OpenAI reopens $200 Pro plan with usage math worth half the old API-equivalent value

OpenAI is reopening its $200-per-month Pro subscription to new subscribers and, at the same time, is changing how usage is calculated so that the plan works out to roughly half the API-equivalent dollar value of the previous version. In a post on X by Tibo (@thsottiaux), which surfaced on Hacker News's front page on September 29, 2026, the company said the subscription reopens the following day.

According to the post, anyone who converts the new plan's allowances into pay-as-you-go API terms will find it delivers about fifty cents on the dollar compared with the old Pro plan. OpenAI frames this as a deliberate rebalancing rather than a downgrade, arguing that the work subscribers can actually get done will keep improving even as the headline value figure falls.

The reasoning OpenAI gave

The post laid out four points behind the change:

• The five-hour usage cap is not coming back. OpenAI says it would rather let subscribers spend the full weekly allowance whenever they want than reintroduce session limits as a cost control.
• The company says subscribers are guaranteed to get more done over time, at rising quality, because model efficiency gains are passed through as API price cuts. In its framing, value per dollar keeps climbing even as the nominal API-equivalent figure shrinks.
• OpenAI says it does not want a structural incentive to keep API list prices inflated so that a subscription looks generous by comparison, with discounts quietly compensating. It points to this week's introduction of GPT-6 Sol and GPT-6 Luna at half their earlier price as the pattern it intends to follow. Its stated long-term view is that API prices fall far enough that most people will simply buy usage as they need it, with little gap between what a subscription and the API deliver per dollar.
• Additional features are being added to the subscription that will not draw on the usage allowance. The post declined to say what those are.

The author described the advance notice as an attempt at transparency ahead of what the post called large announcements scheduled for the following day.

What is still unclear

The post does not spell out how existing subscribers are treated. Its wording describes changing the usage calculation for the plan as a whole, without distinguishing between current customers and new sign-ups, and without quantifying the weekly allowance or listing which models and tools draw on it. Those details may arrive with the announcements OpenAI has teased.

Why it matters

For heavy users this is a terms change in everything but price: the subscription still costs $200 per month, but measured against OpenAI's own API rates it now represents about half the value of the old plan. Whether that amounts to a loss depends on whether the company's throughput-and-quality argument holds. If models keep getting cheaper and better on the API, the same subscription dollar buys more real work even as its nominal value shrinks.

The announcement is also notable for what it rules out. When a plan's worth is measured against a vendor's own list prices, the vendor can make the comparison look better by raising list rates and discounting elsewhere. OpenAI explicitly rejects that playbook and is instead betting that falling API prices will narrow the gap between subscriptions and pay-as-you-go until the distinction mostly stops mattering. If that prediction holds, subscriptions become a convenience wrapper around cheap metered usage rather than the best deal per token.

The timing is the final piece. A de facto value cut announced one day ahead of a batch of product news is a familiar sequencing, with the uncomfortable change landing before the announcements meant to offset it. Subscribers deciding whether to stay will have to weigh the halved accounting against whatever OpenAI adds to the plan next.