HELM AFR - MMLU Clinical Knowledge (Xhosa): leaderboard

Metric: EM. Source: crfm.stanford.edu. 22 models tracked.

Top models

#ModelScore
1Gemini 2.0 Flash81.89
2GPT-4o (2024-08-06)79.62
3Gemini 2.0 Flash Lite77.36
4Claude 3.7 Sonnet (20250219)73.21
5DeepSeek V370.94
6GPT-4 Turbo66.04
7Llama 3.1 405B Instruct56.6
8Llama 3.3 70B Instruct51.7
9Llama 3.1 70B Instruct51.32
10Gemma 2 27B (IT)50.57
11Claude 3.5 Haiku (20241022)47.55
12Qwen 2.5 72B Instruct45.28
13GPT-3.5 Turbo (0125)40
14Gemma 2 9B (IT)38.87
15Llama 3.1 8B Instruct35.47

Interactive version: theaggregate.ai/benchmark?slug=helm-afr-mmlu-clinical-knowledge-xhosa · How It Works · Data refreshed daily, snapshot 2026-09-08.