Open Arabic LLM - Alghafa Multiple Choice Sentiment Task Leaderboard — The Aggregate

Open Arabic LLM - Alghafa Multiple Choice Sentiment Task — leaderboard

Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.

Top models

#	Model	Score
1	Rombos-LLM-V2.5-Qwen-72B	43.78
2	Qwen 2.5 72B	43.66
3	calme-2.2-qwen2.5-72B	43.66
4	Ultiima-72B	43.55
5	calme-2.1-qwen2.5-72B	43.43
6	Qwen 2.5 72B Instruct	43.02
7	Saka-14B	43.02
8	Qwen 2 72B	43.02
9	SeaLLMs-v3-7B-Chat	42.44
10	Gemma 3 27B (IT)	42.33
11	Qwen 2.5 14B	42.27
12	Qwen 2.5 32B	41.63
13	aya-expanse-32B	41.57
14	Llama 3.1 8B Instruct	41.51
15	Gemma 2 27B (IT)	41.51

Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-alghafa-multiple-choice-sentiment-task · How the rankings work · Data refreshed daily, snapshot 2026-07-22.