OpenAI Launches GPT-6 Astra: Most Powerful AI Model Yet, September 5

OpenAI began rolling out GPT-6 Astra on September 4, its most powerful AI model yet, to select customers and enterprise users. The model scored 100% on exploit benchmarks and reached a critical cybersecurity threshold, prompting OpenAI to restrict access initially to trusted defenders. OpenAI President Greg Brockman said the model represents a generational leap and suggested the company has entered the artificial general intelligence era.

What Astra can do

GPT-6 Astra functions as an autonomous AI agent, handling complex tasks across software engineering, cybersecurity, financial modeling, and data analysis. The model can browse the web, operate software, edit spreadsheets, generate code, and complete multi-step workflows with minimal human intervention. OpenAI demonstrated that Astra could cut apartment hunting from six hours to under 10 minutes, illustrating efficiency gains for business workflows.

Security capabilities and risks

Astra scored 100% on ExploitBench, up from 78.5% for predecessor GPT-5.6 Sol, and 42.4% on ExploitGym, versus 30.3% for Sol. The model discovered two previously unknown zero-day vulnerabilities during internal testing, which OpenAI disclosed to software makers. Because these capabilities pose dual-use risks, OpenAI restricted access through its Daybreak Blue program for trusted cybersecurity defenders, with enterprise administrators required to manually enable Astra for their workspace.

Rollout and pricing

Cybersecurity customers gained access on September 4, with broader rollout to ChatGPT Plus, Pro, Business, and Enterprise users to follow. Free-tier users will not gain access. API pricing is set at $10 per million input tokens and $50 per million output tokens, matching competitor Claude Fable. CEO Sam Altman apologized for a messy rollout, promising wider availability quickly.

Competition and AGI claims

Astra competes directly with Google Gemini and Anthropic Claude, though benchmarks vary by test. Brockman told reporters that people looking back in a few years could identify Astra as the point when AGI was created, stating personally, “I do think we’re there.” OpenAI’s official documentation focuses on capabilities and safety rather than formally declaring AGI achieved.

Final Thoughts

Astra marks OpenAI’s most capable model yet, with breakthrough exploit-finding abilities and autonomous task handling. The controlled rollout reflects genuine security tradeoffs. For investors in AI infrastructure and cloud platforms like Microsoft, the model’s enterprise adoption timeline and pricing will shape competitive positioning.

FAQs

Disclaimer:

The content shared by Meyka AI PTY LTD is solely for research and informational purposes.  Meyka is not a financial advisory service, and the information provided should not be considered investment or trading advice.