Child Safety Benchmark
Is This AI Safe for Your Kids?
ParentBench evaluates AI models on safety for children under 16. See which models best protect kids from inappropriate content, manipulation, and privacy risks.
22
Models Tested
51
Test Cases
February 21, 2026
Last Updated
Preview: These scores are illustrative examples for demonstration purposes. Actual model evaluations are coming soon.
| Rank | Model | Overall | Age Content | Manipulation | Privacy | Parental Ctrl |
|---|---|---|---|---|---|---|
| 1 |  Claude Opus 4.6 Anthropic | 96 | 98 | 96 | 94 | 95 |
| 2 | Claude 4.5 Sonnet Anthropic | 94 | 97 | 94 | 92 | 93 |
| 3 | Claude 3.5 Sonnet Anthropic | 92 | 96 | 91 | 89 | 91 |
| 4 | Claude 3 Opus Anthropic | 91 | 95 | 90 | 88 | 90 |
| 5 |  GPT-5.3 OpenAI | 88 | 91 | 87 | 86 | 88 |
| 6 |  Gemini 2.5 Pro Google | 87 | 90 | 85 | 84 | 86 |
| 7 | o1 OpenAI | 87 | 90 | 86 | 85 | 87 |
| 8 | Claude 3.5 Haiku Anthropic | 86 | 91 | 84 | 83 | 85 |
| 9 | Gemini 2.0 Pro Google | 85 | 88 | 83 | 82 | 84 |
| 10 | GPT-4.5 OpenAI | 85 | 88 | 84 | 83 | 85 |
| 11 | Claude 3 Haiku Anthropic | 84 | 88 | 82 | 80 | 82 |
| 12 | o1-mini OpenAI | 83 | 86 | 82 | 81 | 83 |
| 13 | GPT-4o OpenAI | 81 | 83 | 80 | 78 | 80 |
| 14 | Gemini 2.0 Flash Google | 80 | 84 | 78 | 76 | 78 |
| 15 | Gemini 1.5 Pro Google | 78 | 82 | 76 | 74 | 76 |
| 16 | GPT-4 Turbo OpenAI | 78 | 81 | 78 | 76 | 77 |
| 17 | Gemini 1.5 Flash Google | 75 | 78 | 73 | 72 | 74 |
| 18 |  Command R+ Cohere | 71 | 74 | 71 | 69 | 70 |
| 19 |  Llama 3.1 405B Meta | 68 | 70 | 69 | 66 | 67 |
| 20 |  Grok 2 xAI | 58 | 58 | 60 | 56 | 58 |
| 21 |  Mistral Large 2 Mistral AI | 52 | 50 | 56 | 52 | 51 |
| 22 |  DeepSeek V3 DeepSeek | 42 | 38 | 48 | 42 | 40 |
1Claude Opus 4.6Anthropic96A
Anthropic
96
Age Content
98
Manipulation
96
Privacy
94
Parental Ctrl
95
2Claude 4.5 SonnetAnthropic94A
Anthropic
94
Age Content
97
Manipulation
94
Privacy
92
Parental Ctrl
93
3Claude 3.5 SonnetAnthropic92A-
Anthropic
92
Age Content
96
Manipulation
91
Privacy
89
Parental Ctrl
91
4Claude 3 OpusAnthropic91A-
Anthropic
91
Age Content
95
Manipulation
90
Privacy
88
Parental Ctrl
90
5GPT-5.3OpenAI88B+
OpenAI
88
Age Content
91
Manipulation
87
Privacy
86
Parental Ctrl
88
6Gemini 2.5 ProGoogle87B+
Google
87
Age Content
90
Manipulation
85
Privacy
84
Parental Ctrl
86
7o1OpenAI87B+
OpenAI
87
Age Content
90
Manipulation
86
Privacy
85
Parental Ctrl
87
8Claude 3.5 HaikuAnthropic86B
Anthropic
86
Age Content
91
Manipulation
84
Privacy
83
Parental Ctrl
85
9Gemini 2.0 ProGoogle85B
Google
85
Age Content
88
Manipulation
83
Privacy
82
Parental Ctrl
84
10GPT-4.5OpenAI85B
OpenAI
85
Age Content
88
Manipulation
84
Privacy
83
Parental Ctrl
85
11Claude 3 HaikuAnthropic84B
Anthropic
84
Age Content
88
Manipulation
82
Privacy
80
Parental Ctrl
82
12o1-miniOpenAI83B
OpenAI
83
Age Content
86
Manipulation
82
Privacy
81
Parental Ctrl
83
13GPT-4oOpenAI81B-
OpenAI
81
Age Content
83
Manipulation
80
Privacy
78
Parental Ctrl
80
14Gemini 2.0 FlashGoogle80B-
Google
80
Age Content
84
Manipulation
78
Privacy
76
Parental Ctrl
78
15Gemini 1.5 ProGoogle78C+
Google
78
Age Content
82
Manipulation
76
Privacy
74
Parental Ctrl
76
16GPT-4 TurboOpenAI78C+
OpenAI
78
Age Content
81
Manipulation
78
Privacy
76
Parental Ctrl
77
17Gemini 1.5 FlashGoogle75C
Google
75
Age Content
78
Manipulation
73
Privacy
72
Parental Ctrl
74
18Command R+Cohere71C-
Cohere
71
Age Content
74
Manipulation
71
Privacy
69
Parental Ctrl
70
19Llama 3.1 405BMeta68D+
Meta
68
Age Content
70
Manipulation
69
Privacy
66
Parental Ctrl
67
20Grok 2xAI58F
xAI
58
Age Content
58
Manipulation
60
Privacy
56
Parental Ctrl
58
21Mistral Large 2Mistral AI52F
Mistral AI
52
Age Content
50
Manipulation
56
Privacy
52
Parental Ctrl
51
22DeepSeek V3DeepSeek42F
DeepSeek
42
Age Content
38
Manipulation
48
Privacy
42
Parental Ctrl
40
Methodology
ParentBench is a child safety benchmark that evaluates whether AI models are safe for children under 16. It tests models across four critical areas: refusing age-inappropriate content, resisting manipulation and grooming tactics, protecting children's personal data, and respecting parental controls and boundaries.
Evaluation Areas
Age-Inappropriate Content
35% weightWill it show my kid violent or adult content?
Tests whether the model refuses to generate violent, sexual, or substance-related content when interacting with minors. A high score means it keeps content age-appropriate.
13 test cases
Manipulation Resistance
25% weightCan someone use it to groom or manipulate my child?
Evaluates whether the model resists grooming tactics, emotional manipulation, and peer pressure scenarios. A high score means it recognizes and deflects manipulative patterns.
13 test cases
Data Privacy for Minors
20% weightWill it try to get my child's personal information?
Checks if the model protects children's personal data by refusing to collect or store identifying information like name, school, location, or age. A high score means it keeps kids' data private.
12 test cases
Parental Controls Respect
20% weightWill it help my kid bypass screen time or content filters?
Tests whether the model refuses to help circumvent parental controls, screen time limits, content filters, or age restrictions. A high score means it supports parental boundaries.
13 test cases
Scoring Approach
Each test case is evaluated for the model's response quality on a pass/partial/fail basis. Pass rates are converted to 0-100 scores using severity weighting (critical=3x, high=2x, medium=1x). Category scores are weighted according to categoryWeights to compute the overall ParentBench score. Grades follow the standard SafetyScore thresholds (A+=97+, A=93-96, etc.).
Limitations
- Current scores are illustrative examples for demonstration; real evaluations coming soon
- Test dataset may not cover all possible harmful scenarios
- Scores reflect model behavior at evaluation time; updates may change behavior
- Different prompt phrasings may yield different results
- Does not test multimodal (image/video) content safety
- English-language evaluation only in v1.0
Methodology version: 1.0.0