HELM MMLU - Electrical Engineering: leaderboard
Metric: EM. Source: crfm.stanford.edu. 79 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3 | 86.9 |
| 2 | Gemini 1.5 Pro (002) | 85.52 |
| 3 | Claude 3.5 Sonnet (20241022) | 84.83 |
| 4 | Gemini 2.0 Flash (Preview) | 83.45 |
| 5 | Claude 3.5 Sonnet (20240620) | 82.76 |
| 6 | Llama 3.1 405B Instruct | 82.07 |
| 7 | Claude 3 Opus (20240229) | 81.38 |
| 8 | GPT-4o (2024-05-13) | 80.69 |
| 9 | Qwen 2.5 72B Instruct | 80 |
| 10 | Gemini 1.5 Flash (001) | 80 |
| 11 | Nova Pro | 80 |
| 12 | GPT-4o (2024-08-06) | 79.31 |
| 13 | Qwen 2 72B Instruct | 79.31 |
| 14 | Mistral Large 2 (Jul) | 79.31 |
| 15 | Jamba 1.5 Large | 79.31 |
Interactive version: theaggregate.ai/benchmark?slug=helm-mmlu-electrical-engineering · How It Works · Data refreshed daily, snapshot 2026-09-08.