MMLU-by-task - ARC Challenge: leaderboard
The "ARC Challenge" subtask of MMLU-by-task. A single MMLU subject, reporting per-task multiple-choice accuracy from the MMLU-by-task leaderboard across a wide range of academic and professional domains.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 1257 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | StableBeluga2 | 65.96 | #1078 |
| 2 | Xwin-LM-70B-V0.1 | 65.96 | #982 |
| 3 | Llama 2 70B Chat | 64.08 | #1237 |
| 4 | Llama 2 70B Base | 62.63 | #1025 |
| 5 | WizardLM-70B-V1.0 | 61.43 | #1236 |
| 6 | Llama 2 70B Chat (HF) | 60.49 | #1264 |
| 7 | vicuna-33B-v1.3 | 59.9 | #1347 |
| 8 | falcon-40B Instruct | 58.28 | #1458 |
| 9 | MythoMax-L2-13B | 58.28 | #1409 |
| 10 | Mistral-7B-v0.1 | 56.83 | #1326 |
| 11 | OpenHermes-13B | 56.48 | #1393 |
| 12 | Llama 2 13B Chat Base | 55.63 | #1422 |
| 13 | internlm-20B | 55.38 | #1296 |
| 14 | vicuna-13B-v1.5 | 54.95 | #1392 |
| 15 | Nous-Hermes-13B | 54.95 | #1468 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-arc-challenge · How It Works · Data refreshed daily, snapshot 2026-10-20.