HELM Arabic - Mbzuai Human Translated Arabic Mmlu - College Chemistry: leaderboard

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

Top models

#ModelScore
1Claude Opus 4.769
2GPT-5.4 (2026-03-05)67
3Qwen 3.5 397B A17B66
4Claude Sonnet 4.665
5Claude Haiku 4.5 (20251001)63
6Llama 4 Maverick Instruct FP861
7Gemma 4 31B (IT)60
8DeepSeek V3.158
9Qwen 2.5 72B Instruct57
10Mistral Large 357
11Gemini 2.5 Flash (Thinking)57
12GPT-4.1 (2025-04-14)56
13Qwen 3.5 9B56
14Qwen 3 Next 80B A3B Instruct56
15Mistral Large 2 (Nov) Instruct (2411)52

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