OpenAI has launched GPT-6 Astra, its latest AI model that the company says excels at navigating computers and web browsers, writing software, and solving complex maths problems. The model will be available to paid customers in the coming days, starting with enterprise clients before rolling out to ChatGPT Plus, Pro, and Business subscribers. OpenAI president Greg Brockman stated he believes the world has entered the artificial general intelligence era, suggesting GPT-6 Astra may mark the creation of AGI when looking back in a few years. The company claims GPT-6 Astra can book DMV appointments, search for jobs, and hunt for apartments faster than average humans. However, chief scientist Jakub Pachocki acknowledged that monitoring advanced AI models for safety is becoming increasingly challenging and may constrain future development.

OpenAI has launched GPT-6 Astra, a new artificial intelligence model featuring advances in computer use, coding, scientific research and professional work. The San Francisco-based company said Astra is rolling out to enterprise customers with Daybreak access, with ChatGPT Plus, Pro, Business and Enterprise users to follow. The model can perform computer-based tasks such as filling out forms, updating records and organising calendars. It can also create documents, spreadsheets and presentations. OpenAI described Astra as "the best model for software engineering to date" and highlighted its cybersecurity capabilities, which meet the Critical threshold under the company's Preparedness Framework. The company also touted performance improvements in mathematics and science.

OpenAI released GPT-6 Astra on Thursday, calling it the world's most intelligent AI system and the first model to trigger advanced internal safety protections due to its cybersecurity capabilities. The company said Astra can autonomously control computer systems and perform tasks such as filling out spreadsheets and creating websites. "Astra can really do anything a human can do with a computer," said OpenAI co-founder Greg Brockman. The model is initially available to participants in OpenAI's Daybreak programme for cybersecurity defenders, with wider access planned for coming days. OpenAI said Astra can find previously unknown security flaws and develop exploits across well-protected systems without human guidance. The release follows Anthropic's launch of its Fable 5.1 and Mythos 5.1 models earlier this week.