Qwen3-32B-AutoRound-MXFP8-RTN

Model Details

This model is a MXFP8 quantization of Qwen/Qwen3-32B generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Qwen/Qwen3-32B
Quantization Tool	AutoRound
Quantization Scheme	MXFP8
Quantized Size	33659 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6402
mmlu	0.8052
mmlu_abstract_algebra	0.6900
mmlu_anatomy	0.7704
mmlu_astronomy	0.9276
mmlu_business_ethics	0.7600
mmlu_clinical_knowledge	0.8528
mmlu_college_biology	0.9306
mmlu_college_chemistry	0.6700
mmlu_college_computer_science	0.7900
mmlu_college_mathematics	0.6300
mmlu_college_medicine	0.8150
mmlu_college_physics	0.7647
mmlu_computer_security	0.8700
mmlu_conceptual_physics	0.9234
mmlu_econometrics	0.7719
mmlu_electrical_engineering	0.8552
mmlu_elementary_mathematics	0.8545
mmlu_formal_logic	0.7222
mmlu_global_facts	0.5800
mmlu_high_school_biology	0.9613
mmlu_high_school_chemistry	0.8227
mmlu_high_school_computer_science	0.9000
mmlu_high_school_european_history	0.8424
mmlu_high_school_geography	0.9192
mmlu_high_school_government_and_politics	0.9741
mmlu_high_school_macroeconomics	0.8949
mmlu_high_school_mathematics	0.6407
mmlu_high_school_microeconomics	0.9370
mmlu_high_school_physics	0.8013
mmlu_high_school_psychology	0.9376
mmlu_high_school_statistics	0.8333
mmlu_high_school_us_history	0.9167
mmlu_high_school_world_history	0.9072
mmlu_human_aging	0.8251
mmlu_human_sexuality	0.9389
mmlu_humanities	0.7180
mmlu_international_law	0.8512
mmlu_jurisprudence	0.8611
mmlu_logical_fallacies	0.8896
mmlu_machine_learning	0.7321
mmlu_management	0.8641
mmlu_marketing	0.9145
mmlu_medical_genetics	0.9300
mmlu_miscellaneous	0.9093
mmlu_moral_disputes	0.8006
mmlu_moral_scenarios	0.6056
mmlu_nutrition	0.8693
mmlu_other	0.8333
mmlu_philosophy	0.8039
mmlu_prehistory	0.8920
mmlu_professional_accounting	0.6950
mmlu_professional_law	0.5854
mmlu_professional_medicine	0.8897
mmlu_professional_psychology	0.8399
mmlu_public_relations	0.7545
mmlu_security_studies	0.8245
mmlu_social_sciences	0.8911
mmlu_sociology	0.9254
mmlu_stem	0.8237
mmlu_us_foreign_policy	0.9300
mmlu_virology	0.5542
mmlu_world_religions	0.8713
piqa	0.8090

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 = "Qwen3-32B-AutoRound-MXFP8-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 Qwen3-32B-AutoRound-MXFP8-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:

• Intel Neural Compressor
• AutoRound

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.