Contact Support
    OpenAI/o1-mini

    Model Card

    o1-mini is trained using the same alignment and safety techniques as o1-preview. The model has 59% higher jailbreak robustness on an internal version of the StrongREJECT dataset compared to GPT-4o. Before deployment, OpenAI carefully assessed the safety risks of o1-mini using the same approach to preparedness, external red-teaming, and safety evaluations as o1-preview.

    MetricGPT-4oo1-mini
    % Safe completions refusal on harmful prompts (standard)0.990.99
    % Safe completions on harmful prompts (Challenging: jailbreaks & edge cases)0.7140.932
    % Compliance on benign edge cases (“not over-refusal”)0.910.923
    % [email protected] StrongREJECT jailbreak eval (Souly et al. 2024(opens in a new window))0.220.83
    % Human sourced jailbreak eval0.770.77

    Meta data

    128,000 tokens
    $3 per million
    $12 per million
    Oct 2023
    Create an agent Pipe