LLM Trustworthy - Privacy: leaderboard
Metric: Trust Score (%). Source: huggingface.co. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 2 7B Chat | 97.39 |
| 2 | GPT-4o (2024-05-13) | 97.04 |
| 3 | GPT-4o Mini (2024-07-18) | 89.38 |
| 4 | gemma-2B (IT) | 88.77 |
| 5 | Claude 2 | 85.35 |
| 6 | zephyr-7B-beta | 84.18 |
| 7 | gemma-7B (IT) | 83.69 |
| 8 | Llama 3 8B Instruct | 81.59 |
| 9 | Gemini 1.0 Pro | 81.59 |
| 10 | tulu-2-13B | 78.9 |
| 11 | Mistral-7B-OpenOrca | 77.36 |
| 12 | tulu-2-7B | 75.82 |
| 13 | vicuna-7B-v1.3 | 72.96 |
| 14 | falcon-7B Instruct | 70.26 |
| 15 | GPT-3.5 Turbo (0301) | 70.13 |
Interactive version: theaggregate.ai/benchmark?slug=llm-trustworthy-privacy · How It Works · Data refreshed daily, snapshot 2026-09-05.