HELM Arabic - Arabic Mmlu - Civics (Middle School): leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 77.54 |
| 2 | GPT-4.1 (2025-04-14) | 75 |
| 3 | Gemini 2.5 Flash (Thinking) | 74.58 |
| 4 | Claude Sonnet 4.6 | 74.15 |
| 5 | GPT-5.4 (2026-03-05) | 71.19 |
| 6 | Mistral Large 3 | 69.92 |
| 7 | Gemma 4 31B (IT) | 69.07 |
| 8 | Claude Haiku 4.5 (20251001) | 67.8 |
| 9 | Qwen 3.5 397B A17B | 66.95 |
| 10 | Llama 4 Maverick Instruct FP8 | 66.1 |
| 11 | DeepSeek V3.1 | 65.68 |
| 12 | Gemini 2.5 Flash Lite (Thinking) | 65.25 |
| 13 | Qwen 3 Next 80B A3B Instruct | 64.83 |
| 14 | AceGPT-v2-32B-Chat | 64.83 |
| 15 | GPT-4.1 Mini | 64.41 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-arabic-mmlu-civics-middle-school · How It Works · Data refreshed daily, snapshot 2026-09-19.