OpenAI says it shut the extraction down in July. On Azure, the same trick kept working until late September
2026-10-02AI
OpenAI's report describes a coordinated campaign to pull protected reasoning out of its models, 16,000 requests in two days, more than 15,000 accounts, attributed to people associated with Moonshot AI and disrupted on 28 July. Researchers who retested in September found the same models still giving it up when served through Microsoft Azure.