Full Summary
This Saturday, OpenAI is rolling out its newest and most powerful AI model, GPT-6, codenamed Astra, to select customers. Both Asharq Al-Awsat English and Mashable confirm Astra is already launching. OpenAI calls Astra its "most intelligent and aligned model," designed for complex tasks across software engineering, scientific research, and professional work. However, Mashable and Pragativadi.com highlight a critical point: Astra is the first OpenAI model to be designated a "critical" cybersecurity risk. It can identify unknown security flaws and exploit system vulnerabilities without human guidance. OpenAI states they have safeguards in place, and initially, Astra's most advanced cybersecurity skills will only be available to select testing partners. OpenAI President Greg Brockman suggests we are now in the era of artificial general intelligence, with Astra capable of autonomous tasks like website creation and scientific analysis. This model can reportedly cut apartment hunting time from hours to minutes. Pragativadi.com notes Astra discovered two previously unknown vulnerabilities during internal evaluations. OpenAI plans to make Astra available to ChatGPT Plus, Pro, Business, and Enterprise users in the coming days, with API pricing set at $10 per million input tokens and $50 per million output tokens. This matches the pricing of Anthropic’s Claude Fable 5.1, which Astra just surpassed on the Code Arena: WebDev leaderboard, scoring 1,797 points. Both models also offer a one-million-token context window. Meanwhile, other AI advancements are emerging. Moonshot AI's Kimi K3 offers a powerful model with a one-million-token context window at a fraction of the cost, charging $3 per million input tokens. Z.ai's GLM-5.3-Flash, a 320-billion-parameter model, runs 3.3 times faster and is natively multimodal, accepting text, images, and video. This means that while cutting-edge AI like Astra offers unprecedented capabilities, it also brings new security considerations for users. Expect your digital interactions to become significantly more automated and potentially more vulnerable, requiring increased vigilance.