OpenAI Slows Down Astra Development Citing Critical Cyber Risks
moyix · x · 2026-08-08
OpenAI confirmed to Axios that it is slowing down the internal development of its Astra project. The company stated it cannot rule out the model's "critical cyber capabilities," adding that it will scale up safety testing and pause some internal activities.
Commenting on the news, Andrew Curran speculated that Astra is actually finished but was sent back to safety review after being demoed to the administration. He also noted that the slowdown in training new models is real, and that the recent Hugging Face incident has genuinely triggered an internal rethink on current safety protocols.
Related event: OpenAI Delays Astra Model Release Over Cybersecurity Risks(6 posts)→
More from Models
- OpenAI's Upcoming Model Astra Triggers High-Level Safety Protocols — deanwball · 2026-08-08
- Qwen3.8-Max beats Gemini 3.5 Flash by 8.5 points in tests — usamawahabkhan · 2026-08-08
- DeepSeek V4 Flash Appears on ARC Prize Leaderboard — tosh · 2026-08-08
- Pokee AI Launches Isaac Model with 10M-Token Context and API — Kyrannio · 2026-08-08
- Opinion: Google or Meta Could Win the AI Race by Dropping a Better Open-Source Model Than K3 — bindureddy · 2026-08-08
- Why LLMs Can't Count Tokens: The Need for Intermediate Steps — ctjlewis · 2026-08-08