OpenAI made GPT-6 Astra available to paying ChatGPT users on Friday, completing the first phase of distribution a day after the model's public debut. CEO Sam Altman acknowledged the rocky start with a message on X: "Thank you for the patience."
By 6:30 p.m. Eastern on Friday, September 4, GPT-6 Astra had reached all ChatGPT subscribers whose plans included access. Thibault Sottiaux, a leading technical staff member at OpenAI, announced the expansion: "OK nevermind, the team and Astra did a good job and our systems are more scalable than we anticipated. Astra is now rolled out to all Plus and Business users too. Hope you have a blast and let us know how it goes!"
Earlier that day, Altman had indicated a phased approach, stating that GPT-6 Astra was available to Pro, Enterprise, and Business Premium users in Work/Codex and through the API, with Plus and Business users to follow. Subscribers to the $8/month ChatGPT Go plan lack access to both GPT-6 Astra and GPT-5.6, per the company's tier documentation.
The initial rollout proved problematic. Hours after Thursday's announcement, Altman apologized for the deployment difficulties, acknowledging that neither API customers nor ChatGPT subscribers had yet gained broad access.
https://x.com/thsottiaux/status/2096002992046796932?ref_src=twsrc%5Etfw
First, sorry for the messy rollout
Sam Altman, OpenAI CEO
Altman indicated that wider distribution would commence "in the near future," beginning with ChatGPT Pro users. The situation left developers in an awkward position: OpenAI had already published API documentation and pricing details, including the 1.05 million-token context window, support for up to 128,000 output tokens, and rates of $10 per million input tokens and $50 per million output tokens. Yet the actual endpoint remained unavailable, and the company had not disclosed what triggered the rollout problems. Developers and independent analysts seeking to validate Astra's impressive benchmark results against their specific use cases remained unable to do so.
Why developers are waiting
https://x.com/sama/status/2095678759651438887?ref_src=twsrc%5Etfw
Sottiaux provided additional context regarding the deployment challenges. He explained that completing the rollout would require "a few days" and that OpenAI was activating new infrastructure and additional computational resources to expand availability.
We are starting to release GPT-6 Astra, and we are doing it as carefully and quickly as possible
Thibault Sottiaux, OpenAI
Sottiaux noted that "many novel systems will operate at scale for the first time" during the rollout and that OpenAI was "bringing a lot of compute up." While this does not definitively establish that compute limitations caused the initial delay, it illuminates the operational challenges OpenAI faced during the expansion. The company's original announcement had stated that Astra would start with limited access to select organizations, followed by availability for ChatGPT Plus, Pro, Business, and Enterprise users, the OpenAI API, Microsoft Azure, and AWS Bedrock "over the coming days."
Developers who obtained early API access have already discovered additional complications beyond the rollout itself. Astra's API introduces a new category of safety-triggered interruptions that resemble timeouts but function differently, potentially affecting developers constructing production systems around the model.
Banked resets explained briefly
To partially offset the access delays, OpenAI introduced what it terms a "banked reset" compensation mechanism. Sottiaux announced that paid ChatGPT subscribers would receive one banked reset for each day they lacked Astra access, beginning September 3.
Team is moving mountains to give access as fast as we can.
Thibault Sottiaux, OpenAI
OpenAI has previously deployed banked resets with ChatGPT Work and Codex, enabling users to restore usage after reaching limits. These differ from additional API credits or permanent usage increases; they represent resets that users can accumulate and apply when needed. Astra's approach distinguishes itself through the distribution method: one reset per day without access for each paying ChatGPT subscriber. To date, OpenAI has not announced comparable compensation for developers awaiting API access.
The compensation strategy reflects OpenAI's broader experimentation with AI pricing mechanisms. The company recently unveiled an outcome-based pricing structure that charges only when the model delivers correct results, signaling that its pricing approach continues to evolve.
Compute capacity complicates rollout
OpenAI indicated that GPT-6 Astra would reach Plus, Pro, Business and Enterprise users, plus the OpenAI API, Microsoft Azure and AWS Bedrock "over the coming days," with Pro subscribers receiving priority. Sottiaux estimated the full rollout would span several days.
For the developer community, the critical outstanding question concerns when comprehensive API access will materialize and whether it will arrive with usage restrictions. Astra's persistent-agent functionality amplifies the significance of this delay for developers. OpenAI did not provide immediate responses to inquiries regarding the rollout complications, banked resets, or API availability timeline.
Source: The New Stack