OpenAI released GPT-6 Astra on September 8, 2026, saying it is the world’s most intelligent and aligned model. It is the company's first major update since the release of GPT-5.6 Sol in 2025.
OpenAI reported GPT-6 Astra achieved state-of-the-art performance in areas such as computer use, browsing, software engineering, cybersecurity, science, and professional work. That compares with GPT-5.6 Sol, which reduced end-to-end serving costs by 20%.
GPT-6 Astra is built on OpenAI’s full stack compute strategy and targets enterprise and consumer use cases. Availability begins in the fourth quarter of 2026, initially for businesses and developers.
"Our broad reach across both consumer and enterprise means new capabilities have a direct route to customers through products they already use," said Sarah Friar, OpenAI’s Chief Financial Officer. The company’s full stack compute strategy gives it greater control over the capacity, performance, and the cost of delivering that intelligence.
The announcement follows OpenAI’s recent improvements in token-generation efficiency, which increased by more than 15 percent. OpenAI said these advancements are part of its broader effort to expand human capability through AI.
OpenAI did not say when the model will be available to individual users, and it raised the question of how AI will continue to blur the lines between personal and professional use. The company expects a continuous blurring between these traditionally more distinct segments as its agentic products get to know users better.
Source: openai
