Qwen3-4B-Distilled-Claude-4.6-AutoRound-NVFP4-RTN

Model Details

This model is a NVFP4 (NVIDIA FP4) quantization of FlameF0X/Qwen3-4B-Distilled-Claude-4.6 generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	FlameF0X/Qwen3-4B-Distilled-Claude-4.6
Quantization Tool	AutoRound
Quantization Scheme	NVFP4
Original Size	2542 MB
Quantized Size	3985 MB

Evaluation Results

Task	Accuracy
hellaswag	0.4846
mmlu	0.6647
mmlu_abstract_algebra	0.4900
mmlu_anatomy	0.6370
mmlu_astronomy	0.7303
mmlu_business_ethics	0.6500
mmlu_clinical_knowledge	0.7283
mmlu_college_biology	0.8194
mmlu_college_chemistry	0.5100
mmlu_college_computer_science	0.6600
mmlu_college_mathematics	0.5200
mmlu_college_medicine	0.6994
mmlu_college_physics	0.5392
mmlu_computer_security	0.7900
mmlu_conceptual_physics	0.7787
mmlu_econometrics	0.6316
mmlu_electrical_engineering	0.7241
mmlu_elementary_mathematics	0.6508
mmlu_formal_logic	0.5873
mmlu_global_facts	0.3900
mmlu_high_school_biology	0.8419
mmlu_high_school_chemistry	0.6798
mmlu_high_school_computer_science	0.8000
mmlu_high_school_european_history	0.7758
mmlu_high_school_geography	0.7929
mmlu_high_school_government_and_politics	0.8446
mmlu_high_school_macroeconomics	0.7256
mmlu_high_school_mathematics	0.4704
mmlu_high_school_microeconomics	0.7941
mmlu_high_school_physics	0.5894
mmlu_high_school_psychology	0.8661
mmlu_high_school_statistics	0.6574
mmlu_high_school_us_history	0.8088
mmlu_high_school_world_history	0.8186
mmlu_human_aging	0.6996
mmlu_human_sexuality	0.7481
mmlu_humanities	0.5683
mmlu_international_law	0.7521
mmlu_jurisprudence	0.7870
mmlu_logical_fallacies	0.7975
mmlu_machine_learning	0.5089
mmlu_management	0.8252
mmlu_marketing	0.8761
mmlu_medical_genetics	0.7700
mmlu_miscellaneous	0.7867
mmlu_moral_disputes	0.6908
mmlu_moral_scenarios	0.3374
mmlu_nutrition	0.7288
mmlu_other	0.7100
mmlu_philosophy	0.6817
mmlu_prehistory	0.7006
mmlu_professional_accounting	0.5638
mmlu_professional_law	0.4505
mmlu_professional_medicine	0.7022
mmlu_professional_psychology	0.6977
mmlu_public_relations	0.6455
mmlu_security_studies	0.7184
mmlu_social_sciences	0.7667
mmlu_sociology	0.8458
mmlu_stem	0.6644
mmlu_us_foreign_policy	0.8100
mmlu_virology	0.4578
mmlu_world_religions	0.7953
piqa	0.7312

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 = "Qwen3-4B-Distilled-Claude-4.6-AutoRound-NVFP4-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 Qwen3-4B-Distilled-Claude-4.6-AutoRound-NVFP4-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.