Qwen3-8B-AutoRound-MXFP4-RTN

Model Details

This model is a MXFP4 (Microscaling FP4) quantization of Qwen/Qwen3-8B generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Qwen/Qwen3-8B
Quantization Tool	AutoRound
Quantization Scheme	MXFP4
Quantized Size	8009 MB

Evaluation Results

Task	Accuracy
hellaswag	0.5398
mmlu	0.6962
mmlu_abstract_algebra	0.5400
mmlu_anatomy	0.6370
mmlu_astronomy	0.8487
mmlu_business_ethics	0.7200
mmlu_clinical_knowledge	0.7585
mmlu_college_biology	0.8264
mmlu_college_chemistry	0.5400
mmlu_college_computer_science	0.6600
mmlu_college_mathematics	0.5500
mmlu_college_medicine	0.7225
mmlu_college_physics	0.5588
mmlu_computer_security	0.7900
mmlu_conceptual_physics	0.8213
mmlu_econometrics	0.6316
mmlu_electrical_engineering	0.7448
mmlu_elementary_mathematics	0.6984
mmlu_formal_logic	0.6032
mmlu_global_facts	0.3800
mmlu_high_school_biology	0.8806
mmlu_high_school_chemistry	0.6995
mmlu_high_school_computer_science	0.8700
mmlu_high_school_european_history	0.8424
mmlu_high_school_geography	0.8434
mmlu_high_school_government_and_politics	0.9016
mmlu_high_school_macroeconomics	0.7513
mmlu_high_school_mathematics	0.4741
mmlu_high_school_microeconomics	0.8697
mmlu_high_school_physics	0.6358
mmlu_high_school_psychology	0.8789
mmlu_high_school_statistics	0.6898
mmlu_high_school_us_history	0.8137
mmlu_high_school_world_history	0.8692
mmlu_human_aging	0.6816
mmlu_human_sexuality	0.8168
mmlu_humanities	0.6028
mmlu_international_law	0.8264
mmlu_jurisprudence	0.8148
mmlu_logical_fallacies	0.8344
mmlu_machine_learning	0.5536
mmlu_management	0.9029
mmlu_marketing	0.8889
mmlu_medical_genetics	0.8100
mmlu_miscellaneous	0.8148
mmlu_moral_disputes	0.7225
mmlu_moral_scenarios	0.3654
mmlu_nutrition	0.7745
mmlu_other	0.7351
mmlu_philosophy	0.7138
mmlu_prehistory	0.7932
mmlu_professional_accounting	0.5390
mmlu_professional_law	0.4791
mmlu_professional_medicine	0.7316
mmlu_professional_psychology	0.7222
mmlu_public_relations	0.6909
mmlu_security_studies	0.7837
mmlu_social_sciences	0.7979
mmlu_sociology	0.8010
mmlu_stem	0.6981
mmlu_us_foreign_policy	0.8500
mmlu_virology	0.5301
mmlu_world_religions	0.7836
piqa	0.7356

How to Use

HF Usage

Step 1: Install AutoRound

pip install auto-round

Step 2: Load and run the quantized model

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen3-8B-AutoRound-MXFP4-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

vllm serve Qwen3-8B-AutoRound-MXFP4-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.