HELM Arabic - Arabic Mmlu - Law (Professional): leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 87.26 |
| 2 | GPT-5.4 (2026-03-05) | 86.31 |
| 3 | Claude Sonnet 4.6 | 83.76 |
| 4 | Llama 4 Maverick Instruct FP8 | 82.48 |
| 5 | Gemma 4 31B (IT) | 80.57 |
| 6 | Qwen 2.5 72B Instruct | 78.03 |
| 7 | Llama 4 Scout Instruct | 77.39 |
| 8 | GPT-4.1 (2025-04-14) | 76.75 |
| 9 | Mistral Large 3 | 76.11 |
| 10 | DeepSeek V3.1 | 74.84 |
| 11 | Gemini 2.5 Flash (Thinking) | 71.97 |
| 12 | GPT-4.1 Mini | 70.38 |
| 13 | Llama 3.3 70B Instruct | 70.06 |
| 14 | Claude Haiku 4.5 (20251001) | 69.11 |
| 15 | AceGPT-v2-32B-Chat | 67.83 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-arabic-mmlu-law-professional · How It Works · Data refreshed daily, snapshot 2026-09-19.