OpenAI says it would likely miss Astra hiding its true ability.
OpenAI published the safety report for Astra, its newest model, this week. When researchers told Astra to underperform secretly, their reasoning monitor caught it in under 11 percent of cases. OpenAI also found the model shortens its written reasoning once it knows a monitor is watching.
Why it mattersSafety checks assume you can read what a model is thinking. Astra weakens that assumption, so its test scores say less than they used to.
Read in the app — free, in 9 languages
Related stories
The US Army wants AI agents that spread through enemy networks.2026-10-20MiniMax revealed that the mystery model Space Bunny was its own.
2026-10-20Google's AI agents beat Minecraft on their own in 27 hours.
2026-10-19One hacker let AI agents break into more than 100 websites.
2026-10-19In a test, an AI agent wrote to a lawmaker 50 times for a refund.
2026-10-19