HELM AFR - MMLU Clinical Knowledge (Xhosa): leaderboard
Metric: EM. Source: crfm.stanford.edu. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.0 Flash | 81.89 |
| 2 | GPT-4o (2024-08-06) | 79.62 |
| 3 | Gemini 2.0 Flash Lite | 77.36 |
| 4 | Claude 3.7 Sonnet (20250219) | 73.21 |
| 5 | DeepSeek V3 | 70.94 |
| 6 | GPT-4 Turbo | 66.04 |
| 7 | Llama 3.1 405B Instruct | 56.6 |
| 8 | Llama 3.3 70B Instruct | 51.7 |
| 9 | Llama 3.1 70B Instruct | 51.32 |
| 10 | Gemma 2 27B (IT) | 50.57 |
| 11 | Claude 3.5 Haiku (20241022) | 47.55 |
| 12 | Qwen 2.5 72B Instruct | 45.28 |
| 13 | GPT-3.5 Turbo (0125) | 40 |
| 14 | Gemma 2 9B (IT) | 38.87 |
| 15 | Llama 3.1 8B Instruct | 35.47 |
Interactive version: theaggregate.ai/benchmark?slug=helm-afr-mmlu-clinical-knowledge-xhosa · How It Works · Data refreshed daily, snapshot 2026-09-08.