license, datasets, language, base_model, pipeline_tag

license	datasets	language	base_model	pipeline_tag
cc-by-nc-4.0	hammh0a/Hala-4.6M-SFT	ar	QCRI/Fanar-1-9B-Instruct	text-generation

Hala: Arabic‑Centric Instruction & Translation Models

Paper: Hala Technical Report: Building Arabic‑Centric Instruction & Translation Models at Scale

Authors: Hasan Abed Al Kader Hammoud*, Mohammad Zbeeb*, Bernard Ghanem

Affiliation: King Abdullah University of Science and Technology (KAUST)

*Equal contribution

In Arabic, حلا (Hala) conveys sweetness and beauty—qualities long associated with the language itself. In this spirit, we call our models Hala.

---

🔗 Quick Links

• Models & Data (Hugging Face collection): https://huggingface.co/collections/hammh0a/hala-68bf02b34a14b9f22305ab3a
• Contact: hasanabedalkader.hammoud@kaust.edu.sa

---

Example

from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline

model_id = "hammh0a/Hala-9B"  # pick a released Hala model

tok = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id, torch_dtype="auto", device_map="auto"
)

# Use chat template
messages = [
    {"role": "system", "content": "أنت مساعد خبير في الفيزياء."},
    {"role": "user", "content": "اشرح بإيجاز مبدأ الانحفاظ في الفيزياء، وأعطني مثالاً يومياً."},
]

prompt = tok.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

pipe = pipeline("text-generation", model=model, tokenizer=tok)
out = pipe(prompt, max_new_tokens=256, do_sample=False)

print(out[0]["generated_text"])

---

📊 Results

Hala models are placed at the end of each size category; best Average per category is in bold.

≤2B parameters

Size	Model Name	Params	AlGhafa	ArabicMMLU	EXAMS	MadinahQA	AraTrust	ArbMMLU‑HT	Average
≤2B	meta-llama/Llama-3.2-1B	1B	33.9	26.5	21.2	25.7	37.1	23.9	28.0
≤2B	Qwen/Qwen2-1.5B-Instruct	1.5B	53.1	49.2	35.2	45.5	68.9	37.4	48.2
≤2B	Qwen/Qwen2.5-1.5B-Instruct	1.5B	48.4	43.5	31.8	38.2	70.8	35.9	44.8
≤2B	Sakalti/Saka-1.5B	1.5B	51.4	40.0	31.3	31.5	47.5	33.5	39.2
≤2B	Qwen/Qwen3-1.7B-Base	1.7B	56.8	49.7	38.2	40.0	75.6	43.9	50.7
≤2B	Qwen/Qwen1.5-1.8B	1.8B	32.7	26.7	23.8	26.0	31.5	23.6	27.4
≤2B	silma-ai/SILMA-Kashif-2B-Instruct-v1.0	2B	59.7	45.6	33.1	38.8	73.3	35.8	47.7
≤2B	google/gemma-2-2b-it	2B	34.1	30.1	23.6	20.1	31.2	23.4	27.1
≤2B	LiquidAI/LFM2-350M	350M	39.0	35.2	30.9	28.3	43.3	29.1	34.3
≤2B	Hala‑350M	350M	51.4	41.2	36.9	34.5	52.1	35.4	41.9
≤2B	LiquidAI/LFM2-700M	700M	50.1	38.3	34.3	32.5	56.3	37.2	41.4
≤2B	Hala‑700M	700M	55.5	45.9	40.6	34.7	65.2	39.4	46.9
≤2B	LiquidAI/LFM2-1.2B	1.2B	53.8	45.2	35.0	34.7	65.6	43.4	46.3
≤2B	Hala‑1.2B	1.2B	59.2	48.6	43.4	41.6	71.7	44.2	51.4

7B–9B parameters

Size	Model Name	Params	AlGhafa	ArabicMMLU	EXAMS	MadinahQA	AraTrust	ArbMMLU‑HT	Average
7B–9B	CohereForAI/c4ai-command-r7b-arabic-02-2025	7B	74.8	59.3	65.0	63.8	80.5	50.1	65.6
7B–9B	JasperV13/Yehia-7B-DPO-Reasoning-preview	7B	75.1	66.3	51.8	54.9	81.9	55.1	64.2
7B–9B	Navid-AI/Yehia-7B-preview	7B	70.8	64.9	52.1	54.4	87.5	53.4	63.9
7B–9B	JasperV13/Yehia-7B-Reasoning-preview	7B	75.2	66.3	52.7	55.0	80.8	55.2	64.2
7B–9B	ALLaM-AI/ALLaM-7B-Instruct-preview	7B	69.5	64.9	51.6	54.2	86.9	52.8	63.3
7B–9B	Qwen/Qwen2-7B-Instruct	7B	73.2	60.0	47.3	59.5	82.8	51.3	62.4
7B–9B	Qwen/Qwen3-8B-Base	8B	74.8	65.0	52.5	52.2	83.4	61.5	64.9
7B–9B	QCRI/Fanar-1-9B-Instruct	9B	76.4	65.8	52.7	73.3	88.3	58.6	69.2
7B–9B	Hala‑9B	9B	78.3	65.6	53.8	70.4	89.6	61.4	69.9

Evaluation protocol: lighteval on ArabicMMLU (OALL‑2) excluding AlRage.

---

📚 Citation

If you find Hala useful, please cite:

@misc{hammoud2025halatechnicalreportbuilding,
      title={Hala Technical Report: Building Arabic-Centric Instruction & Translation Models at Scale}, 
      author={Hasan Abed Al Kader Hammoud and Mohammad Zbeeb and Bernard Ghanem},
      year={2025},
      url={https://arxiv.org/abs/2509.14008}, 
}