HELM Arabic - Arabic Mmlu - Computer Science (University): leaderboard
HELM Arabic (release v2.2.0) ArabicMMLU Computer Science (University): native Arabic multiple-choice questions from ArabicMMLU, which draws on school, university and professional exams across Arabic-speaking countries; 64 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 | GPT-5.1 (2025-11-13) | 89.06 | #107 |
| 2 | Claude Opus 4.7 | 85.94 | #51 |
| 3 | GPT-4.1 (2025-04-14) | 85.94 | #257 |
| 4 | Claude Sonnet 4.6 | 84.38 | #94 |
| 5 | Gemma 4 31B (IT) | 84.38 | #178 |
| 6 | GPT-5.4 (2026-03-05) | 84.38 | #63 |
| 7 | Qwen 3.5 397B A17B | 82.81 | #154 |
| 8 | Llama 4 Scout Instruct | 82.81 | #547 |
| 9 | Gemini 2.5 Flash (Non-reasoning) | 82.81 | #255 (Gemini 2.5 Flash) |
| 10 | Llama 3.3 70B Instruct | 81.25 | #541 |
| 11 | Qwen 3 Next 80B A3B Instruct | 81.25 | #381 |
| 12 | Mistral Large 3 | 81.25 | #410 |
| 13 | GPT-5.4 Mini (2026-03-17) | 81.25 | #180 |
| 14 | GPT-4.1 Mini | 79.69 | #363 |
| 15 | AceGPT-v2-70B-Chat | 79.69 | #421 |
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-computer-science-university · How It Works · Data refreshed daily, snapshot 2026-11-05.