OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections

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 Xinthas shown. OpenAI's reasoning models work through a problem before answering, and the service normally returns only the raw reasoning asan encrypted itemthe user cannot read. The method got past