On September 3, 2026, OpenAI unveiled its new flagship model GPT-6 “Astra,” calling it the company’s “most intelligent and aligned” system to date. The rollout began with selected organizations, with broader availability following a day later.
What OpenAI Promises
According to the vendor, GPT-6 Astra sets new highs in computer use, web browsing, software engineering, cybersecurity, science and professional work. OpenAI cites benchmark scores of 98% on FrontierMath (Tier 4), 99.9% on ARC-AGI-3 and 100% on ExploitBench. These figures are vendor-reported and have not yet been independently verified. For computer use, OpenAI puts the speed at nearly double that of its predecessor.
Availability and Pricing
Astra is included in the ChatGPT Plus, Pro, Business and Enterprise plans and is also accessible via the OpenAI API and Amazon Web Services. For enterprise customers, access is disabled by default at launch and must be enabled manually. Usage is priced above the previous model, GPT-5.6 “Sol.”
A Safety Debate in the Background
The launch comes during a period of heightened attention to AI safety. OpenAI itself points to the model’s potential for harm, while experts warn that capabilities are advancing faster than an understanding of the associated risks. Whether Astra lives up to the lofty expectations in everyday use will only become clear through independent testing.
Source: OpenAI, “GPT-6 Astra: A new generation of intelligence” (openai.com); additional reporting by Al Jazeera.
