OpenAI CEO Sam Altman Apologizes for Chaotic GPT-6 Astra Rollout
OpenAI CEO Sam Altman apologized for the chaotic rollout of the new GPT-6 Astra model and announced that access has now expanded to all paid subscribers across Plus, Pro, Business, and Enterprise tiers. The model, released on September 3, was touted as OpenAI's most intelligent version to date with state-of-the-art capabilities in software engineering and computer usage. The launch marks a major milestone in frontier AI technology, setting new performance standards for coding and complex professional tasks. However, prioritizing enterprise security accounts over paying Pro users triggered strong pushback, highlighting the operational friction when balancing retail subscribers and enterprise enterprise commitments. To appease dissatisfied subscribers, OpenAI introduced a compensation mechanism offering a usage limit reset for each day a paying user was delayed access to Astra. The new model is being distributed across ChatGPT Work, Codex, the OpenAI API, as well as cloud platforms such as Microsoft Azure and Amazon AWS Bedrock.
## BACKGROUND
OpenAI regularly deploys flagship foundation models to power its conversational AI services and APIs, offering tiered access across individual consumer and corporate subscriptions. Enterprise cloud platforms like Amazon Bedrock provide managed environments that enable businesses to integrate these foundation models into their software stack securely.