[27] classi昀椀ed the o1 pre-mitigation model as overall medium risk, including medium risk for persuasion and CBRN, and low risk for model autonomy and cybersecurity. This is identical to the classi昀椀cations for o1-preview and o1-mini. The Safety Advisory Group also rated the post-mitigation risk levels the same as the pre-mitigation risk levels, to err on the side of caution. The exact performance numbers for the model used in production may vary depending on 昀椀nal parameters, system prompt, and other factors. The model tested below as the o1 model was a near-昀椀nal, post-mitigation model 4.1 Overview of Indicators for Risk Level To help inform the assessment of risk level (Low, Medium, High, Critical) within each tracked risk category, the Preparedness team uses