OpenAI has launched GPT-6 Astra, calling it a “generational leap” and the start of the “AGI era,” with record scores in math, reasoning and cybersecurity—but access is still rolling out in stages. The model is first going to a small set of partners, with broader ChatGPT and API access expected within days.
Astra shows a big jump in abstract problem-solving and hard math, while also scoring 100% on a tough cyber benchmark, prompting OpenAI to label it “Critical” for cyber risk and gate some capabilities behind safety checks. In practice, this means Astra can more reliably tackle unfamiliar puzzles, advanced equations and security tasks that earlier models would often fail.
On aggregate leaderboards it trails a few top Anthropic and Meta models, but leads on several headline tests, making the frontier a tight three-way race between OpenAI, Anthropic and Meta. Companies will now pick models not just by “who is best overall,” but by which one performs strongest on their specific mix of math, coding, reasoning and safety needs.
Over the next two years, expect AI to move from smart chatbots to semi-autonomous agents that can browse, code and run workflows, even as governments and companies impose stricter safety rules on these powerful systems. For most people, this will feel less like a smarter search box and more like having a tireless digital co-worker that can actually get things done.
AI IS STARTING TO ACT, NOT JUST ANSWER.
Sanjay Sahay
Have a nice evening.