OpenAI Slows Down Next-Gen Astra Model Development Over Cybersecurity Risks
APPSO · wechat · 2026-08-08
During internal testing, OpenAI discovered that its unreleased next-generation model, Astra, exhibits critical cybersecurity capabilities, including finding vulnerabilities and autonomously planning attack steps. Consequently, the company has proactively slowed its development and placed the model in a fully isolated testing environment with universal monitoring.
CEO Sam Altman acknowledged that while restricting access to powerful models isn't ideal, more time is needed to ensure Astra's safety before a public release. OpenAI also voluntarily reported this delay to the US government. The article contrasts this move with Anthropic's previous policy reversals regarding safety pauses and highlights the ongoing confusion within the industry as the US government drafts its own AI evaluation framework.
Related event: OpenAI Slows Astra Development Over Cyber Risk, Limits Initial Release(30 posts)→
More from Models
- GLM 5.5 rumored to drop soon, coding ability expected to match frontier models — bindureddy · 2026-10-16
- Five rounds of adversarial prompting make Opus 5.5 and Argon converge on identical output — LanceCharger · 2026-10-16
- DIY benchmark ranks 18 models on subjective judgments, GLM-5.3 tops consensus match — arkuto · 2026-10-16
- Bindu Reddy: Google's Argon excels at complex doc writing at half the cost of Opus — bindureddy · 2026-10-16
- MAI-2 is 2x bigger than V4-Pro while V4.1F is 3x smaller: Microsoft's scaling ladder questioned — teortaxesTex · 2026-10-16
- Step 5 Preview hits DeepInfra day-zero: 600B MoE, 1M context, $1/M input tokens — StepFun_ai · 2026-10-16