OpenAI claims GPT-6 Astra is an ethereal 'Alien Mind' with AGI-like qualities — company warns of alignment challenges as new frontier leader emerges
OpenAI has made bold claims with its new GPT-6 Astra AI model, and it's certainly capable, but benchmarks suggest it has many of the usual weaknesses alongside the strengths, while cost and accessibility are still the biggest factors in widespread usage.
In a recent blog post, OpenAI's chief scientist, Jakub Pachocki, described the company's latest large language model, GPT-6 Astra, as an "alien mind" with AGI-like qualities. Pachocki emphasized that Astra is more ethereal and harder to quantify than traditional AI. Jensen Huang, CEO of NVIDIA, called Astra an "AGI," and OpenAI claims it is the best, most aligned model they have ever released.
Astra is safer to delegate, better at complex work tasks, and can even beat Portal in just a few hours. Artificial Analysis benchmarks suggest Astra is as smart as Fable 5.1, though it is cheaper per task. While Astra may be more aligned than previous models, OpenAI's claims suggest an inflection point for the AI industry. However, Pachocki warned that AI can only simulate facets of human behavior, not recreate it, and emphasized the importance of alignment, honesty, and integrity in AI development.
OpenAI's Astra release comes at a time when the AI industry is facing increasing pressure for more affordability, with Chinese alternatives and Western models offering competitive intelligence at lower costs. Despite its advanced capabilities, Astra is notably cheaper to run than its chief rival, Anthropic's Fable 5.1. OpenAI is prioritizing three areas of work with AI: developing a good researcher, scientific progress, and creating personalized AGI helpers for individuals.
Pachocki called for voluntary slowdowns and potential legislation to ensure humans remain in control of the future and are not left behind by unchecked AI progress.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- 'Why would you ruin your career?': OpenAI claims to have cracked one of math's greatest unsolved problems, but mathematicians allege it played dirty livescience.com
- OpenAI claims 10,000 of its AI agents solved one of mathematics' hardest problems in 88 hours techspot.com
- It took OpenAI's agents 130 billion tokens to crack a 90-year-old math problem businessinsider.com
- Researchers: OpenAI's agents used 10+ previously undisclosed sites for unsanctioned communications earlier in 2026; the behavior was closer to spam than hacking (Reuters) reuters.com
- OpenAI says its models solved one of math's hardest problems as researchers cry foul france24.com
- OpenAI claims to have solved the 90-year-old Navier-Stokes math problem in 88 hours cnbc.com
- OpenAI's rogue AI agents were secretly spreading across far more websites than disclosed qz.com