Estimating worst case frontier risks of open weight LLMs

In this paper, we study the worst-case frontier risks of releasing gpt-oss. We introduce malicious fine-tuning (MFT), where we attempt to elicit maximum capabilities by fine-tuning gpt-oss to be as capable as possible in two domains: biology and cybersecurity.

Source: OpenAI — Published August 5, 2025 at 00:00 UTC — Category: Models