aiXamine - Model & Data Privacy: leaderboard
aiXamine Model & Data Privacy service score: safeguarding sensitive information against memorization-based leakage and privacy-violating requests (aiXamine, arXiv 2608.20554v1, Appendix Table 7). Normalized 0-100 score aggregated from the service's tests; higher is better. 42 models evaluated under identical black-box conditions via OpenRouter or containerized vLLM.
Metric: Service score (0-100; normalized pass rate). Source: arxiv.org. 42 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 91.1 | #357 |
| 2 | Llama 3.2 3B | 88.1 | #1376 |
| 3 | Llama 3.2 1B | 86.2 | #1540 |
| 4 | DeepSeek R1 0528 | 84.7 | #232 |
| 5 | Phi-4 | 84.2 | #707 |
| 6 | Claude 3.5 Haiku | 80.8 | #571 |
| 7 | cogito-v1-preview-qwen-14B | 79.1 | |
| 8 | Claude Haiku 4.5 | 78.7 | #287 |
| 9 | Llama 3.1 8B | 78.6 | #1130 |
| 10 | GPT-OSS-120B | 76.9 | #346 |
| 11 | Llama 4 Scout | 74.5 | #648 |
| 12 | DeepSeek R1 Distill Qwen 14B | 74.4 | #860 |
| 13 | Llama 3.3 70B | 74.4 | #578 |
| 14 | DeepSeek R1 Distill Llama 8B | 73.5 | #1295 |
| 15 | DeepSeek R1 0528 Qwen3 8B | 73.4 | #765 |
Interactive version: theaggregate.ai/benchmark?slug=aixamine-model-data-privacy · How It Works · Data refreshed daily, snapshot 2026-11-05.