HELM Arabic - Arabic Mmlu - Law (Professional): leaderboard
HELM Arabic (release v2.2.0) ArabicMMLU Law (Professional): native Arabic multiple-choice questions from ArabicMMLU, which draws on school, university and professional exams across Arabic-speaking countries; 314 questions, exact match. Percent correct, 0-100, higher is better.
Metric: Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Around February 2027. 41 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 87.26 | #51 |
| 2 | GPT-5.4 (2026-03-05) | 86.31 | #63 |
| 3 | Claude Sonnet 4.6 | 83.76 | #94 |
| 4 | Llama 4 Maverick Instruct FP8 | 82.48 | #424 |
| 5 | Gemma 4 31B (IT) | 80.57 | #178 |
| 6 | GPT-5.1 (2025-11-13) | 80.57 | #107 |
| 7 | GPT-5.4 Mini (2026-03-17) | 79.62 | #180 |
| 8 | Qwen 2.5 72B Instruct | 78.03 | #469 |
| 9 | Llama 4 Scout Instruct | 77.39 | #547 |
| 10 | GPT-4.1 (2025-04-14) | 76.75 | #257 |
| 11 | Mistral Large 3 | 76.11 | #410 |
| 12 | DeepSeek V3.1 | 74.84 | #275 |
| 13 | Gemini 2.5 Flash (Non-reasoning) | 71.97 | #255 (Gemini 2.5 Flash) |
| 14 | GPT-4.1 Mini | 70.38 | #363 |
| 15 | Llama 3.3 70B Instruct | 70.06 | #541 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-arabic-mmlu-law-professional · How It Works · Data refreshed daily, snapshot 2026-11-05.