HELM MMLU - High School Statistics: leaderboard
The "High School Statistics" subtask of HELM MMLU. HELM Classic MMLU group leaderboard from the public HELM release artifacts.
Metric: EM. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 79 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 1.5 Pro (002) | 88.43 | #350 |
| 2 | DeepSeek V3 | 86.11 | #323 |
| 3 | Gemini 1.5 Pro (001) | 85.65 | #446 |
| 4 | Claude 3.5 Sonnet (20241022) | 85.19 | #291 |
| 5 | Claude 3.5 Sonnet (20240620) | 84.72 | #298 |
| 6 | Llama 3.1 405B Instruct | 83.8 | #461 |
| 7 | Gemini 2.0 Flash (Preview) | 80.09 | #366 |
| 8 | Qwen 2.5 72B Instruct | 79.63 | #458 |
| 9 | Claude 3 Opus (20240229) | 79.63 | #413 |
| 10 | GPT-4 (0613) | 79.17 | #494 |
| 11 | GPT-4o (2024-05-13) | 78.7 | #326 |
| 12 | Nova Pro | 77.78 | #573 |
| 13 | Llama 3.3 70B Instruct | 76.85 | #536 |
| 14 | GPT-4o (2024-08-06) | 76.85 | #335 |
| 15 | Llama 3.2 90B Vision Instruct | 76.39 | #626 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-mmlu-high-school-statistics · How It Works · Data refreshed daily, snapshot 2026-10-23.