HELM Arabic - Alghafa - MCQ Exams Test AR: leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.3 70B Instruct | 64.99 |
| 2 | Claude Opus 4.7 | 62.66 |
| 3 | GPT-4.1 (2025-04-14) | 61.58 |
| 4 | Claude Sonnet 4.6 | 61.4 |
| 5 | GPT-5.4 (2026-03-05) | 61.4 |
| 6 | Llama 4 Maverick Instruct FP8 | 60.86 |
| 7 | Gemini 2.5 Flash (Thinking) | 60.32 |
| 8 | Llama 4 Scout Instruct | 59.96 |
| 9 | Qwen 3.5 397B A17B | 59.78 |
| 10 | Mistral Large 3 | 59.78 |
| 11 | Gemini 2.5 Flash Lite (Thinking) | 59.61 |
| 12 | Qwen 2.5 72B Instruct | 59.43 |
| 13 | Gemma 4 31B (IT) | 59.25 |
| 14 | Claude Haiku 4.5 (20251001) | 59.07 |
| 15 | Qwen 3 Next 80B A3B Instruct | 58.89 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-alghafa-mcq-exams-test-ar · How It Works · Data refreshed daily, snapshot 2026-09-19.