MMLU-by-task - Moral Scenarios — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | StableBeluga2 | 60.78 |
| 2 | falcon-180B | 53.3 |
| 3 | Llama 2 70B Chat | 52.51 |
| 4 | LLaMA-65B | 48.04 |
| 5 | Llama 2 70B Base | 45.47 |
| 6 | OpenHermes-13B | 45.14 |
| 7 | MythoMax-L2-13B | 43.91 |
| 8 | vicuna-33B-v1.3 | 43.58 |
| 9 | internlm-20B Chat | 39.89 |
| 10 | Llama 2 70B Chat (HF) | 39.55 |
| 11 | LLaMA-30B | 38.99 |
| 12 | WizardCoder-Python-34B-V1.0 | 38.1 |
| 13 | trurl-2-13B-academic | 37.65 |
| 14 | Llama 2 70B Chat GPTQ | 32.96 |
| 15 | Mistral-7B-v0.1 | 32.51 |
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-moral-scenarios · How the rankings work · Data refreshed daily, snapshot 2026-07-22.