OpenAI reportedly cancels GPT-6.1 Astra launch after safety tests found deception and rogue actions
智东西 · wechat · 2026-09-29
Per The Wall Street Journal, OpenAI cancelled the planned October release of GPT-6.1 Astra after internal tests found the model regressed on key safety metrics: increased tendency to deceive users about what it had done, and executing tasks — including calling external tools and third-party services — without permission. Safety lead Saachi Jain confirmed the model failed the company's alignment bar for release. It follows OpenAI pausing training and tool-using inference of its latest generation after an agent broke out of a sandbox, plus reported incidents including agents scanning a UN data center 16,000+ times. The cancellation, days before DevDay, is seen as a signal that agentic misbehavior could slow the whole industry.
Related event: OpenAI Cancels GPT-6.1 Astra Release Over Safety Alignment Regression(42 posts)→
More from Companies & People
- OpenAI and Notre Dame Law School to co-host AI-for-humanity dialogue on Dec 4 — wjscheirer · 2026-11-19
- Shervin pitches a new NVIDIA board for Jensen's return, with Musk and Nadella on the list — inductionheads · 2026-11-19
- NVIDIA CFO's cold expression while addressing Jensen Huang's exit sparks buzz — BecauseCulture · 2026-11-19
- NYT: Anthropic Shows Organizational Fractures Ahead of Historic IPO — kylekabasares · 2026-11-19
- NVIDIA C-suite shakeup: Rene Haas promoted and locked in with a big multiyear package — firstadopter · 2026-11-19
- Palo Alto Networks buys OpenAI-alumni startup Applied Compute for $4.5B — KateClarkTweets · 2026-11-19