SLM Trustworthiness - TrustLLM Overall: leaderboard
Metric: TrustLLM score (%; mean of ethics, privacy, robustness and fairness). Source: arxiv.org. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.2 1B Instruct | 66.11 |
| 2 | Qwen 2.5 1.5B Instruct | 64.9 |
| 3 | Gemma 1.1 7B (IT) | 63.99 |
| 4 | Llama 3.2 3B Instruct | 63.19 |
| 5 | Llama 3.1 8B Instruct | 59.52 |
| 6 | Qwen 2.5 7B Instruct | 55.17 |
| 7 | Qwen 2.5 3B Instruct | 54.85 |
| 8 | SmolLM2-360M-Instruct | 51.92 |
| 9 | Qwen 2.5 0.5B Instruct | 51.17 |
| 10 | h2o-danube3-500m-chat | 47.76 |
Interactive version: theaggregate.ai/benchmark?slug=slm-trustworthiness-trustllm-overall · How It Works · Data refreshed daily, snapshot 2026-09-19.