Qwen3.6 35B A3B NVFP4
简介
Qwen3.6 35B A3B NVFP4 是由 NVIDIA 提供的基于通义千问系列的优化版本。该模型采用了 NVFP4 低精度量化技术，旨在通过大幅降低显存占用，在不牺牲太多性能的前提下，提升在 NVIDIA GPU 上的推理吞吐量。对于开发者而言，它在保持 30B 级别模型逻辑推理能力的同时，极大降低了部署门槛，非常适合需要兼顾响应速度与智能程度的本地化部署或企业级私有化场景，是追求极致推理能效比的理想选择。
核心亮点
- NVFP4 量化大幅降低显存占用，提升推理速度
- 继承 Qwen 系列强悍的中文理解与逻辑推理能力
- 极低部署门槛，适合在 NVIDIA 显卡上高效运行
- Apache-2.0 协议，支持灵活的商业化部署
使用方法
安装依赖
# 安装 Hugging Face transformers
pip install transformers torchSDK 使用
# 使用 transformers 加载模型
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("nvidia/Qwen3.6-35B-A3B-NVFP4")
tokenizer = AutoTokenizer.from_pretrained("nvidia/Qwen3.6-35B-A3B-NVFP4")Hugging Face 下载
我们推荐使用命令行或者 Hugging Face Hub SDK 来进行模型的下载。
操作指引：在下载前，请先通过如下命令安装 huggingface_hub：
操作指引
pip install -U huggingface_hub命令行下载
下载完整模型库
下载完整模型库
huggingface-cli download nvidia/Qwen3.6-35B-A3B-NVFP4下载单个文件到指定本地文件夹（以下载 config.json 到当前路径下 ./dir 目录为例）
下载单个文件到指定本地文件夹（以下载 config.json 到当前路径下 ./dir 目录为例）
huggingface-cli download nvidia/Qwen3.6-35B-A3B-NVFP4 config.json --local-dir ./dirSDK 下载
SDK 下载
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('nvidia/Qwen3.6-35B-A3B-NVFP4')Git 下载
请确保 lfs 已经被正确安装
Git 下载
git lfs install
git clone https://huggingface.co/nvidia/Qwen3.6-35B-A3B-NVFP4如果您希望跳过 lfs 大文件下载，可以使用如下命令
跳过 LFS
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/nvidia/Qwen3.6-35B-A3B-NVFP4模型文件托管在 Hugging Face Hub，使用 HF CLI / SDK / Git 直接下载，不经过本站。
PyTorch / Transformers 使用
安装 Transformers
安装 Transformers
pip install -U transformers torch模型加载和推理
模型加载和推理
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('nvidia/Qwen3.6-35B-A3B-NVFP4')
tokenizer = AutoTokenizer.from_pretrained('nvidia/Qwen3.6-35B-A3B-NVFP4')完整文档
来源: HuggingFace
---
pipeline_tag: text-generation
base_model:
- Qwen/Qwen3.6-35B-A3B
license: apache-2.0
library_name: Model Optimizer
tags:
- nvidia
- ModelOpt
- Qwen3.6
- quantized
- FP4
- fp4
---
Model Overview
Description:
The NVIDIA Qwen3.6-35B-A3B-NVFP4 model is the quantized version of Alibaba's Qwen3.6-35B-A3B model, which is an auto-regressive language model that uses an optimized transformer architecture. For more information, please check here. The NVIDIA Qwen3.6-35B-A3B-NVFP4 model is quantized with Model Optimizer.This model is ready for commercial/non-commercial use. <br>
Third-Party Community Consideration
This model is not owned or developed by NVIDIA. This model has been developed and built to a third-party’s requirements for this application and use case; see link to Non-NVIDIA (Qwen3.6-35B-A3B) Model Card from Alibaba.References
NVIDIA Model Optimizer: https://github.com/NVIDIA/Model-OptimizerLicense/Terms of Use:
Apache license 2.0Deployment Geography:
Global <br>Use Case: <br>
Developers looking to take off-the-shelf, pre-quantized models for deployment in AI Agent systems, chatbots, RAG systems, and other AI-powered applications. <br>Release Date: <br>
Hugging Face on 05/28/2026 via https://huggingface.co/nvidia/Qwen3.6-35B-A3B-NVFP4 <br>Model Architecture:
Architecture Type: Transformers <br> Network Architecture: Mixture-of-Experts (MoE) with Hybrid Attention <br> Number of Model Parameters: 35B in total and 3B activated <br>Input:
Input Type(s): Text, Image, Video <br> Input Format(s): String, Red, Green, Blue (RGB), Video (MP4/WebM) <br> Input Parameters: One-Dimensional (1D), Two-Dimensional (2D), Three-Dimensional (3D) <br> Other Properties Related to Input: Context length up to 262K <br>Output:
Output Type(s): Text <br> Output Format: String <br> Output Parameters: One-Dimensional(1D): Sequences <br> Other Properties Related to Output: None <br>Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA’s hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions. <br>
Software Integration:
Supported Runtime Engine(s): <br>- vLLM <br>
Supported Hardware Microarchitecture Compatibility: <br>
- NVIDIA Hopper, NVIDIA Blackwell <br>
Preferred Operating System(s): <br>
- Linux <br>
The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.
Model Version(s):
The model version is NVFP4 1.0 version and is Quantized with nvidia-modelopt v0.44.0 <br>Training and Evaluation Datasets:
Calibration Dataset:
Link: cnn_dailymail, Nemotron-Post-Training-Dataset-v2 <br> Data Collection Method by dataset: Automated. <br> Labeling Method by dataset: Automated. <br> Properties: The cnn_dailymail dataset is an English-language dataset containing just over 300k unique news articles as written by journalists at CNN and the Daily Mail. The Nemotron-Post-Training-Dataset-v2 is a post-training dataset curated by NVIDIA containing multi-turn conversations across diverse topics. <br>Training Dataset:
Data Modality: Undisclosed <br> Data Collection Method by dataset: Undisclosed <br> Labeling Method by dataset: Undisclosed <br> Data Size: Undisclosed <br> Properties: UndisclosedEvaluation Dataset:
Datasets: MMLU Pro, GPQA Diamond, τ²-Bench Telecom, MMMU Pro, SciCode, AIME 2025, AA-LCR, IFBench <br> Data Collection Method by dataset: Hybrid: Automated, Human <br> Labeling Method by dataset: Hybrid: Human, Automated <br> Properties: We evaluated the model on text-based reasoning and coding benchmarks: MMLU Pro is a multi-task language understanding benchmark with challenging multiple-choice questions across diverse academic domains; GPQA Diamond contains 448 graduate-level multiple-choice questions written by domain experts in biology, physics, and chemistry; τ²-Bench Telecom evaluates agentic tool-use and policy-adherence capabilities in dual-control telecom customer-service scenarios where the model interacts with a simulated user and external tools to resolve account issues; MMMU Pro is the more challenging version of the Massive Multi-discipline Multimodal Understanding benchmark, measuring college-level multimodal reasoning across diverse disciplines with expanded answer choices and a vision-only input setting; SciCode evaluates scientific coding capabilities; AIME 2025 contains problems from the American Invitational Mathematics Examination; AA-LCR (Artificial Analysis Long Context Recall) evaluates a model's ability to accurately retrieve and recall information from long input contexts; IFBench is a benchmark for evaluating instruction-following capabilities across diverse and structured task constraints. <br>Inference:
Acceleration Engine: vLLM <br> Test Hardware: NVIDIA GB300 <br>Post Training Quantization
This model was obtained by quantizing the weights of Qwen3.6-35B-A3B to NVFP4 data type, ready for inference with vLLM. Only the weights and activations of the linear operators within transformer blocks in MoE are quantized. This optimization reduces the number of bits per parameter from 16 to 4, reducing the disk size and GPU memory requirements by approximately 3.06x.Usage
To serve this checkpoint with vLLM, you can start the docker vllm/vllm-openai:nightly and run the sample command below:
sh
vllm serve nvidia/Qwen3.6-35B-A3B-NVFP4 --port 8000 --quantization modelopt --max-model-len 262144 --reasoning-parser qwen3For NVIDIA DGX Spark, we recommend using this vllm serve command:
sh
vllm serve nvidia/Qwen3.6-35B-A3B-NVFP4 \
--host 0.0.0.0 \
--port 8000 \
--tensor-parallel-size 1 \
--trust-remote-code \
--kv-cache-dtype fp8 \
--attention-backend flashinfer \
--moe-backend marlin \
--gpu-memory-utilization 0.4 \
--max-model-len 262144 \
--max-num-seqs 4 \
--max-num-batched-tokens 8192 \
--enable-chunked-prefill \
--async-scheduling \
--enable-prefix-caching \
--speculative-config '{"method":"mtp","num_speculative_tokens":3,"moe_backend":"triton"}' \
--load-format fastsafetensors \
--reasoning-parser qwen3 \
--tool-call-parser qwen3_xml \
--enable-auto-tool-choice