MMLU (CoT)
Chain-of-Thought variant of the Massive Multitask Language Understanding benchmark, evaluating language models across 57 tasks including elementary mathematics, US history, computer science, law, and other professional and academic subjects. This version uses chain-of-thought prompting to elicit step-by-step reasoning.
3rows
scoreprimary metric
2026-05-06sampled
Metadata
Metrics
Score, Normalized Score
| Rank | Subject | Score | Model Match | Provenance | Sampled |
|---|---|---|---|---|---|
| 1 | Llama 3.1 405B Instruct | 0.89 | — | Self-reported | 2026-05-06 |
| 2 | Llama 3.1 70B Instruct | 0.86 |  Llama 3.1 70B Instruct meta-llama-llama-3.1-70b-instruct | Self-reported | 2026-05-06 |
| 3 | Llama 3.1 8B Instruct | 0.73 | Llama 3.1 8B Instruct meta-llama-llama-3.1-8b-instruct | Self-reported | 2026-05-06 |
No matching rows.