HELM Arabic - Mbzuai Human Translated Arabic Mmlu - Formal Logic: leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 84.92 |
| 2 | GPT-5.4 (2026-03-05) | 73.02 |
| 3 | Claude Haiku 4.5 (20251001) | 72.22 |
| 4 | Gemini 2.5 Flash (Thinking) | 72.22 |
| 5 | Claude Sonnet 4.6 | 70.63 |
| 6 | Qwen 3.5 397B A17B | 70.63 |
| 7 | Gemma 4 31B (IT) | 69.84 |
| 8 | Qwen 3 Next 80B A3B Instruct | 69.05 |
| 9 | GPT-4.1 (2025-04-14) | 66.67 |
| 10 | Mistral Large 3 | 65.08 |
| 11 | Llama 4 Maverick Instruct FP8 | 64.29 |
| 12 | GPT-4.1 Mini | 61.11 |
| 13 | Qwen 2.5 72B Instruct | 61.11 |
| 14 | DeepSeek V3.1 | 59.52 |
| 15 | Gemini 2.5 Flash Lite (Thinking) | 59.52 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-mbzuai-human-translated-arabic-mmlu-formal-logic · How It Works · Data refreshed daily, snapshot 2026-09-19.