OpenAI has launched its latest AI model, GPT-6 Astra, with claims that it can navigate computers and web browsers, write software, and solve complex math problems with ease. The company hopes to make it available to paid customers soon, starting with enterprise clients and those in its Daybreak early access program.

OpenAI's president, Greg Brockman, is optimistic, declaring that the world is now in the era of artificial general intelligence. He believes that in a couple of years, people might look back and see this time as the moment AGI was truly created.

The timing is crucial for the company, as it aims to advance its technology and sales before its planned initial public offering, while also ensuring safety and security. OpenAI faces competition from Anthropic, which is also planning an IPO. In a blog post, the company describes GPT-6 Astra as the 'world’s best computer use model,' capable of performing tasks like booking appointments, searching for job listings, and apartment hunting more efficiently than humans.

However, ensuring the safe deployment of such advanced models is becoming increasingly challenging. Jakub Pachocki, OpenAI's chief scientist, emphasizes the importance of monitoring the model's internal processes to avoid unintended harm. He notes that as AI capabilities grow, so does the difficulty in monitoring them, which might become a bottleneck for further progress.