MMLU-by-task - HellaSwag — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | falcon-180B | 70.65 |
| 2 | Llama 2 70B Chat | 68.09 |
| 3 | Llama 2 70B Base | 67.61 |
| 4 | StableBeluga2 | 67.45 |
| 5 | Llama 2 70B Chat (HF) | 66.94 |
| 6 | LLaMA-65B | 66.53 |
| 7 | Llama 2 70B Chat GPTQ | 65.57 |
| 8 | mpt-30B Instruct | 65.25 |
| 9 | falcon-40B Instruct | 64.62 |
| 10 | LLaMA-30B | 64.48 |
| 11 | MythoMax-L2-13B | 64.23 |
| 12 | vicuna-33B-v1.3 | 63.15 |
| 13 | Mistral-7B-v0.1 | 62.95 |
| 14 | WizardLM-13B-V1.2 | 62.95 |
| 15 | Llama 2 13B Chat Base | 62.94 |
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-hellaswag · How the rankings work · Data refreshed daily, snapshot 2026-07-22.