OpenAI slows Astra work over cybersecurity concerns
The company says preliminary evaluations found advanced agentic coding and cybersecurity capabilities.
Why it matters
The move shows frontier AI developers are treating advanced cyber capabilities as a release-gating safety issue, not only a post-launch monitoring problem. It also comes amid broader scrutiny after OpenAI, Anthropic and Meta disclosed incidents involving models breaching other organizations.
The key points
- 1.OpenAI paused some Astra work after cybersecurity evaluations.
- 2.Astra showed advances in agentic coding and cybersecurity.
- 3.The company is strengthening safeguards and security controls.
OpenAI said it has suspended or paused some internal work on Astra, an in-development model, after preliminary cybersecurity evaluations and expert assessments raised security concerns. The company said Astra showed significant advances in agentic coding and cybersecurity and does not yet meet new security standards it is putting in place.
⚡ Try this today
Review OpenAI’s Astra cybersecurity evaluation before building or deploying agentic coding systems with cyber-relevant capabilities.
Sources & original reporting
This brief summarizes and links to reporting from the publishers below.
Enjoyed this brief? Get the next one in your inbox.
More in Models
Qwen3.8 27B scores 52 on Artificial Analysis
Reddit and Hacker News users flagged a large Agentic Index jump for Qwen3.8 27B.
Google introduces Gemini 3.7 Flash
Google and DeepMind published posts introducing the Gemini 3.7 Flash model.
Meta releases open-weight Glimmer AI model
The downloadable model contrasts with Meta’s more powerful API-only Muse Spark.