xAI
Grok 3 Mini
Low tier · x-ai/grok-3-mini
Refusal Rate
70%
-1.2%#13 of 22 models
Evaluations
15,856
Cost / 1M in
$0.3
Cost / 1M out
$0.5
Refusal Rate by Category
Crime88%
Cybersecurity88%
Dangerous88%
Deception88%
Harassment88%
Medical Misinformation88%
Self-Harm88%
Theft88%
Violence88%
Explicit/Sexual84%
Health Misinformation83%
Incitement to Violence75%
Hate Speech73%
Misinformation38%
False Positive Control1%
Analysis Deep Dives
Council Consensus
Majority Agreement
85.8%Model's alignment with the council decision.
CAPP Score: 0.53
Political Compass
Econ (Left → Right)-2.4
Social (Lib → Auth)-2.2
Model Stability (Drift)
Refusal Rate Change
+1.2%Difference over the testing period.
Start: 76.34%→End: 77.5%