Written by Admin Alex · Fact-Checked by M.Ali · Info Verified September 2026
We review and update this article regularly as new information becomes available.
TL;DR: OpenAI has stopped accepting new sign-ups for ChatGPT Pro, its $200-a-month tier, because demand for its new Astra model is straining company servers. Existing Pro subscribers keep their access. Everyone else is locked out until further notice.

Imagine launching a product so popular you have to stop selling it. That’s roughly where OpenAI landed this week.
GPT-6 Astra went live on September 3. It’s the company’s most capable model yet, built for heavier reasoning, coding, and computer-use tasks than anything OpenAI has shipped before. The marketing leaned hard into the idea that this is the model that starts closing the gap toward AGI. Whether or not you buy that framing, plenty of people wanted in immediately.
By September 9, OpenAI was warning that its systems were under real strain. By the 10th, new sign-ups for the $200-a-month Pro plan were off the table entirely.
Thibault Sottiaux, one of OpenAI’s product leads, didn’t dress it up much. “Pro plan puts the most strain on its systems,” he said, adding that “demand for Astra is really unprecedented” and that he hadn’t “seen anything like it until now.” That’s a notable admission from a company that has weathered plenty of viral moments already.
Who’s affected, and who isn’t
If you already pay for Pro, nothing changes. You keep your access, your usage limits, everything. The freeze only applies to new customers trying to buy in at the top tier.
Plus, Go, the API, and the Enterprise and Business plans remain open. So the practical effect is narrower than the headlines suggest: this is a Pro-specific bottleneck, not a company-wide outage. But Pro is also the plan aimed at the heaviest individual users, the people running the most demanding Astra workloads, which is exactly why it’s the one buckling first.
There’s a pattern here if you look back a month. OpenAI quietly raised usage limits for Codex, its coding assistant, back in August. That’s usually a sign that internal capacity planning was already getting tight before Astra even shipped. Astra just pushed it over the edge.
No timeline, no numbers
OpenAI hasn’t said how long the pause will last, or published any figures on how many people tried to sign up, or how much compute headroom it actually needs before reopening the tier. That’s fairly typical of how the company handles capacity crunches. It says just enough to explain the decision, and nothing that would let outsiders benchmark how bad the strain really is.
For a company sitting on hundreds of billions in valuation and backing from some of the deepest-pocketed investors on the planet, running out of server room for a $200 subscription tier is an odd kind of problem to have. It’s also, in a strange way, the best kind of problem: too many customers, not too few.
Bottom Line
OpenAI built a model impressive enough that its own infrastructure couldn’t keep pace with demand for it, and rather than let service degrade for existing Pro users, it chose to just stop the line. That’s a reasonable call operationally. It’s also a reminder that even the best-funded AI lab in the world is still, at times, building the plane while flying it.


