The NVIDIA Qwen3.5-397B-A17B NVFP4 V2 model is the quantized version of Alibaba's Qwen3.5-397B-A17B model, which is an auto-regressive language model that uses an optimized transformer architecture. For more information, please check here. The NVIDIA Qwen3.5-397B-A17B NVFP4 V2 model is quantized with Model Optimizer.
This model is ready for commercial/non-commercial use.
Developers looking to take off-the-shelf, pre-quantized models for deployment in AI Agent systems, chatbots, RAG systems, and other AI-powered applications.
Architecture Type: Transformers Network Architecture: Qwen3.5-397B-A17B Number of Model Parameters: 397B in total and 17B activated
Input:
Input Type(s): Text, Image, Video Input Format(s): String, Red, Green, Blue (RGB), Video (MP4/WebM) Input Parameters: One-Dimensional (1D), Two-Dimensional (2D), Three-Dimensional (3D) Other Properties Related to Input: Context length up to 262K
Output:
Output Type(s): Text Output Format: String Output Parameters: 1D (One-Dimensional): Sequences Other Properties Related to Output: Context length up to 262K
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.
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 2.0 version and is quantized with nvidia-modelopt v0.45.0.dev173+g52f1ccbee
Training and Evaluation Datasets:
Calibration Dataset:
Link:cnn_dailymail, Nemotron-Post-Training-Dataset-v2 Data Collection Method by dataset: Automated. Labeling Method by dataset: Automated. 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.
Training Dataset:
Data Modality: Undisclosed Data Collection Method by dataset: Undisclosed Labeling Method by dataset: Undisclosed Properties: Undisclosed Audio Training Data Size: Undisclosed Image Training Data Size: Undisclosed Text Training Data Size: Undisclosed Video Training Data Size: Undisclosed Non-Audio, Image, Text Training Data Size: Undisclosed
Evaluation Dataset:
Datasets: MMMU Pro, GPQA Diamond, SciCode, AA-LCR, IFBench, Tau2 Bench Telecom
** Data Collection Method by dataset: Hybrid: Automated, Human
** Labeling Method by dataset: Hybrid: Human, Automated
** Properties: We evaluated the model on multimodal reasoning, coding, agentic tool-use, and instruction-following benchmarks: 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; GPQA Diamond contains 448 graduate-level multiple-choice questions written by domain experts in biology, physics, and chemistry; SciCode evaluates scientific coding capabilities; 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; Tau2 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.
Inference:
Acceleration Engine: SGLang Test Hardware: B300
Post Training Quantization
This model was obtained by quantizing the weights and activations of Qwen3.5-397B-A17B to NVFP4 data type, ready for inference with SGLang. The routed experts are quantized to NVFP4 using MSE based scale setting while the attention and shared experts are quantized to per-tensor FP8 format. Only the weights and activations of the linear operators within transformer blocks are quantized.
Usage
To serve this checkpoint with SGLang, you can start the docker lmsysorg/sglang:v0.5.12.post1-cu130 and run the sample command below:
Tip: add --mamba-ssm-dtype bfloat16 to store the Mamba SSM state in BF16 (default is FP32) for faster decode on Blackwell.
Evaluation
The accuracy benchmark results are presented in the table below:
Precision
MMMU Pro
GPQA Diamond
SciCode
AA-LCR
IFBench
Tau2 Bench Telecom
FP8
0.787
0.872
0.467
0.688
0.761
0.954
NVFP4 V2
0.784
0.877
0.481
0.678
0.765
0.952
Baseline: Qwen3.5-397B-A17B-FP8.
Benchmarked with temperature=0.6, top_p=0.95, max num tokens 64000. For tau2 bench telecom we use max num tokens 128000
Model Limitations:
The base model was trained on data that contains toxic language and societal biases originally crawled from the internet. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts. The model may generate answers that may be inaccurate, omit key information, or include irrelevant or redundant text producing socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive.
Ethical Considerations
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
Please report model quality, risk, security vulnerabilities or NVIDIA AI Concerns here.
SUBCARDS:
Explainability
Field:
Response:
Intended Task/Domain:
Text generation, reasoning, summarization, and question answering.
Model Type:
Text and Image-to-text transformer
Intended Users:
This model is intended for developers, researchers, and customers building/utilizing LLMs, while balancing accuracy and efficiency.
Output:
Text String(s)
Describe how the model works:
Generates text by predicting the next word or token based on the context provided in the input sequence using multiple self-attention layers
Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of:
Not Applicable
Technical Limitations & Mitigation:
The model was trained on data that contains toxic language and societal biases originally crawled from the internet. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts. Therefore, before deploying any applications of this model, developers should perform safety testing and tuning tailored to their specific applications of the model.
Verified to have met prescribed quality standards?
Yes
Performance Metrics:
Accuracy, Throughput, and user-side throughput
Potential Known Risk
The model may generate answers that may be inaccurate, omit key information, or include irrelevant or redundant text producing socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive.
The Principle of least privilege (PoLP) is applied limiting access for dataset generation. Restrictions enforce dataset access during training, and dataset license constraints adhered to. Model checkpoints are made available on Hugging Face, and may become available on cloud providers' model catalog.
Privacy
Field:
Response:
Generatable or Reverse engineerable personal data?
No
Personal data used to create this model?
No
Was consent obtained for any personal data used?
Not Applicable
How often is dataset reviewed?
Before Release
Was data from user interactions with the AI model (e.g. user input and prompts) used to train the model?
No
Is there provenance for all datasets used in training?
Yes
Does data labeling (annotation, metadata) comply with privacy laws?
Yes
Is data compliant with data subject requests for data correction or removal, if such a request was made?
