Back to events
Model releaseCritical global significanceConfirmed confidence

OpenAI introduces GPT-6 Astra with limited access for trusted enterprises

GPT-6 Astra opens to Trusted Access Program enterprises with a 1.05M-token context window. Broader API and paid ChatGPT access is expected in the coming days.

Event details

On September 3, 2026, OpenAI's model directory and documentation listed GPT-6 Astra for complex reasoning, coding, computer use, research and document creation, with initial access that day for enterprises in the Trusted Access Program. OpenAI also published a Playco customer story featuring GPT-6 Astra. The model documentation lists a 1,050,000-token context window and a maximum output of 128,000 tokens. It supports text and image input, text output, streaming, function calling and structured output. Standard text API prices are $10 per million input tokens, $1 per million cached input tokens and $50 per million output tokens. Entire requests with more than 272K input tokens are charged at higher rates. This is a limited release. OpenAI says broader access through the API and Plus, Pro, Business and Enterprise plans will follow in the coming days, without specifying a date. A September 1 safety statement says Astra has reached the Critical cybersecurity capability threshold, with stricter access controls for advanced cybersecurity work. Earlier rumors of an August 2026 Astra release did not materialize within that window and remain in a separate forecast archive.

Why it matters

OpenAI's next major model enters limited availability, extending complex professional work and multimodal reasoning while introducing access boundaries consistent with Critical cybersecurity capabilities.

88/100Global significance score. Regional effects are recorded only when the evidence supports a meaningful difference.

Access notes

Initial access is limited to Trusted Access Program enterprises. Broader API and Plus, Pro, Business and Enterprise access remains forthcoming.