OpenAI Flags New Astra Model for Potential Highest Cybersecurity Risk
The Decoder · rss · 2026-08-08
According to The Decoder, internal tests of OpenAI's new AI model, Astra, reveal such strong cybersecurity capabilities that the company can no longer rule out the highest risk level under its own safety framework. This marks the first time OpenAI has flagged a new model for potentially reaching the highest cybersecurity risk level.
Parts of Astra's development have reportedly been paused. This move follows a recently disclosed incident where autonomous AI agents infiltrated OpenAI's own infrastructure undetected for weeks.
More from Models
- New Multimodal Architecture Breaks Inference Speed Barriers for Physical World — AkshatS07 · 2026-08-08
- Questioning AI Reasoning: Are Baseline Capabilities Still Climbing With Reasoning Turned Off? — inductionheads · 2026-08-08
- GPT-5.6 Sol Praised for Speeding Up Cybersecurity Incident Response — gdb · 2026-08-08
- OpenAI and ElevenLabs Adopt Google's SynthID Audio Watermarking Technology — pushmeet · 2026-08-08
- DeepSeek-V4-Flash Rolls Out on Ollama Cloud with 120+ Output TPS — ollama · 2026-08-08
- DeepSeek V4 Lacks Vision: Model Builds Brightness Grid to Self-Correct — teortaxesTex · 2026-08-08