MMLU-by-task - World Religions — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 2 70B Chat | 86.55 |
| 2 | Llama 2 70B Base | 85.38 |
| 3 | StableBeluga2 | 84.8 |
| 4 | falcon-180B | 84.21 |
| 5 | Mistral-7B-v0.1 | 83.04 |
| 6 | Llama 2 70B Chat (HF) | 81.87 |
| 7 | Llama 2 70B Chat GPTQ | 81.29 |
| 8 | LLaMA-65B | 80.7 |
| 9 | LLaMA-30B | 80.12 |
| 10 | vicuna-13B-v1.3 | 78.36 |
| 11 | falcon-40B Instruct | 77.78 |
| 12 | MythoMax-L2-13B | 77.78 |
| 13 | vicuna-13B-v1.5 | 77.78 |
| 14 | OpenHermes-13B | 77.19 |
| 15 | internlm-20B Chat | 77.19 |
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-world-religions · How the rankings work · Data refreshed daily, snapshot 2026-07-22.