OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections
The Hacker News · Oct 27, 2026
OpenAI's compaction feature, which shrinks a long conversation to save space, could be turned into a way to copy a model's hidden reasoning out in plain text, security firm Xint has shown. OpenAI's reasoning models work through a problem before answering, and the service normally returns only the raw reasoning as an encrypted item the user cannot read. The method got past
Read original on The Hacker News
Want to stay informed about new business solutions?
Follow us
Ready to build a better digital experience?
We create modern multilingual websites, integrate external services and automate content workflows for growing businesses.