This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of Qwen/Qwen3.5-27B generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Qwen/Qwen3.5-27B
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	18117 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6352
mmlu	0.8408
mmlu_abstract_algebra	0.7500
mmlu_anatomy	0.8444
mmlu_astronomy	0.9605
mmlu_business_ethics	0.8400
mmlu_clinical_knowledge	0.9132
mmlu_college_biology	0.9653
mmlu_college_chemistry	0.6700
mmlu_college_computer_science	0.8200
mmlu_college_mathematics	0.6700
mmlu_college_medicine	0.8381
mmlu_college_physics	0.7647
mmlu_computer_security	0.8500
mmlu_conceptual_physics	0.9277
mmlu_econometrics	0.7281
mmlu_electrical_engineering	0.8000
mmlu_elementary_mathematics	0.8677
mmlu_formal_logic	0.7381
mmlu_global_facts	0.5600
mmlu_high_school_biology	0.9581
mmlu_high_school_chemistry	0.8424
mmlu_high_school_computer_science	0.9400
mmlu_high_school_european_history	0.8727
mmlu_high_school_geography	0.9343
mmlu_high_school_government_and_politics	1.0000
mmlu_high_school_macroeconomics	0.9308
mmlu_high_school_mathematics	0.6259
mmlu_high_school_microeconomics	0.9496
mmlu_high_school_physics	0.8212
mmlu_high_school_psychology	0.9541
mmlu_high_school_statistics	0.8611
mmlu_high_school_us_history	0.9412
mmlu_high_school_world_history	0.9241
mmlu_human_aging	0.8296
mmlu_human_sexuality	0.9237
mmlu_humanities	0.7819
mmlu_international_law	0.9421
mmlu_jurisprudence	0.9074
mmlu_logical_fallacies	0.8650
mmlu_machine_learning	0.7768
mmlu_management	0.8738
mmlu_marketing	0.9402
mmlu_medical_genetics	0.9600
mmlu_miscellaneous	0.9170
mmlu_moral_disputes	0.8179
mmlu_moral_scenarios	0.6905
mmlu_nutrition	0.9085
mmlu_other	0.8648
mmlu_philosophy	0.8328
mmlu_prehistory	0.9105
mmlu_professional_accounting	0.7624
mmlu_professional_law	0.6962
mmlu_professional_medicine	0.9669
mmlu_professional_psychology	0.8807
mmlu_public_relations	0.7545
mmlu_security_studies	0.8327
mmlu_social_sciences	0.9090
mmlu_sociology	0.9303
mmlu_stem	0.8382
mmlu_us_foreign_policy	0.9300
mmlu_virology	0.5723
mmlu_world_religions	0.9064
piqa	0.8101

How to Use

HF Usage

Step 1: Install AutoRound

pip install auto-round

Step 2: Load and run the quantized model

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen3.5-27B-AutoRound-W4A16-RTN"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")

# prepare the model input
prompt = "Write a quick sort algorithm."
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(**model_inputs, max_new_tokens=512)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]) :].tolist()

content = tokenizer.decode(output_ids, skip_special_tokens=True)
print("content:", content)

VLLM Usage

bash

vllm serve Qwen3.5-27B-AutoRound-W4A16-RTN \
    --trust-remote-code \
    --dtype bfloat16 \
    --tensor_parallel_size 1

If you encounter any issues, feel free to open an issue on the AutoRound GitHub repo or provide feedback on the Low-Bit Open LLM Leaderboard.

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs. Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Here are a couple of useful links to learn more about Intel's AI software:

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize,
  title={Optimize weight rounding via signed gradient descent for the quantization of llms},
  author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
  journal={arXiv preprint arXiv:2309.05516},
  year={2023}
}

arxiv github

This model is part of the Intel Low-Bit Open LLM Leaderboard initiative.

LeaderboardModel1/Qwen3.5-27B-AutoRound-W4A16-RTN

Qwen3.5-27B-AutoRound-W4A16-RTN

Model Details