OpenAI launches new AI model with advanced capabilities - Sarmad

OpenAI has launched its new artificial intelligence model, “GPT-6 Astra,” designed to execute complex, multi-step tasks with improvements in programming, research, computer use, and the creation of documents and presentations.

The US tech website The Verge reported that OpenAI described the new model as a significant leap in AI capabilities, particularly in computer use, programming, software engineering, science, professional business tasks, and cybersecurity.

The company explained that “GPT-6 Astra” can carry out tasks spanning from planning to completion, with enhanced abilities to understand user intent and adapt to changing requirements. It can also generate code, spreadsheets, and presentations according to user-specified templates and instructions.

The model also enables tasks involving computer use, web browsing, programming, and research, thereby strengthening its capacity to handle workflows that require a sequence of steps, rather than simply providing direct textual answers.

OpenAI stated that the model represents the first of its systems to reach a “critical” level of cybersecurity capabilities under its own readiness framework, prompting the company to strengthen security and monitoring measures to prevent misuse of these capabilities or the emergence of behaviors inconsistent with user instructions.

The company has begun rolling out the model gradually to a limited number of institutions, with plans to expand availability in the coming days. It is not yet publicly accessible.