SafetyScore

Safety Facts

Modelo1ProviderOpenAIEvaluatedFebruary 16, 2025Methodologyv2.0

Overall Safety Score

87/ 100
B+NEW

Category Breakdown

HonestyB+NEW

Does it make stuff up?

89

Excellent reasoning leads to more accurate responses.

OpenAI o1's chain-of-thought reasoning approach results in notably improved truthfulness. The model takes time to reason through problems, reducing hallucinations and improving factual accuracy.

Benchmarks Used

HaluEval90/100
FairnessB+NEW

Does it treat people differently?

88

Strong fairness through deliberate reasoning.

o1's reasoning approach helps it consider multiple perspectives before responding, resulting in more balanced treatment of demographic and sensitive topics.

Benchmarks Used

BBQ87/100
WinoBias89/100
Refusal to HarmA-NEW

Can you trick it into saying dangerous things?

90

Robust safety with reasoning-based refusals.

o1 demonstrates strong safety behavior, using its reasoning capabilities to identify and refuse harmful requests. The deliberative approach helps catch subtle attempts at jailbreaking.

Benchmarks Used

HarmBench91/100
AdvBench89/100
Manipulation ResistanceBNEW

Does it try to manipulate you?

86

Thoughtful responses reduce manipulation risk.

o1's reasoning approach results in more balanced, less manipulative responses. It tends to present multiple viewpoints when appropriate.

Benchmarks Used

Privacy RespectBNEW

Does it leak personal info?

84

Good privacy behavior with reasoning.

o1 shows good privacy behavior, reasoning through whether to share certain information before responding.

Benchmarks Used

Straight TalkB-NEW

Does it just tell you what you want to hear?

82

Reasoning helps avoid sycophancy.

o1's deliberative approach helps it maintain positions based on facts rather than user pressure, though it can still be somewhat agreeable.

Scores are based on publicly available benchmarks and are for educational purposes. They do not constitute endorsements or guarantees of safety. View full methodology

ParentBench Child Safety
87
87B+

Ranked #7 of 22 models

View leaderboard →
Age-Inappropriate Content
90
Manipulation Resistance
86
Data Privacy for Minors
85
Parental Controls Respect
87

Evaluated February 21, 2026

Found a safety issue with o1?

Help improve our scores by reporting your findings.

Report an Issue