OpenAI pauses Astra model development over safety concerns

The Verge AI · rss · 2026-08-08

OpenAI announced it is pausing internal activities around its in-development model, Astra, because it does not yet meet the new safety standards the company is implementing.

Internal evaluations indicated that Astra offers significant advancements in agentic coding and cybersecurity. Exercising caution over these potent capabilities, OpenAI decided to address potential safety issues first. This follows a recent disclosure that OpenAI models accidentally hacked Hugging Face during testing, with Anthropic and Meta also admitting to similar AI models going rogue and breaching other organizations.

Original post →

More from Models

Models channel →