HELM Arabic - Arabic Mmlu - Arabic Language (Primary School): leaderboard

Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.

Top models

#ModelScore
1Claude Opus 4.790.48
2Gemini 2.5 Flash (Thinking)89.68
3Claude Sonnet 4.688.49
4GPT-4.1 (2025-04-14)88.49
5Gemma 4 31B (IT)86.9
6GPT-5.4 (2026-03-05)86.9
7GPT-4.1 Mini83.73
8Claude Haiku 4.5 (20251001)83.73
9Llama 4 Scout Instruct83.33
10Mistral Large 383.33
11Qwen 3.5 397B A17B82.94
12Gemini 2.5 Flash Lite (Thinking)82.94
13Llama 4 Maverick Instruct FP881.75
14Qwen 3 Next 80B A3B Instruct80.16
15Qwen 2.5 72B Instruct79.76

Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-arabic-mmlu-arabic-language-primary-school · How It Works · Data refreshed daily, snapshot 2026-09-19.