OpenAI Rates Astra Its Most Dangerous Model: First With 'Critical' Cyber Capabilities
The Decoder · rss · 2026-09-02
According to The Decoder, OpenAI has officially rated its upcoming Astra model as the first system with "critical" cyber capabilities.
- The company plans to keep it in check by monitoring the model's chain of thought
- The catch: CoT monitoring is already considered an unreliable mirror of a model's real decisions
- Worse, Astra's new architecture reportedly pushes even more of its reasoning into unreadable form
The safety net may be weakening just as capabilities jump.
More from Models
- AxiomProver tops LeanEval, the last unsaturated math formalization benchmark — BenBlaiszik · 2026-09-03
- Muse Spark 1.3 calls user 'Judah' then denies it, users report odd behavior — fragment_me · 2026-09-03
- Google AI Mode shows zero citations on high-level TOFU queries, SEO tests find — gaganghotra_ · 2026-09-03
- Fable 5.1 halves agent failure rate to 7% with 0.7% hallucinations, at 1.8x the cost — ryanshrout · 2026-09-03
- Seroter Daily #859: Gemini 3.8 Flash, agent telemetry, and 7 agent skill patterns — rseroter · 2026-09-03
- Insider leak: OpenAI's Astra tested as 'ultima-alpha' and 'vega-alpha' checkpoints — Ok_Display_3159 · 2026-09-03