GPT-6 Astra Just Went CRITICAL...
Summary
OpenAI's new Astra model reached a "Critical" cybersecurity threshold, according to the company. The Information reports Astra uses "looped transformers" for latent space reasoning, a technique that could obscure model thinking. This development follows warnings from researchers about AI safety and the potential for rogue agents. The new architecture raises concerns about understanding AI decision-making processes.
Why it matters
Why it matters: OpenAI's Astra model's "Critical" cybersecurity rating and its use of "looped transformers" mark a significant step in AI development. This new architecture, which reasons in latent space rather than text, could make AI behavior harder to monitor, echoing concerns from a recent paper on "Chain of Thought Monitorability." This development affects AI safety researchers and developers across the field, including competitors like Google Gemini and Anthropic. Future focus should be on how this latent reasoning impacts transparency and safety protocols.
Related: OpenAI: OpenAI’s next big AI model has ‘entered the AGI era’ · Anthropic: GPT-6 Astra Just Went CRITICAL... · Microsoft: What's new in AI? · Meta: Get the full story behind the light · xAI: Sam Altman "AGI by December"
Rated high: a launch, deal or policy change that changes what you can do or what it costs.
- GPT-6 Astra Is Here—and OpenAI Thinks It May Kick Off the AGI EraWIRED AI · Web · Sep 3, 2026
- OpenAI’s new reasoning technique alarms AI safety expertsTechCrunch AI · Web · Sep 2, 2026
- Researchers fear safety disaster ahead of OpenAI’s Astra releaseThe Verge AI · Web · Sep 2, 2026
- OpenAI hails ‘new era of artificial general intelligence’ with Astra model releaseThe Guardian AI · Web · Sep 3, 2026
- The Most Overhyped and Underhyped New AI ModelsMatt Wolfe · Web · Sep 2, 2026
- OpenAI’s Altman Unveils Astra as a New Step Toward AGI | The Close 9/3/2026Bloomberg Technology · Web · Sep 3, 2026
- LWiAI Podcast #255 - Gemini 3.7, Jalapeño, Qwen 3.8, DronesLast Week in AI
- Whose Assessment of Distress? Community Perspectives and LLM Alignment on Well-Being PostsarXiv cs.AI
- Understanding the inner thoughts of AIGoogle DeepMind on YouTube
- When millions of AI agents meetGoogle DeepMind on YouTube
- Claude Fable 5.1 made me a really nice animated pelicanSimon Willison
- Mapping global methane emissions from space with deep learningGoogle Research Blog
Questions people ask
- What is OpenAI's Astra model?
- Astra is OpenAI's upcoming model that has reached a "Critical" cybersecurity threshold under its Preparedness Framework. It reportedly uses a new technique called "looped transformers."
- What are "looped transformers"?
- Looped transformers, or recurrent depth, is a new technique reportedly used by Astra. It allows the model to reason in latent space instead of readable text, which raises concerns about transparency.
More from Wes Roth
This is the first item Subvolts has from Wes Roth. More arrive with each daily run.
Page generated Sep 3, 2026. Summaries are Subvolts' own; the story belongs to Wes Roth.