MMLU-by-task - Abstract Algebra: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | MythoMax-L2-13B | 37 |
| 2 | Llama 2 70B Chat | 36 |
| 3 | CodeLlama-34B-Instruct-hf | 36 |
| 4 | Llama 2 70B Chat (HF) | 35 |
| 5 | LLaMA-30B | 35 |
| 6 | Phi-1.5 | 34 |
| 7 | CodeLlama-34B-hf | 34 |
| 8 | Llama 2 70B Base | 33 |
| 9 | CodeLlama-7B-Instruct-hf | 33 |
| 10 | CodeLlama-7B-hf | 33 |
| 11 | vicuna-33B-v1.3 | 33 |
| 12 | mpt-30B Instruct | 33 |
| 13 | LLaMA-13B | 32 |
| 14 | mpt-30B | 32 |
| 15 | GPT-NeoXT-Chat-Base-20B | 32 |
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-abstract-algebra · How It Works · Data refreshed daily, snapshot 2026-09-05.