Evaluate and compare the security and safety of leading AI models. The higher the score, the more secure the model is.
| Rank | Model |
cloudsineAI's Attack dataset
|
Prompt Injection Security
|
Jailbreak Security
|
Output Safety
|
Sensitive Information
|
LLM Native Protection
|
Protector Plus Protection
|
|---|---|---|---|---|---|---|---|---|
| 1 | Claude-3.5-Sonnet | 86.9 | 97.3 | 100.0 | 99.4 | 99.9 | 93.0 |
97.9
|
| 2 | Claude-3.7-Sonnet | 85.6 | 94.6 | 99.2 | 94.8 | 97.5 | 91.1 |
98.1
|
| 3 | GPT-4o | 72.4 | 98.3 | 100.0 | 95.4 | 99.4 | 85.3 |
94.6
|
| 4 | Llama3.3:70b | 74.9 | 92.9 | 98.2 | 91.3 | 95.7 | 84.7 |
97.2
|
| 5 | grok-2-1212 | 72.8 | 92.8 | 99.8 | 86.1 | 97.4 | 83.4 |
97.1
|
| 6 | Llama3.2 | 82.6 | 78.8 | 92.8 | 80.2 | 79.1 | 82.7 |
97.5
|
| 7 | GPT-4.1 | 72.1 | 91.5 | 98.9 | 93.1 | 84.8 | 82.1 |
94.8
|
| 8 | Gemini-2.0-flash | 66.8 | 88.7 | 96.2 | 89.0 | 87.3 | 78.5 |
95.3
|
| 9 | GPT4o-mini | 70.5 | 67.7 | 97.8 | 83.1 | 84.5 | 77.0 |
96.0
|
| 10 | Grok-3 | 56.3 | 89.3 | 92.1 | 74.4 | 79.9 | 70.1 |
87.9
|
| 11 | Gemini-2.5-Flash | 34.8 | 76.8 | 88.6 | 65.6 | 71.3 | 55.2 |
90.2
|
| 12 | DeepSeek-R1-Distill-Qwen-32B | 54.2 | 68.0 | 48.6 | 39.8 | 44.0 | 52.1 |
95.0
|
| 13 | Llama2-Uncensored | 65.5 | 73.7 | 11.5 | 10.3 | 17.1 | 41.9 |
96.1
|
| 14 | DeepSeek-R1-Distill-Llama-8B | 43.8 | 52.9 | 5.2 | 7.5 | 9.0 | 31.3 |
91.5
|
Explore the leading models in our cloudsineAI attack dataset, Prompt Injection Security, Jailbreak Security, LLM Output Safety and protection against sensitive information.
Our leaderboard ranks models based on performance across these key categories, with an overall score ranging from 0.00 to 100.00. The higher the score, the more secure the model. For example, Claude 3.5 Sonnet is the most secure model we tested, with its native protection able to block 93% of all attacks. When enhanced with our GenAI firewall, the model can block almost 98% of all attacks.
We evaluated the models on our leaderboard using a combination of curated datasets and our proprietary dataset.
Evaluation Criteria: For each model, we analysed the proportion of outputs classified as harmful or safe under these conditions.
Testing with WebOrion® Protector Plus: We retested the models with WebOrion® Protector Plus protection, to assess the model’s effectiveness in blocking harmful prompts and outputs while measuring the safety of the resulting output.
Measures the model's ability to resist malicious prompt injection attempts.
Our proprietary dataset for evaluating LLM security, designed from a hacker's perspective.
Measures the model's consistency in generating safe and appropriate outputs. It covers harmful content, crime, and hate speech.
Measures the model's ability to resist advanced Jailbreak Techniques such as Analyzing-Based Jailbreak and Adaptive Jailbreak Attacks.
Evaluates the model's resistance against unethical and confidential information.
Contact our team of experts to learn how cloudsineAI can enhance your GenAI security. Leave the security to us and focus on innovating.
Fill out the form below, and we will be in touch shortly.
