HELM Arabic - Mbzuai Human Translated Arabic Mmlu - High School Psychology: leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 95.41 |
| 2 | Claude Sonnet 4.6 | 94.5 |
| 3 | GPT-5.4 (2026-03-05) | 93.21 |
| 4 | Gemma 4 31B (IT) | 93.03 |
| 5 | GPT-4.1 (2025-04-14) | 92.48 |
| 6 | Gemini 2.5 Flash (Thinking) | 91.74 |
| 7 | Qwen 3.5 397B A17B | 91.19 |
| 8 | Mistral Large 3 | 90.83 |
| 9 | Claude Haiku 4.5 (20251001) | 89.54 |
| 10 | Llama 4 Maverick Instruct FP8 | 88.99 |
| 11 | GPT-4.1 Mini | 88.44 |
| 12 | Gemini 2.5 Flash Lite (Thinking) | 87.16 |
| 13 | DeepSeek V3.1 | 86.97 |
| 14 | Qwen 2.5 72B Instruct | 86.24 |
| 15 | Qwen 3 Next 80B A3B Instruct | 85.87 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-mbzuai-human-translated-arabic-mmlu-high-school-psychology · How It Works · Data refreshed daily, snapshot 2026-09-19.