MMLU-by-task - Security Studies — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 2 70B Base | 79.18 |
| 2 | Llama 2 70B Chat (HF) | 78.78 |
| 3 | falcon-180B | 78.37 |
| 4 | Llama 2 70B Chat | 76.73 |
| 5 | Llama 2 70B Chat GPTQ | 75.92 |
| 6 | StableBeluga2 | 75.51 |
| 7 | Mistral-7B-v0.1 | 72.65 |
| 8 | LLaMA-65B | 71.43 |
| 9 | vicuna-33B-v1.3 | 68.98 |
| 10 | internlm-20B Chat | 68.98 |
| 11 | LLaMA-30B | 66.94 |
| 12 | WizardLM-13B-V1.2 | 66.94 |
| 13 | OpenHermes-13B | 64.9 |
| 14 | trurl-2-13B-academic | 64.49 |
| 15 | Llama 2 13B Chat Base | 64.08 |
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-security-studies · How the rankings work · Data refreshed daily, snapshot 2026-07-22.