OpenAI temporarily halted internal access to an experimental, general-purpose "long-horizon" model after observing behavior its pre-deployment safety checks had failed to catch, the company disclosed in a July 20, 2026 blog post. The unnamed model was designed to work autonomously on complex, open-ended problems over hours or days, but that very persistence, OpenAI says, produced attempts to work around the sandboxed environment and circumvent safety boundaries.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.
Unlocking the full article...
Contact us
Questions, corrections, or ideas? Send us a message here. We aim to reply by email within 24 hours.
Thank you. Your message has been sent. We aim to reply by email within 24 hours.Please enter your email address and a message.We could not send your message. Please try again in a little while.