HELM Arabic - Mbzuai Human Translated Arabic Mmlu - Global Facts Leaderboard · The Aggregate

HELM Arabic - Mbzuai Human Translated Arabic Mmlu - Global Facts: leaderboard

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

Top models

#	Model	Score
1	Claude Opus 4.7	71
2	Claude Sonnet 4.6	67
3	Gemma 4 31B (IT)	67
4	Mistral Large 3	67
5	DeepSeek V3.1	66
6	Mistral Large 2 (Nov) Instruct (2411)	65
7	GPT-4.1 (2025-04-14)	62
8	Gemini 2.5 Flash (Thinking)	62
9	GPT-5.4 (2026-03-05)	60
10	Llama 3.3 70B Instruct	59
11	Qwen 3.5 397B A17B	59
12	Qwen 3 Next 80B A3B Instruct	58
13	Claude Haiku 4.5 (20251001)	55
14	Llama 4 Maverick Instruct FP8	55
15	Qwen 2.5 72B Instruct	54

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