HELM Arabic - Arabic Mmlu - Natural Science (Primary School): leaderboard
HELM Arabic (release v2.2.0) ArabicMMLU Natural Science (Primary School): native Arabic multiple-choice questions from ArabicMMLU, which draws on school, university and professional exams across Arabic-speaking countries; 336 questions, exact match. Percent correct, 0-100, higher is better.
Metric: Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 41 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 94.94 | #51 |
| 2 | GPT-5.4 (2026-03-05) | 94.64 | #63 |
| 3 | Qwen 3.5 397B A17B | 93.75 | #154 |
| 4 | Llama 4 Maverick Instruct FP8 | 93.75 | #424 |
| 5 | Claude Sonnet 4.6 | 93.45 | #94 |
| 6 | GPT-5.1 (2025-11-13) | 93.45 | #107 |
| 7 | Gemini 2.5 Flash Lite (Non-reasoning) | 93.45 | #432 (Gemini 2.5 Flash Lite) |
| 8 | GPT-4.1 (2025-04-14) | 93.15 | #257 |
| 9 | Mistral Large 3 | 93.15 | #410 |
| 10 | Gemma 4 31B (IT) | 92.86 | #178 |
| 11 | Gemini 2.5 Flash (Non-reasoning) | 92.86 | #255 (Gemini 2.5 Flash) |
| 12 | GPT-5.4 Mini (2026-03-17) | 92.86 | #180 |
| 13 | DeepSeek V3.1 | 92.26 | #275 |
| 14 | GPT-4.1 Mini | 91.96 | #363 |
| 15 | Qwen 2.5 72B Instruct | 91.67 | #469 |
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-natural-science-primary-school · How It Works · Data refreshed daily, snapshot 2026-11-05.