HELM Arabic - Mbzuai Human Translated Arabic Mmlu - Prehistory: leaderboard
HELM Arabic (release v2.2.0) MBZUAI human-translated Arabic MMLU, Prehistory: the MMLU four-option questions for this subject translated into Arabic by human translators; 324 questions, exact match. Percent correct, 0-100, higher is better; guessing scores 25.
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 | 90.74 | #51 |
| 2 | GPT-5.4 (2026-03-05) | 88.58 | #63 |
| 3 | Claude Sonnet 4.6 | 87.35 | #94 |
| 4 | GPT-5.1 (2025-11-13) | 86.73 | #107 |
| 5 | GPT-4.1 (2025-04-14) | 86.11 | #257 |
| 6 | Gemini 2.5 Flash (Non-reasoning) | 85.8 | #255 (Gemini 2.5 Flash) |
| 7 | Gemma 4 31B (IT) | 81.79 | #178 |
| 8 | Mistral Large 3 | 81.17 | #410 |
| 9 | Qwen 3.5 397B A17B | 80.86 | #154 |
| 10 | GPT-5.4 Mini (2026-03-17) | 79.01 | #180 |
| 11 | Qwen 3 Next 80B A3B Instruct | 78.7 | #381 |
| 12 | Claude Haiku 4.5 (20251001) | 78.4 | #264 |
| 13 | Qwen 2.5 72B Instruct | 76.85 | #469 |
| 14 | Llama 4 Maverick Instruct FP8 | 76.85 | #424 |
| 15 | DeepSeek V3.1 | 75.62 | #275 |
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-mbzuai-human-translated-arabic-mmlu-prehistory · How It Works · Data refreshed daily, snapshot 2026-11-05.