing exactly how far above 20X a possible Pro Max tier would sit.

Why OpenAI Might Need a Plan Above $200

At first glance, a $500 consumer AI subscription sounds extreme.

The economics become easier to understand when ChatGPT is treated not as a chatbot but as an agent execution platform.

Modern ChatGPT workloads can involve significantly more compute than sending a prompt and receiving a short answer. A single complex task may require:

• extended reasoning;
• hundreds of thousands of tokens of context;
• repeated tool calls;
• browser interaction;
• code execution;
• repository exploration;
• file processing;
• web retrieval;
• multiple model calls;
• retries and verification;
• long-running cloud environments.

OpenAI's own documentation explains that ChatGPT Work and Codex share usage and that consumption varies based on factors such as model selection, context size, reasoning, tool use, retrieval, caching, and task complexity. Large projects and long-running tasks can therefore consume dramatically more allowance than short coding questions.

This creates a fundamentally different pricing problem.

A user asking 500 lightweight questions is not necessarily as expensive to serve as a developer repeatedly assigning autonomous agents to large repositories for hours at a time.

GPT-6 Astra Shows How Wide the Usage Gap Already Is

The current Codex and ChatGPT Work limits provide a useful example of how aggressively OpenAI is already segmenting access by compute consumption.

OpenAI currently publishes estimated local messages per five-hour period for several models. For GPT-6 Astra, the ranges are approximately:

These are not fixed message caps. OpenAI explicitly notes that actual consumption depends on workload complexity and that weekly limits can also apply.

The important signal is the size of the difference.

Pro is increasingly becoming a way to purchase more frontier-model compute, not merely a collection of premium UI features.

That makes another higher tier commercially plausible even if the rumored $500 figure ultimately proves inaccurate.

Why Pro 20X Is Currently Paused for New Subscribers

There is another major piece of context behind the Pro Max rumor.

OpenAI says that, beginning September 10, 2026, it temporarily paused new subscriptions and upgrades to the $200 Pro 20X tier. Existing Pro 20X subscribers remain unaffected, and users can still purchase the $100 Pro tier.

This does not prove that Pro Max is coming.

It does show that OpenAI is actively managing access to its highest-consumption individual tier rather than treating compute capacity as effectively unlimited.

That makes any future Pro Max product especially interesting.

A higher-priced tier could potentially function as a form of compute segmentation: users with workloads valuable enough to justify substantially higher monthly spending receive access to more scarce frontier-model capacity.

What Would a $500 Pro Max Need to Include?

If OpenAI simply offered the current $200 experience with a slightly larger message limit, a $500 price would be difficult to differentiate.

A serious Pro Max product would likely need to change the type of work users can perform, not merely the number of prompts they can send.

The following features would create meaningful separation. These are possible product directions, not confirmed Pro Max features.

1. Much Higher GPT-6 Astra Allowances

Frontier reasoning models are among the most compute-intensive parts of an AI subscription.

A premium tier could provide significantly larger Astra allowances, particularly for:

• difficult software engineering;
• scientific reasoning;
• large document analysis;
• mathematical work;
• autonomous research;
• multi-stage planning.

The important metric would not necessarily be raw message count. It could instead be a larger pool of compute credits or fewer restrictions on high-effort reasoning.

2. Longer ChatGPT Work Tasks

Long-running work may become one of the strongest reasons for a high-end subscription.

Instead of keeping a browser session open and manually prompting an assistant every few minutes, a user could potentially delegate a substantial job and allow ChatGPT to continue operating across many steps.

Examples include:

• researching a market and producing a finished report;
• auditing hundreds of pages on a website;
• processing a large collection of files;
• investigating a complex engineering problem;
• navigating websites and collecting structured results;
• performing multi-stage competitive research.

Reporting around the Pro Max leak has specifically connected the possible plan with longer-running tasks, although the exact implementation remains unknown.

3. More Concurrent Codex Agents

Coding agents become considerably more powerful when work can happen in parallel.

Instead of assigning one task at a time, a high-end developer might want separate agents simultaneously handling:

• a frontend implementation;
• backend API changes;
• tests;
• dependency upgrades;
• bug investigation;
• documentation;
• code review.

Concurrency consumes infrastructure faster than sequential chat usage, making it a logical premium feature.

4. Priority Compute During Peak Demand

At $500 per month, users would reasonably expect more predictable access to frontier models when demand is high.

Possible differentiation could include:

• shorter queues;
• higher task concurrency;
• faster agent execution;
• priority GPU capacity;
• fewer temporary capacity restrictions.

Again, none of these capabilities has been confirmed for Pro Max.

5. Larger Effective Context for Real Projects

High-end users increasingly work with complete repositories, research archives, spreadsheets, PDFs, and long-running project histories.

A premium plan could potentially differentiate through larger practical working context, higher file allowances, or more persistent project state rather than simply increasing conventional chat messages.

How $500 Would Compare With Other Premium AI Plans

A $500 ChatGPT subscription would sit well above most major consumer AI plans currently marketed to individuals.

Anthropic officially offers Claude Max at $100 for Max 5X and $200 for Max 20X.

Google announced in May 2026 that Google AI Ultra now starts at $100 per month, while its higher 20X Ultra tier costs $200 per month.

Perplexity Max is officially priced at $200 per month or $2,000 annually.

A genuine $500 ChatGPT Pro Max tier would therefore move OpenAI into a notably different pricing category for a self-service individual AI subscription.

The important competitive question would not be whether $500 is expensive. It clearly would be relative to today's major individual subscriptions.

The real question would be whether OpenAI could offer enough usable compute and autonomous work capacity to make the economics attractive to professionals whose time is worth substantially more than the subscription fee.

Who Could Actually Need ChatGPT Pro Max?

A $500 tier would probably make little sense for casual ChatGPT users.

The potential audience would instead include people who are already using AI as production infrastructure.

Professional Software Developers

Developers using Codex continuously across multiple repositories can reach usage limits far faster than users performing occasional coding assistance.

For these users, higher concurrency and longer cloud tasks could directly reduce development time.

AI-Native Founders

Small teams increasingly use AI agents for research, coding, analytics, customer operations, content production, and internal tooling.

For a founder effectively replacing several hours of repetitive work each week, subscription economics look different from those of a casual consumer.

Researchers and Analysts

Deep research workflows can involve large amounts of browsing, retrieval, synthesis, and reasoning.

A tier allowing longer uninterrupted jobs could be useful for competitive intelligence, financial research, technical literature reviews, and market analysis.

Heavy ChatGPT Work Users

As ChatGPT expands from answering questions into operating software and completing multi-step jobs, a new category of user emerges: people buying AI labor time rather than chatbot access.

That group is the strongest potential market for a Pro Max-style subscription.

Why Selling More Usage Is Becoming Harder

The AI subscription market originally copied familiar SaaS pricing: pay a monthly fee and receive access to a product.

Frontier AI breaks that model because marginal usage costs can vary enormously.

Two customers paying the same subscription price can generate very different infrastructure costs.

One may use a lightweight model for short writing tasks. Another may run frontier reasoning models against large repositories through autonomous coding agents all day.

The cost difference can be substantial.

This explains the industry's increasing use of:

• usage multipliers;
• model-specific limits;
• rolling windows;
• weekly allowances;
• usage credits;
• premium reasoning modes;
• paid extra usage.

Anthropic, for example, now offers Max subscribers additional prepaid usage bundles on top of their subscriptions, with credits usable across Claude, Claude Code, and Cowork.

OpenAI's own current structure similarly separates Plus, Pro 5X, and Pro 20X based heavily on usage rather than completely different product capabilities.

Pro Max would fit this broader industry transition from simple software subscriptions toward bundled compute subscriptions.

Could Pro Max Be Designed Around Credits Instead of Unlimited Usage?

This is another possibility worth watching.

The word Max does not necessarily imply unlimited usage.

In fact, truly unlimited frontier reasoning and agent execution would be difficult to price predictably because users can create extremely expensive workloads.

A future plan could instead combine:

Base subscription
+
large included compute allowance
+
priority access
+
optional paid credits

This model would give OpenAI predictable minimum revenue while still placing a ceiling on extremely compute-intensive usage.

OpenAI already describes ChatGPT Work and Codex consumption in terms of shared allowances and credits in applicable environments, making a compute-oriented structure technically consistent with the current product direction.

What the Frontend Leak Still Cannot Tell Us

Frontend code is valuable for spotting upcoming products, but it is also easy to overinterpret.

Finding promax in JavaScript does not answer several critical questions.

Is Pro Max the Final Name?

Not necessarily.

Internal plan identifiers frequently differ from public marketing names.

Is $500 the Final Price?

There is currently no reliable first-party evidence confirming that price.

Will It Be Available Globally?

Unknown.

OpenAI often rolls out products, models, and features gradually across regions and account groups.

Will Existing Pro Users Automatically Receive Any New Features?

Unknown.

A feature discovered alongside a plan identifier may still be tested across several subscription tiers before launch.

Could OpenAI Cancel the Plan?

Yes.

Companies routinely ship experimental code that never becomes a public product.

That is why the strongest confirmation would be one of the following:

• an OpenAI announcement;
• an official Help Center page;
• a live ChatGPT pricing page;
• a real checkout flow;
• a subscription SKU with a visible price;
• account-level rollout to paying users.

Until one of those appears, Pro Max should remain categorized as an unreleased and unconfirmed subscription tier.

Why the Timing Before OpenAI DevDay 2026 Matters

OpenAI has officially scheduled DevDay 2026 for September 29 in San Francisco.

The Pro Max references began receiving broader attention only days before the event.

That timing naturally creates speculation about a possible announcement.

However, timing alone is not evidence.

OpenAI's official DevDay page currently confirms the event but does not mention ChatGPT Pro Max or a $500 subscription.

The safest interpretation is therefore:

DevDay is a logical event to watch, but there is no confirmed Pro Max announcement scheduled for DevDay.

What to Watch Next

Several signals would dramatically increase confidence that Pro Max is approaching a real launch.

A Checkout Price Appears

The strongest evidence would be a live checkout flow exposing an actual monthly price.

A New Usage Multiplier Appears

Existing Pro plans use clear 5X and 20X positioning. A new public entitlement such as 40X or 50X would reveal much more about Pro Max economics.

OpenAI Updates Its Help Center

OpenAI's Help Center currently documents both existing Pro tiers in detail. A dedicated Pro Max article would effectively confirm the plan.

The Pricing Page Adds Another Tier

The official ChatGPT pricing page is the clearest customer-facing source. At present, it does not list a separate Pro Max subscription.

Work or Codex Gets New Long-Running Task Limits

Because long-running tasks appear central to reporting around the leak, changes to Work or Codex limits may provide clues even before a formal plan announcement.

Is a $500 AI Subscription Actually Plausible?

Yes, but only for a different category of AI usage than today's mainstream chatbot subscriptions.

The strongest argument for such a plan is not access to a slightly smarter model.

It is access to more machine work.

The value proposition becomes much easier to understand if a subscription can reliably complete hours of engineering, research, browser automation, analysis, or other professional work every month.

Under that model, the competition is no longer simply:

ChatGPT vs Claude vs Gemini

It increasingly becomes:

AI subscription cost
vs
human time saved
vs
API compute cost
vs
additional employee or contractor hours

That is a much larger economic market than traditional chatbot subscriptions.

If Pro Max is real, its most important feature may therefore be neither a new model nor a larger context window. It may be the amount of autonomous compute OpenAI is willing to bundle into a predictable monthly price.

Conclusion

The evidence currently supports a careful conclusion: OpenAI appears to be experimenting with a higher ChatGPT Pro tier associated with promax or chatgptpromax, but a $500 monthly price has not been officially confirmed.

What makes the leak important is not the name itself.

OpenAI already sells Pro 5X for $100 and Pro 20X for $200, and its latest products increasingly consume compute through coding agents, deep research, ChatGPT Work, tool execution, and long-running workflows.

A higher tier would therefore represent the next logical step in AI subscription economics: charging power users for substantially more agent execution and frontier-model compute.

The rumored $500 number should remain labeled as unconfirmed until it appears in an official OpenAI announcement, pricing page, Help Center document, or live checkout flow.

With OpenAI DevDay scheduled for September 29, the next few days are likely to provide a clearer picture of whether promax is merely an internal experiment or the beginning of a new ultra-premium category of ChatGPT subscription.