LLM Trustworthy - Adversarial Demo: leaderboard
Metric: Trust Score (%). Source: huggingface.co. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o Mini (2024-07-18) | 88.49 |
| 2 | GPT-4o (2024-05-13) | 88.1 |
| 3 | GPT-3.5 Turbo (0301) | 81.28 |
| 4 | GPT-4 (0314) | 77.94 |
| 5 | Llama 3 8B Instruct | 75.54 |
| 6 | Gemini 1.0 Pro | 75.54 |
| 7 | Claude 2 | 72.97 |
| 8 | tulu-2-13B | 71.17 |
| 9 | zephyr-7B-beta | 68.68 |
| 10 | Mistral-7B-OpenOrca | 62.15 |
| 11 | tulu-2-7B | 60.49 |
| 12 | vicuna-7B-v1.3 | 57.99 |
| 13 | Llama 2 7B Chat | 55.54 |
| 14 | gemma-2B (IT) | 35.55 |
| 15 | falcon-7B Instruct | 33.95 |
Interactive version: theaggregate.ai/benchmark?slug=llm-trustworthy-adversarial-demo · How It Works · Data refreshed daily, snapshot 2026-09-05.