All model comparisons
Gemini 2.0 Flash logo by Google

Google

Gemini 2.0 Flash

Mid tier · google/gemini-2.0-flash-001

Refusal Rate

68%

+23.1%

#15 of 22 models

Evaluations

2,756

Cost / 1M in

$0.1

Cost / 1M out

$0.4

Refusal Rate by Category

Crime100%
Cybersecurity100%
Dangerous100%
Deception100%
Harassment100%
Medical Misinformation100%
Self-Harm100%
Theft100%
Violence100%
Health Misinformation91%
Explicit/Sexual78%
Hate Speech75%
Incitement to Violence70%
Misinformation29%
False Positive Control3%

Analysis Deep Dives

Council Consensus

Majority Agreement

85.2%

Model's alignment with the council decision.

CAPP Score: 0.55

Political Compass
Econ (Left → Right)-1.4
Social (Lib → Auth)+2.4
Model Stability (Drift)

Refusal Rate Change

+23.4%

Difference over the testing period.

Start: 52.26%End: 75.68%
Paternalism Audit

Persona Refusal Rate

68.0%

Refusals for sensitive user personas.