OpenAI CEO Admits 'Messy' GPT-6 Astra AI Launch

OpenAI CEO Sam Altman has acknowledged the "messy" launch of the company's GPT-6 Astra artificial intelligence model, according to a post on X. The company had begun rolling out the model on September 3, touting it as its "smartest" version yet, promising state-of-the-art performance in computing, browsing, software engineering, science, and professional work.

OpenAI's initial plan was to make Astra available to all ChatGPT Plus, Pro, Business, and Enterprise users within "a few days," also planning to open it via OpenAI API, Microsoft Azure, and Amazon AWS Bedrock. However, the actual rollout saw enterprise security clients gain access before Pro subscribers, leading to significant dissatisfaction among higher-paying Pro users who expected immediate access.

To address the user frustration, OpenAI proposed a compensation mechanism. Paid ChatGPT users would receive a quota reset for future use for each day they lacked access to Astra, starting September 4. This offer aimed to mitigate discontent and ensure eventual access for all users.

By the time of reporting, Altman announced that GPT-6 Astra was being made available to all Plus, Pro, Enterprise, Business, and Business Premium users within Work and Codex. This expanded rollout now covers a wider range of paid customers who can utilize the model's enhanced capabilities.