OpenAI Unveils GPT-6 Astra
News Desk
San Francisco: What if an AI could do more than answer your questions, and actually use a computer, browse the web, write software and complete complicated tasks on its own?
That is the direction OpenAI is taking with GPT-6 Astra, its latest artificial intelligence model, which the company says represents a major step toward AI systems capable of carrying out complex, multi-stage work with far less human intervention.
OpenAI says Astra has been significantly upgraded in computer use, web browsing, software engineering, cybersecurity, scientific research and professional tasks. Unlike traditional chatbots that primarily generate responses, the new model is designed to interact with digital tools and take action.
In demonstrations, Astra has been shown handling demanding assignments such as developing video games, creating websites, conducting online research and working through technical problems. The model can also perform sequences of tasks rather than stopping after producing a single answer.
OpenAI reports strong results across several advanced evaluations, including a 99.9 percent score on ARC-AGI-3 and 97.6 percent on FrontierMath Tier 4.
But Astra’s most striking capability may also be its most controversial.
OpenAI has classified the model at a “Critical” level for cybersecurity capability. The company says Astra can identify previously unknown vulnerabilities and develop ways to exploit them when provided with the necessary tools and access.
That raises a bigger question about the future of AI: when a model can operate computers and make decisions across multiple steps, how much control should humans retain over what it does?
OpenAI says it has introduced additional safeguards, monitoring and security measures alongside the model’s expanded capabilities. The company is initially making Astra available to a limited group of organisations before expanding access to ChatGPT Plus, Pro, Business and Enterprise users, as well as through its API and selected cloud platforms.
The launch also signals a broader change in the AI race. OpenAI, Anthropic and Google are increasingly competing not simply over which system can produce the most convincing answer, but over which AI can actually get the job done.
With GPT-6 Astra, OpenAI is betting that the next generation of AI will look less like a chatbot waiting for instructions and more like a digital worker capable of taking those instructions and turning them into action.
The technology could reshape how people interact with computers, but it also puts a new spotlight on the risks of giving increasingly capable AI systems greater independence.
