OpenAI is facing an unprecedented surge in demand for its latest model, GPT-6 Astra, prompting the company to potentially halt new Pro subscriptions to prioritize existing users.
Massive Demand for GPT-6 Astra
Tibo Sottiaux, Head of OpenAI Codex, took to X (formerly Twitter) to address the overwhelming user interest in Astra. He stated that the demand is “unprecedented” and has necessitated the reallocation of all available resources to ensure supply. Sottiaux emphasized that even during previous periods of rapid growth, the company had not witnessed such a high level of demand.
To safeguard the user experience for its current subscriber base, OpenAI has announced that if demand continues to escalate, they may temporarily suspend new Pro subscriptions. This measure is aimed at ensuring that existing users can continue to access and utilize the model without performance degradation.
Astra’s Rollout and Initial Challenges
GPT-6 Astra, which initially became available to select institutions on September 3rd, has since been progressively rolled out to users subscribed to ChatGPT Plus, Pro, Business, and Enterprise tiers. It is also accessible via API, Azure, and the Bedrock platform.
The launch of Astra was not without its hurdles. Shortly after its release, the model experienced capacity bottlenecks, preventing some paid users from accessing it. In response to these early issues, Tibo Sottiaux announced on September 3rd that any user who was unable to use the model for a full day would have their pre-loaded usage quota reset. OpenAI CEO Sam Altman described the initial launch as “a mess.”
Subsequently, Astra was integrated into the regular quota system. OpenAI stated that optimizations for long-tail usage scenarios have resulted in a reduction in quota consumption by as much as 75% to 67% compared to previous levels. It’s worth noting that the OpenAI Codex service previously had approximately 20 million weekly active users.
Quota Management and User Feedback
Despite optimizations, users continue to report issues such as overloaded servers and rapid quota depletion. Some users have observed that Astra consumes more quota for the same tasks compared to the Sol model. These ongoing concerns highlight the strain on OpenAI’s infrastructure.
Prioritizing Existing Subscribers
The current warning specifically targets new Pro subscriptions, not the termination of services for existing users. Tibo Sottiaux previously assured that Codex would remain available in both free and Plus subscription tiers. This quota-based restriction is a mechanism to manage resources, not a downgrade in model capabilities.
The core value proposition for paid users is access to advanced AI agents capable of coding and computer control. OpenAI, as the service provider, packages its most powerful models within fixed monthly subscription plans. By pausing new Pro subscriptions, the company aims to allocate additional computing resources towards stabilizing the experience for current users, preventing a slowdown for the entire user base. This decision ultimately benefits existing Pro and Enterprise plan subscribers by ensuring service continuity and quality.
