All model comparisons
Claude 3 Haiku logo by Anthropic

Anthropic

Claude 3 Haiku

Low tier · anthropic/claude-3-haiku

Refusal Rate

88%

+40.4%

#3 of 24 models

Evaluations

7,141

Cost / 1M in

$0.25

Cost / 1M out

$1.25

Refusal Rate by Category

Crime100%
Cybersecurity100%
Dangerous100%
Deception100%
Harassment100%
Medical Misinformation100%
Self-Harm100%
Theft100%
Violence100%
Health Misinformation95%
Incitement to Violence93%
Hate Speech88%
Misinformation86%
Explicit/Sexual85%
False Positive Control10%
International Controversy0%

Analysis Deep Dives

Council Consensus

Majority Agreement

92.9%

Model's alignment with the council decision.

CAPP Score: 0.56

Political Compass
Econ (Left → Right)0.0
Social (Lib → Auth)0.0
Model Stability (Drift)

Refusal Rate Change

+39.5%

Difference over the testing period.

Start: 58.06%End: 97.52%
Paternalism Audit

Persona Refusal Rate

85.3%

Refusals for sensitive user personas.