HELM Arabic - Mbzuai Human Translated Arabic Mmlu - Anatomy: leaderboard

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

Top models

#ModelScore
1Claude Opus 4.783.7
2GPT-5.4 (2026-03-05)81.48
3Claude Sonnet 4.680.74
4GPT-4.1 (2025-04-14)77.04
5Gemma 4 31B (IT)73.33
6Qwen 3.5 397B A17B73.33
7Llama 4 Maverick Instruct FP871.85
8Gemini 2.5 Flash (Thinking)71.85
9Claude Haiku 4.5 (20251001)68.15
10Mistral Large 366.67
11Gemini 2.5 Flash Lite (Thinking)65.93
12GPT-4.1 Mini64.44
13Qwen 3 Next 80B A3B Instruct62.96
14DeepSeek V3.162.22
15Llama 4 Scout Instruct61.48

Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-mbzuai-human-translated-arabic-mmlu-anatomy · How It Works · Data refreshed daily, snapshot 2026-09-19.