HELM Arabic - Arabic Mmlu - Political Science (University): leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 85.24 |
| 2 | Claude Opus 4.7 | 85.24 |
| 3 | GPT-5.4 (2026-03-05) | 82.86 |
| 4 | Gemma 4 31B (IT) | 81.43 |
| 5 | Gemini 2.5 Flash (Thinking) | 80.48 |
| 6 | Claude Haiku 4.5 (20251001) | 79.05 |
| 7 | Qwen 3 Next 80B A3B Instruct | 78.1 |
| 8 | Gemini 2.5 Flash Lite (Thinking) | 77.62 |
| 9 | Llama 4 Maverick Instruct FP8 | 77.14 |
| 10 | GPT-4.1 (2025-04-14) | 76.67 |
| 11 | Qwen 3.5 397B A17B | 75.71 |
| 12 | Qwen 2.5 72B Instruct | 73.81 |
| 13 | Llama 4 Scout Instruct | 73.33 |
| 14 | Llama 3.3 70B Instruct | 72.86 |
| 15 | AceGPT-v2-70B-Chat | 71.43 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-arabic-mmlu-political-science-university · How It Works · Data refreshed daily, snapshot 2026-09-19.