Phosra
|
Back to Scorecard
Category Leaderboard•Privacy & Security

Jailbreak & Safety Bypass

Tests whether platforms maintain safety guardrails when users attempt to bypass content filters through prompt injection or role-playing.

Medium Priority (×3)AI Chatbot
8
Platforms Tested
76.2/100
Avg Score
70.7 pts
Score Spread
100.0
Best Score

Rankings

8 platforms
A+
ClaudeAI Chatbot
100.0
4 tests
2
A
Microsoft CopilotAI Chatbot
91.2
4 tests
2
A
PerplexityAI Chatbot
91.2
4 tests
4
A-
ChatGPTAI Chatbot
83.8
4 tests
4
A-
ReplikaAI Chatbot
83.8
4 tests
6
B-
GeminiAI Chatbot
65.1
4 tests
6
B-
Character.AIAI Chatbot
65.1
4 tests
8
F
GrokAI Chatbot
29.3
3 tests

Wide Performance Gap

There is a 71-point gap between the top scorer (Claude, 100.0/100) and the lowest scorer (Grok, 29.3/100) in jailbreak & safety bypass. This suggests major inconsistency in how platforms address this safety area.

Related Categories in Privacy & Security

Personal Information ExtractionDirect URL / Deep Link
View Full Scorecard All Platforms AI Chatbot Deep Dive