HELM Arabic - Arabic Mmlu - Arabic Language (High School): leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 74.1 |
| 2 | Gemini 2.5 Flash (Thinking) | 69.49 |
| 3 | Claude Sonnet 4.6 | 68.21 |
| 4 | Gemma 4 31B (IT) | 64.1 |
| 5 | GPT-5.4 (2026-03-05) | 60.77 |
| 6 | Gemini 2.5 Flash Lite (Thinking) | 57.95 |
| 7 | GPT-4.1 (2025-04-14) | 56.41 |
| 8 | Mistral Large 3 | 56.15 |
| 9 | Claude Haiku 4.5 (20251001) | 55.38 |
| 10 | Qwen 3.5 397B A17B | 53.85 |
| 11 | Llama 4 Maverick Instruct FP8 | 52.82 |
| 12 | Qwen 3 Next 80B A3B Instruct | 52.56 |
| 13 | ALLaM-7B-Instruct-preview | 49.49 |
| 14 | Llama 3.3 70B Instruct | 48.97 |
| 15 | Qwen 2.5 72B Instruct | 48.72 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-arabic-mmlu-arabic-language-high-school · How It Works · Data refreshed daily, snapshot 2026-09-19.