nvidia / Llama-4-Scout-17B-16E-Instruct-FP8

Model Overview

Description:

The NVIDIA Llama 4 Scout 17B 16E Instruct FP8 model is the quantized language model of the Meta's Llama 4 Scout 17B 16E model, which is an auto-regressive language model that uses a mixture-of-experts (MoE) architecture and incorporate early fusion for native multimodality. For more information, please check here .

This model is ready for commercial and non-commercial use.

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 (Llama-4-Scout-17B-16E) Model Card .

License/Terms of Use:

GOVERNING TERMS: Use of this model is governed by the NVIDIA Open Model License .
ADDITIONAL INFORMATION: Llama4 Community License Agreement. Built with Llama.

Deployment Geography:

Global, except in European Union

Use Case:

Developers looking to take off the shelf pre-quantized models for deployment in AI Agent systems, chatbots, RAG systems, and other AI-powered applications. .

Release Date:

Huggingface: Jul 10th, 2025 via [ https://huggingface.co/nvidia/Llama-4-Scout-17B-16E-Instruct-FP8]

Model Architecture:

Architecture Type: Transformers
Network Architecture: Llama4

Input:

Input Type(s): Multilingual text, and up to 5 images
Input Format(s): String, Images
Input Parameters: 1D, 2D
Other Properties Related to Input: Context length up to 1M

Output:

Output Type(s): Multilingual text and code
Output Format: String
Output Parameters: 1D
Other Properties Related to Output: Context length up to 1M

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.

Software Integration :

Supported Runtime Engine(s):

• TensorRT-LLM
  
• SGLang
  

Supported Hardware Microarchitecture Compatibility:

• NVIDIA Blackwell
  
• NVIDIA Hopper
  
• NVIDIA Ampere
  

[Preferred/Supported] Operating System(s):

• Linux
  

Model Version(s):

The model is quantized with nvidia-modelopt v0.33.0

Datasets:

• Calibration Dataset: cnn_dailymail
  ** Data collection method: Automated
  ** Labeling method: The dataset is labeled by having the news articles as input and the corresponding highlight as the gold label summary.
  

• Evaluation Dataset: MMMU Pro, GPQA Diamond, HLE, LiveCodeBench, SciCode, HumanEval, AIME 2024, MATH-500
  ** Data collection method: Hybrid: Automated, Human
  ** Labeling method: Hybrid: Human, Automated.
  

Inference:

Engine: TensorRT-LLM
Test Hardware: H100

Post Training Quantization

This model was obtained by quantizing the weights and activations of Llama 4 Scout 17B 16E Instruct to FP8 data type, ready for inference with TensorRT-LLM. Only the weights and activations of the linear operators within transformer blocks are quantized. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%.

Usage

Deploy with TensorRT-LLM

To serve the quantized checkpoint with TensorRT-LLM , follow the sample commands below with the TensorRT-LLM GitHub repo:

• LLM API sample usage:

import asyncio
from tensorrt_llm import LLM, SamplingParams


def main():

    llm = LLM(model="nvidia/Llama-4-Scout-17B-16E-Instruct-FP8", attn_backend="FLASHINFER", backend="pytorch", tensor_parallel_size=8)

    prompts = [
        "Hello, my name is",
        "The president of the United States is",
        "The capital of France is",
        "The future of AI is",
    ]

    sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
    async def task(prompt: str):
        output = await llm.generate_async(prompt, sampling_params)
        print(
            f"Prompt: {output.prompt!r}, Generated text: {output.outputs[0].text!r}"
        )

    async def main():
        tasks = [task(prompt) for prompt in prompts]
        await asyncio.gather(*tasks)

    asyncio.run(main())


# The entry point of the program need to be protected for spawning processes.
if __name__ == '__main__':
    main()

Evaluation

The accuracy benchmark results are presented in the table below:

Precision	MMMU Pro	GPQA Diamond	HLE Challenge	LiveCodeBench	MATH-500	AIME 2024
Llama-4-Scout-17B-16E-Instruct 1	75	57	4	36	82	30
Llama-4-Scout-17B-16E-Instruct-FP8	74	56	4	32	82	31

¹ Reference scores for Llama-4-Scout-17B-16E-Instruct sourced from artificialanalysis .

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. Developers should perform safety testing and tuning tailored to their specific applications of the model. 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.

For more detailed information on ethical considerations for this model, please see the Model Card++ Explainability, Bias, Safety & Security, and Privacy Subcards.

Please report security vulnerabilities or NVIDIA AI Concerns here .

SUBCARDS:

Explainability

Bias

Safety & Security

Privacy