HELM AFR - MMLU Clinical Knowledge: leaderboard
Metric: EM. Source: crfm.stanford.edu. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.0 Flash | 77.53 |
| 2 | Claude 3.7 Sonnet (20250219) | 71.6 |
| 3 | Gemini 2.0 Flash Lite | 71.05 |
| 4 | GPT-4o (2024-08-06) | 68.3 |
| 5 | DeepSeek V3 | 67.24 |
| 6 | GPT-4 Turbo | 56.02 |
| 7 | Llama 3.1 70B Instruct | 55.47 |
| 8 | Llama 3.1 405B Instruct | 52.62 |
| 9 | Claude 3.5 Haiku (20241022) | 50.77 |
| 10 | Gemma 2 27B (IT) | 48.03 |
| 11 | Llama 3.3 70B Instruct | 47.62 |
| 12 | Qwen 2.5 72B Instruct | 47.58 |
| 13 | GPT-3.5 Turbo (0125) | 40.27 |
| 14 | Llama 3.1 8B Instruct | 38.18 |
| 15 | Gemma 2 9B (IT) | 37.22 |
Interactive version: theaggregate.ai/benchmark?slug=helm-afr-mmlu-clinical-knowledge · How It Works · Data refreshed daily, snapshot 2026-09-08.