Qwen3-0.6B-AutoRound-MXFP4-ModelFree

Model Details

This model is a MXFP4 (Microscaling FP4) quantization of Qwen/Qwen3-0.6B generated by agent_optimize. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Qwen/Qwen3-0.6B
Quantization Tool	agent_optimize
Quantization Scheme	MXFP4

Evaluation Results

Task	Accuracy
hellaswag	0.3762
mmlu	0.4021
mmlu_abstract_algebra	0.3100
mmlu_anatomy	0.3704
mmlu_astronomy	0.4539
mmlu_business_ethics	0.4500
mmlu_clinical_knowledge	0.3283
mmlu_college_biology	0.4792
mmlu_college_chemistry	0.3400
mmlu_college_computer_science	0.2900
mmlu_college_mathematics	0.3400
mmlu_college_medicine	0.2890
mmlu_college_physics	0.2647
mmlu_computer_security	0.6200
mmlu_conceptual_physics	0.3915
mmlu_econometrics	0.2895
mmlu_electrical_engineering	0.4207
mmlu_elementary_mathematics	0.3466
mmlu_formal_logic	0.4286
mmlu_global_facts	0.2600
mmlu_high_school_biology	0.4419
mmlu_high_school_chemistry	0.3399
mmlu_high_school_computer_science	0.4400
mmlu_high_school_european_history	0.5394
mmlu_high_school_geography	0.4697
mmlu_high_school_government_and_politics	0.5389
mmlu_high_school_macroeconomics	0.4077
mmlu_high_school_mathematics	0.2778
mmlu_high_school_microeconomics	0.4034
mmlu_high_school_physics	0.2318
mmlu_high_school_psychology	0.5651
mmlu_high_school_statistics	0.2361
mmlu_high_school_us_history	0.5147
mmlu_high_school_world_history	0.5992
mmlu_human_aging	0.4753
mmlu_human_sexuality	0.5038
mmlu_humanities	0.3654
mmlu_international_law	0.5702
mmlu_jurisprudence	0.4074
mmlu_logical_fallacies	0.4663
mmlu_machine_learning	0.3750
mmlu_management	0.5534
mmlu_marketing	0.6368
mmlu_medical_genetics	0.3700
mmlu_miscellaneous	0.4891
mmlu_moral_disputes	0.3208
mmlu_moral_scenarios	0.2380
mmlu_nutrition	0.4575
mmlu_other	0.4239
mmlu_philosophy	0.4212
mmlu_prehistory	0.4321
mmlu_professional_accounting	0.2908
mmlu_professional_law	0.2966
mmlu_professional_medicine	0.3125
mmlu_professional_psychology	0.4101
mmlu_public_relations	0.4545
mmlu_security_studies	0.4939
mmlu_social_sciences	0.4771
mmlu_sociology	0.6418
mmlu_stem	0.3622
mmlu_us_foreign_policy	0.5800
mmlu_virology	0.4217
mmlu_world_religions	0.5263
piqa	0.6763

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 = "Qwen3-0.6B-AutoRound-MXFP4-ModelFree"

# 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 Qwen3-0.6B-AutoRound-MXFP4-ModelFree \
    --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.