Journalism begins where hype ends

,,

I do not fear computers. I fear the lack of them."

—Isaac Asimov

OpenAI Launches GPT-6 Astra, Claims Most Intelligent and Aligned AI Model Yet

OpenAI has launched GPT-6 Astra, its latest frontier AI model, claiming unprecedented capabilities across coding, mathematics, cybersecurity, science, computer use and personalized agentic task execution.
OpenAI GPT-6 Astra AI model represented by a glowing spiral of artificial intelligence particles
September 4, 2026 11:38 AM IST | Written by Vaibhav Jha

Asserting it as their most “intelligent and aligned model” yet, OpenAI released GPT-6 Astra, their much hyped frontier AI model on Friday, claiming the LLM possess unprecedented capabilities in coding, mathematics, cybersecurity, science evaluation, contextual understanding, personalized agentic task execution and computer use/UI interaction.

The launch of GPT-6 Astra comes after multiple delays by OpenAI that had claimed that reinforcement learning (RL) training of the LLM was intentionally paused to align it with human developers.

“GPT-6 Astra is here. We hope it will begin to enable a new generation of entrepreneurship, scientific discovery, and building. We believe it is the best model in the world for computer use, professional work, science, coding, cybersecurity, and more,” said Sam Altman, CEO, OpenAI.

 

According to OpenAI, the GPT-6 Astra has been rolled out on Friday only to a limited set of organizations and in the coming days, it will be available to all ChatGPT Plus, Pro, Business and Enterprise users, as well as through the OpenAI API and AWS.

The AI company claimed that Astra is “state-of-the-art” model that has scored 98% in FrontierMath Tier 4, 99.9% score in ARC-AGI-3 and 100% score in ExploitBench, three different benchmarks to score an LLM on advanced mathematical reasoning, solving unfamiliar interactive tasks and exploiting real software vulnerabilities respectively.

OpenAI has also claimed that GPT-6 Astra is their most aligned model yet, as it scored 0% in misalignment compared to 48% scored by its previous frontier AI model GPT-5.6 Sol.

“GPT-6 Astra marks a new frontier in the speed, accuracy and safety of computer use. It can take care of tedious tasks like filing out online forms, updating customer records in a CRM, and organizing your calendar. It can conduct online research and draft summaries in your email. It can analyze scientific data, generate plots, create a website and run frontend QA checks as well as help users autonomously install and test software and troubleshoot problems on screen,” read a blogpost by OpenAI.

The company also claimed that Astra is trained to specifically pull only the context that matters making outputs more usable on an immediate basis.

Also Read: OpenAI’s Model Astra Solves Ten Longstanding Mathematical Problems

Author

  • Vaibhav Jha, editor and co-founder at AI FrontPage

    Vaibhav Jha is an Editor and Co-founder of AI FrontPage. In his decade long career in journalism, Vaibhav has reported for publications including The Indian Express, Hindustan Times, and The New York Times, covering the intersection of technology, policy, and society. Outside work, he’s usually trying to persuade people to watch Anurag Kashyap films.

    LinkedIn

Related Posts

LATEST NEWS